Home |
Search |
Today's Posts |
#1
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz, here
I don't know if any of my fellow audio enthusiasts out there have
noticed this, but even the best recordings always seem to "lack" something. Uncompressed digital (even Red Book), promises wide dynamic range, excellent frequency response and low distortion. It should be possible to make recordings so good that, given a halfway decent playback system, the musicians are in the room with you. It is technically possible and surprisingly easy to do this, but it rarely happens with commercial recordings. Why is it that still, in this digital age, audiophiles cling to performances recorded more than fifty years ago as the pinnacle of the recording arts? Recordings made in the late 1950's and early 1960's by such people as Mercury Record's C. Robert Fine, or RCA Victor's Lewis Leyton in the classical recording world, and Rudy Van Gelder of Riverside, and Impulse Records fame in the world of jazz are held in such high esteem, that even CD and SACD re-releases of their recordings still sell very well today. It's as if no progress has been made in the art and science of recording in the last 55 years or so. I have found, in building my stereo system, that this has become a dog chasing his tail endeavor. My playback equipment gets better and better and yet the recordings to which I listen, ranging from terrible to OK seldom get any better than just OK. Even so-called audiophile recordings from labels such as Telarc and Reference and Naxos, to name a few, never sound quite as good as I think they should. Any ideas, other opinions, criticisms or nasty comments? All of the above would be welcomed. Audio_Empire |
#2
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz, here
Audio_Empire wrote:
I don't know if any of my fellow audio enthusiasts out there have noticed this, but even the best recordings always seem to "lack" something. Uncompressed digital (even Red Book), promises wide dynamic range, excellent frequency response and low distortion. It should be possible to make recordings so good that, given a halfway decent playback system, the musicians are in the room with you. It is technically possible and surprisingly easy to do this, but it rarely happens with commercial recordings. Why is it that still, in this digital age, audiophiles cling to performances recorded more than fifty years ago as the pinnacle of the recording arts? Because that sense of palpable realism is not the goal. Nobody other than a few hi-fi buffs cares about it, and that's not enough to create a mass market. Everybody else just wants to listen to the music, whether that's via headphones on the bus, over a car stereo, or whatever. In adverse listening conditions, all that compression and equalization helps. I have found, in building my stereo system, that this has become a dog chasing his tail endeavor. My playback equipment gets better and better and yet the recordings to which I listen, ranging from terrible to OK seldom get any better than just OK. Even so-called audiophile recordings from labels such as Telarc and Reference and Naxos, to name a few, never sound quite as good as I think they should. I think you're jaded. I'm listening to the Diamond Dogs remaster, and it sounds fabulous. But that's not "acoustic instruments in a real space", so I suppose it doesn't count. :-( Andrew. |
#3
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
On 4/2/2013 9:20 PM, Audio_Empire wrote:
Uncompressed digital (even Red Book), promises wide dynamic range, excellent frequency response and low distortion. It should be possible to make recordings so good that, given a halfway decent playback system, the musicians are in the room with you. It is technically possible and surprisingly easy to do this, but it rarely happens with commercial recordings. Why is it that still, in this digital age, audiophiles cling to performances recorded more than fifty years ago as the pinnacle of the recording arts? Its because producers don't want it. But there is a caveat: "the musicians are in the room with you" is only a realistic goal if they will fit. Few people have rooms big enough for an orchestra to fit, let alone produce a concert hall sized ambiance. For those cases, one wants ... and I presume you certainly do want ... to bring the sound of the recording concert hall (assuming its not Avery Fisher Hall) into your room. Both are possible, as I'm sure you will agree. Most real audiophiles agree that it is, more or less, with a good room and just the right equipment and recordings. I've never heard a concert hall in my room, but its too small; a friend has a system that really does the job (on the best recordings.) In most cases the problem is not compression or equalization, at least in most modern classical recordings. Large numbers of recordings are not compressed. And bad equalization seems a thing of the past: even recordings made in the 60s and pressed onto vinyl with horrendous cuts in the bass have reasonable balance on CD reissues. The problem is, as I'm sure you as a hyper-audiophile know, is microphone technique, pure and simple. You ask for a nasty comment, I'll make one: I'll bet if you look you can find nasty posts BY YOU complaining of mike technique. Doug McDonald |
#4
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz, here
On Wednesday, April 3, 2013 4:41:19 AM UTC-7, Andrew Haley wrote:
Audio_Empire wrote: I don't know if any of my fellow audio enthusiasts out there have noticed this, but even the best recordings always seem to "lack" something. Uncompressed digital (even Red Book), promises wide dynamic range, excellent frequency response and low distortion. It should be possible to make recordings so good that, given a halfway decent playback system, the musicians are in the room with you. It is technically possible and surprisingly easy to do this, but it rarely happens with commercial recordings. Why is it that still, in this digital age, audiophiles cling to performances recorded more than fifty years ago as the pinnacle of the recording arts? Because that sense of palpable realism is not the goal. Nobody other than a few hi-fi buffs cares about it, and that's not enough to create a mass market. Everybody else just wants to listen to the music, whether that's via headphones on the bus, over a car stereo, or whatever. In adverse listening conditions, all that compression and equalization helps. That may be true with regard to the various "pop" music categories, but as I said in my OP, for this discussion, those don't count. But, I'm sure that what you say, above, is right when discussing pop genres. OTOH, with classical and jazz, it's SO SIMPLE AND EASY to do it right and so complex (not to mention expensive) to do it wrong that I have to wonder why record companies continue to do it wrong and I've never heard a reasonable explanation of "why". I have found, in building my stereo system, that this has become a dog chasing his tail endeavor. My playback equipment gets better and better and yet the recordings to which I listen, ranging from terrible to OK seldom get any better than just OK. Even so-called audiophile recordings from labels such as Telarc and Reference and Naxos, to name a few, never sound quite as good as I think they should. I think you're jaded. I'm listening to the Diamond Dogs remaster, and it sounds fabulous. But that's not "acoustic instruments in a real space", so I suppose it doesn't count. :-( That's correct. It doesn't count. In fact, pop music is such a product of the studio, that what they do in the studio to create each group's "unique sound" actually amounts to more than just recording. It's not the same thing at all. Liken actual recording to the function of a court stenographer. It takes skill to do it and the result must be a true record of the everything said in court. On the other hand, a studio pop recording is like a novelist or playwright who attends a court trial and then creates a fictional play or novel based upon the incidents of that trial. The results may or may not be truthful to what the musicians were actually doing during their recording session. Audio_Empire |
#5
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
On Wednesday, April 3, 2013 10:57:02 AM UTC-7, Doug McDonald wrote:
On 4/2/2013 9:20 PM, Audio_Empire wrote: Uncompressed digital (even Red Book), promises wide dynamic range, excellent frequency response and low distortion. It should be possible to make recordings so good that, given a halfway decent playback system, the musicians are in the room with you. It is technically possible and surprisingly easy to do this, but it rarely happens with commercial recordings. Why is it that still, in this digital age, audiophiles cling to performances recorded more than fifty years ago as the pinnacle of the recording arts? Its because producers don't want it. But there is a caveat: "the musicians are in the room with you" is only a realistic goal if they will fit. Obviously I was being a bit hyperbolic, there. I was referring to the you-are-there sense that good recordings can provide if correctly done. With a small ensemble, like a string quartet or jazz group, it is possible to put them in the room with you. With an orchestra, of course, a good recording transports you to the venue where the recording takes place. Few people have rooms big enough for an orchestra to fit, let alone produce a concert hall sized ambiance. For those cases, one wants ... and I presume you certainly do want ... to bring the sound of the recording concert hall (assuming its not Avery Fisher Hall) into your room. Of course. Both are possible, as I'm sure you will agree. Most real audiophiles agree that it is, more or less, with a good room and just the right equipment and recordings. I've never heard a concert hall in my room, but its too small; a friend has a system that really does the job (on the best recordings.) In most cases the problem is not compression or equalization, at least in most modern classical recordings. Large numbers of recordings are not compressed. And bad equalization seems a thing of the past: even recordings made in the 60s and pressed onto vinyl with horrendous cuts in the bass have reasonable balance on CD reissues. Of course, and I agree fully that over-compression and hard-limiting are largely things of the past in classical recording. Most of these excesses started in the late '60's and went through the 70's and much of the 1980's before saner practices prevailed. It's a shame too as this time period became the last chance to capture some of the great artists of the 20th century before they left us. Those who were captured were recorded by what were (to me) incompetent knob twiddlers who simply had no idea (or didn't care) what they were doing to the music. Case in point. Ever heard Sir Adrian Boult's last 'recording of Holst's "The Planets" Made by EMI in either the late 70's or early 80's it was a great performance, perhaps the best ever - ruined by multimiking, multi-channeling and knob-twiddling. The problem is, as I'm sure you as a hyper-audiophile know, is microphone technique, pure and simple. You ask for a nasty comment, I'll make one: I'll bet if you look you can find nasty posts BY YOU complaining of mike technique. I don't doubt it. I welcome nasty comments. They add spice to the proceedings. |
#6
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz, here
Believe me AE, you can use your recordings as a benchmark for quality in
comparing to both old and new commercial recordings. What I mean is, this is sort of a chicken and egg question, similar to Floyd Toole's Circle of Confusion. You go out and make a recording, you come home and play it back. You make improvements to your technique for next time. You make improvements in your system so the recordings will sound better. But what is the reference, your system or the recordings? I suppose it will always be an iterative process, making adjustments like this until it sounds as close to what you heard live as possible in a smaller environment. But the commercial recordists have to make them sound good on any system that they may be played on, including boom boxes, iPods, and super systems. Quite a problem, and the result is boost and compression and multi-miking. Notice that they USED to advertise "no equalization or compression was used at any point in this recording." But we don't see that much any more. The biggest difference I notice since I started recording is that my stuff is a lot lower in volume than the commercial stuff. I go out and do my darndest, come home and master a disc, and it sounds terrific on my home system because I am listening in a quiet environment and I can adjust the gain to perfection and I have over 1500 watts of power and so on. THEN I decide to play it in my car - big mistake. I need to crank the gain up a lot more than the other sources that I have available, and it may sound OK but I wish I had learned a little more about compression and processing in Audition so it is a LITTLE louder to level the playing field. When my disc stops the FM or other discs BLAST me out and sound fairly good, but leave me scratching my head. I had been giving my recordings a touch of bass boost and of course gain setting in mastering, but lately I have discovered that my AT 2050s in omni mode do not need any bass boost, and my record levels are pretty damn good as is, at least on the screen during editing. If I adjust the gain up any, it might clip the peaks, tho it would raise the overall gain nicely, so I don't do it, and then I have the above problem. So what would John Eargle do? Bob Katz? Scott Dorsey? Gary Eickmeier |
#7
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz, here
Audio_Empire wrote:
On Wednesday, April 3, 2013 4:41:19 AM UTC-7, Andrew Haley wrote: Audio_Empire wrote: I don't know if any of my fellow audio enthusiasts out there have noticed this, but even the best recordings always seem to "lack" something. Uncompressed digital (even Red Book), promises wide dynamic range, excellent frequency response and low distortion. It should be possible to make recordings so good that, given a halfway decent playback system, the musicians are in the room with you. It is technically possible and surprisingly easy to do this, but it rarely happens with commercial recordings. Why is it that still, in this digital age, audiophiles cling to performances recorded more than fifty years ago as the pinnacle of the recording arts? Because that sense of palpable realism is not the goal. Nobody other than a few hi-fi buffs cares about it, and that's not enough to create a mass market. Everybody else just wants to listen to the music, whether that's via headphones on the bus, over a car stereo, or whatever. In adverse listening conditions, all that compression and equalization helps. That may be true with regard to the various "pop" music categories, but as I said in my OP, for this discussion, those don't count. But, I'm sure that what you say, above, is right when discussing pop genres. OTOH, with classical and jazz, it's SO SIMPLE AND EASY to do it right and so complex (not to mention expensive) to do it wrong that I have to wonder why record companies continue to do it wrong and I've never heard a reasonable explanation of "why". I think that my comment above applies to all forms of music. There is a conflict, expecially when listening in less than perfect conditions, between audibility and pure realism, with the dynamic range that implies. For example: when listening to a concerto in a live concert audibility of the soloist can suffer, but to a large extent visual cues make up for that. When recording a concerto it's usual, and IMO praiseworthy, to separately mike the soloist, and indeed each section of the orchestra, so that the balance can be adjusted afterwards. This isn't purist recording, and it may not capture the sound of the hall with absolute realism, but it is a better way to capture the music. And it's going to be a lot easier to listen to while on a train. Everyone but a hi-fi buff (who cares more about the ambience than the music) wins. Andrew. |
#8
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz, here
Gary Eickmeier wrote:
Believe me AE, you can use your recordings as a benchmark for quality in comparing to both old and new commercial recordings. What I mean is, this is sort of a chicken and egg question, similar to Floyd Toole's Circle of Confusion. You go out and make a recording, you come home and play it back. You make improvements to your technique for next time. You make improvements in your system so the recordings will sound better. But what is the reference, your system or the recordings? I suppose it will always be an iterative process, making adjustments like this until it sounds as close to what you heard live as possible in a smaller environment. Including the famous "seat-dip effect", which is a broad and maybe more than 10dB deep notch in the bass around 100Hz. I wouldn't have thought that was at all desirable, but some might. Andrew. |
#9
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz, here
"Audio_Empire" wrote in message
... I don't know if any of my fellow audio enthusiasts out there have noticed this, but even the best recordings always seem to "lack" something. Yes, indeed. Uncompressed digital (even Red Book), promises wide dynamic range, excellent frequency response and low distortion. Subject to the limitations of stereo recordings. Stereo recording has pronounced inherent limitiations. No way does one capture enough information to recreate the original sound field. It should be possible to make recordings so good that, given a halfway decent playback system, the musicians are in the room with you. How, given all of the audible information that is lost during recording with microphones? It is technically possible and surprisingly easy to do this, but it rarely happens with commercial recordings. I don't know about that. Why is it that still, in this digital age, audiophiles cling to performances recorded more than fifty years ago as the pinnacle of the recording arts? Recordings made in the late 1950's and early 1960's by such people as Mercury Record's C. Robert Fine, or RCA Victor's Lewis Leyton in the classical recording world, and Rudy Van Gelder of Riverside, and Impulse Records fame in the world of jazz are held in such high esteem, that even CD and SACD re-releases of their recordings still sell very well today. Sentimentality. Musical works and performances that can't possibly be re-recorded today. It's as if no progress has been made in the art and science of recording in the last 55 years or so. 2013 - 55 = 1958 Since 1958 there has essentially been one signfiicant upgrade to stereo recording which is digital media. I have found, in building my stereo system, that this has become a dog chasing his tail endeavor. My playback equipment gets better and better and yet the recordings to which I listen, ranging from terrible to OK seldom get any better than just OK. Even so-called audiophile recordings from labels such as Telarc and Reference and Naxos, to name a few, never sound quite as good as I think they should. Any ideas, other opinions, criticisms or nasty comments? All of the above would be welcomed. We need a better recording technology than stereo. We have a ready alternative called multitrack, 5.1, 7.1, whatever, but while it adds something, it still doesn't provide us with that something that only exists at the live recording. |
#10
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz, here
Arny Krueger wrote:
"Audio_Empire" wrote in message ... It should be possible to make recordings so good that, given a halfway decent playback system, the musicians are in the room with you. How, given all of the audible information that is lost during recording with microphones? Of course we have heard this said a few times before, but how about pinning that down a little better Arn? Exactly what "information" are you saying gets lost during recording? Perhaps you could approach this by dividing up the total sound "picture" that we experience into its component parts and analyzing which parts are missing (?). We need a better recording technology than stereo. We have a ready alternative called multitrack, 5.1, 7.1, whatever, but while it adds something, it still doesn't provide us with that something that only exists at the live recording. Multitrack is fine, and I play everything in surround, even if it is just DPL II trying to extract whatever ambience is contained in the recording. I also record in surrond sound sometimes, but I haven't succeeded completely in enhancing the just stereo recordings enough to bother with it. I think my mistake is recording from a single point in space, and it would be much better to record the surround from further back in the hall. Lot of reasons, but practical difficulties have prevented me from trying it yet. But anyway, I think that the main difference between the live and the playback is an acoustic one, not necessarily that some info gets lost during recording. To put it simply, no matter how many channels you have you cannot make your listening room sound like a much larger space by playing a recording of a larger space in it. Your speakers become another real sound source that interracts with your room and gives the game away, no matter what has been recorded. There is no simple solution to this outside of a technical laboratory, but I don't believe it is correct to say that the problem is that not enough info has been recorded. Very interesting sidebar, Floyd Toole told me once that he had heard full periphony Ambisonics played back in an anechoic chamber, and it sounded like IHL (Inside the Head Locatedness) rather than superb realism. I don't know if they had added many many more speakers if this problem would be fixed, but it is apparent to me that without a real room to support the sound localization it just cannot place itself outside the head in a real space. This is an acoustical problem, not one of "accuracy" or insufficient information. Gary Eickmeier |
#11
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz, here
"Gary Eickmeier" wrote in message
... Arny Krueger wrote: "Audio_Empire" wrote in message ... It should be possible to make recordings so good that, given a halfway decent playback system, the musicians are in the room with you. How, given all of the audible information that is lost during recording with microphones? Of course we have heard this said a few times before, but how about pinning that down a little better Arn? Exactly what "information" are you saying gets lost during recording? Consider a microphone sitting in a horizontal plane that is marked up with radial lines like a protractor. Approaching the microphone along each radial line is a different sound from a different part of the room. There are obviously dozens or hundreds of different lines depending on how different we want them to be. Each of these lines represents a separate channel of information. The output of the microphone is a single channel that represents all of these sounds multiplied by the sensitivity of the microphone in each direction. Obviously a tremendous amount of information has been lost. |
#12
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
Gary Eickmeier wrote:
Arny Krueger wrote: "Audio_Empire" wrote in message ... It should be possible to make recordings so good that, given a halfway decent playback system, the musicians are in the room with you. How, given all of the audible information that is lost during recording with microphones? Of course we have heard this said a few times before, but how about pinning that down a little better Arn? Exactly what "information" are you saying gets lost during recording? Perhaps you could approach this by dividing up the total sound "picture" that we experience into its component parts and analyzing which parts are missing (?). You're kidding, right? This stuff is hardly new, and wwas known back in the 1930's and before. ALL directional information iis lost in any single microphone. The output of the microphone is a simply two-dimensional record of instantaneous pressure or velocity amplitude vs time. That's it. There is no information in that electrical signal as to where the sound that caused it came from. None. As I said, even in a directional microphone, that information is irretrievably lost. Say a directional microphone is down 20 db 120 degrees relative to the principle axis. There's nothing in the resulting electrical stream that unambiguously (or even vaguely) provides a clue as to whether that signal was due to an 80 dB SPL sound on the principal axis or a 100 dB SPL sound 120 degrees off axis. And when you start to talk about recording in a complex sound field, the electrical output has NO indication AT ALL whether a direct sound came from there, while the reverberent sound came form over there. Now, take a stereo pair. The situation is really not any better It is geometrically impossible to disambiguate, for example, by any property in the elctrical signals, whether a source of a sound is anywhere on a circle whose center is defined by the line between the two microphones and whose plane is at right angles to that circle. Two omnis some distance apart will generate the SAME electrical signals whether the source is 20 feet ahead, 20 feet above, 20 feet behind or anywhere else on the circle. The same is true of any other mike position. The only position that can be unambiguously recorded is somewhere EXACTLY in between the two, which is arguably not very useful. Consider also the reciprocity principle as a gedanken (and, as a real-world excercise, if your want). Record something from a complex sound field with a microphone of your choosing. Now, play it back through the same microphone. While you're thinking about it, go study up on the reciprocity principle. Now, if your assertion were correct, recording with a single mike of several sound sources in different direction, should result, if you insist there is no loss of information, in the sound that emenates from that same microphone finding their way back to the original location they were emitted from. Now, does that happen? Does even the simplest version of that happen? But anyway, I think that the main difference between the live and the playback is an acoustic one, not necessarily that some info gets lost during recording. Uh, sorry, but it is the 3-dimensional aspect of the original acoustical field that is provably lost. The fact is that the HRTF of the original sound field is eliminate from the listening chain is precisely the problem. To put it simply, no matter how many channels you have you cannot make your listening room sound like a much larger space by playing a recording of a larger space in it. Your speakers become another real sound source that interracts with your room and gives the game away, no matter what has been recorded. There is no simple solution to this outside of a technical laboratory, but I don't believe it is correct to say that the problem is that not enough info has been recorded. No, sorry, this might not be the only problem, but it is the FIRST probloem, and unles syou solve it, everything else is parlor trick. The reason carefully done (and VERY inconvenient) binaural works is because it works VERY hard to try to preserve as much of the utility of the listener's HRTF as possible. But to do it right is VERY hard and only works, when done extremely well, for one specific listener. -- +--------------------------------+ + Dick Pierce | + Professional Audio Development | +--------------------------------+ |
#13
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
Dick Pierce wrote:
ALL directional information iis lost in any single microphone. The output of the microphone is a simply two-dimensional record of instantaneous pressure or velocity amplitude vs time. That's it. There is no information in that electrical signal as to where the sound that caused it came from. None. Now, take a stereo pair. The situation is really not any better It is geometrically impossible to disambiguate, for example, by any property in the elctrical signals, whether a source of a sound is anywhere on a circle whose center is defined by the line between the two microphones and whose plane is at right angles to that circle. Two omnis some distance apart will generate the SAME electrical signals whether the source is 20 feet ahead, 20 feet above, 20 feet behind or anywhere else on the circle. The same is true of any other mike position. The only position that can be unambiguously recorded is somewhere EXACTLY in between the two, which is arguably not very useful. Mr. Pierce, I can appreciate your confusion on this, and Arnie's statement that we can't record every direction from the microphones. But summing localization works, and works well to enable us to encode direction along a line between the microphones. We can tell where every instrument is along that line, and we can even get a good sense of depth and spaciousness if the recording was made right and the playback technique is good. The fact that we can't encode all directions of all sounds emitted during a performance is not a real problem if we don't NEED to record all that. Stay with me. Now, if your assertion were correct, recording with a single mike of several sound sources in different direction, should result, if you insist there is no loss of information, in the sound that emenates from that same microphone finding their way back to the original location they were emitted from. What assertion? You may have misread something I said. Now, does that happen? Does even the simplest version of that happen? YES! Of course it does! The fact is that the HRTF of the original sound field is eliminate from the listening chain is precisely the problem. Uh-oh - the binaural confusion rears its ugly "head" - so to speak. I would ask you (both) to try and flip your thinking around for a minute. What you are saying in essence is that it is impossible to take a 3D picture of a snowfall because you cannot see 4 ? steradians all around you. Well, that is not a problem if you don't need to see every direction around you to accomplish your intent. Consider the live vs recorded demos that were so successful. The reason for their success was that the acoustical problems were overcome by recording the instruments anechoically (outdoors) and then relying on the same acoustic space as the live instruments to make them sound alike. The change in thinking that this example illustrates is that there was no attempt or need to encode all directions of each instrument in another acoustic space, nor does HRTF have anything to do with stereophonic (field-type) reproduction. It would be possible, for example, to record each instrument individually and reproduce it by means of its own dedicated speaker, placing those speakers similarly to the original locations, in a room that is similar in size to the original. In this way you didn't need to encode every direction of arriving or acoustically resultant sound because you are going to be playing it back in a real acoustic space, which will in turn cause all of the effects of the complex sound field to occur in that space. You will use your own natural hearing to listen to all of those complex sounds, and apply your own exact HRTF to the sound in the process. We can then simplify the system down to as few as two speakers without losing all that much, because summing localization can place the whole frontal soundstage along a line between the speakers and the support of the playback room acoustic is still active. The fact that we have had a two channel system for so long and we happen to have two ears has screwed us up and caused so much confusion it may be impossible to overcome. Anyway, it's late, I'm tired and got another bid day ahead of me. Gary Eickmeier |
#14
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
Gary Eickmeier wrote:
Dick Pierce wrote: ALL directional information iis lost in any single microphone. The output of the microphone is a simply two-dimensional record of instantaneous pressure or velocity amplitude vs time. That's it. There is no information in that electrical signal as to where the sound that caused it came from. None. Now, take a stereo pair. The situation is really not any better It is geometrically impossible to disambiguate, for example, by any property in the elctrical signals, whether a source of a sound is anywhere on a circle whose center is defined by the line between the two microphones and whose plane is at right angles to that circle. Two omnis some distance apart will generate the SAME electrical signals whether the source is 20 feet ahead, 20 feet above, 20 feet behind or anywhere else on the circle. The same is true of any other mike position. The only position that can be unambiguously recorded is somewhere EXACTLY in between the two, which is arguably not very useful. Mr. Pierce, I can appreciate your confusion on this, No, you can't becasue you're viewing the world through the fog of your own self-created confusion. Now, if your assertion were correct, recording with a single mike of several sound sources in different direction, should result, if you insist there is no loss of information, in the sound that emenates from that same microphone finding their way back to the original location they were emitted from. What assertion? You may have misread something I said. No, you seem to dismiss that recording loses information. Now, does that happen? Does even the simplest version of that happen? YES! Of course it does! Extraordinary assertion: prove it. The fact is that the HRTF of the original sound field is eliminate from the listening chain is precisely the problem. Uh-oh - the binaural confusion rears its ugly "head" - so to speak. I would ask you (both) to try and flip your thinking around for a minute. What you are saying in essence is that it is impossible to take a 3D picture of a snowfall because you cannot see 4 ? steradians all around you. No, that's your confused interpretation: that's not what I said. If you want to play the optical analogy game, fine. Consider that a single microphone is NOT anything remotely like a single camera, rather it's FAR closer to a single photocell. The electrical output of the photocell is simply a record of the instantaneous light intensity vs time. Okay, so it's snowing. What information being emitted from the photocell places each snowflake in 3-d space. Assume you replace the photocell with a lightbulb. Now, play the photocell's electrical output back through the light bulb. According to YOUR statement: Now, if your assertion were correct, recording with a single mike of several sound sources in different direction, should result, if you insist there is no loss of information, in the sound that emenates from that same microphone finding their way back to the original location they were emitted from. Now, does that happen? Does even the simplest version of that happen? YES! Of course it does! that flickering light bulb should be painting pretty scene of our little snowstorm. Does it? Please, if you're going to answer, "of course it does," spare us your embarassment, if you will. -- +--------------------------------+ + Dick Pierce | + Professional Audio Development | +--------------------------------+ |
#15
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz, here
On Thursday, April 4, 2013 7:02:11 AM UTC-7, Arny Krueger wrote:
"Audio_Empire" wrote in message I don't know if any of my fellow audio enthusiasts out there have noticed this, but even the best recordings always seem to "lack" something. Yes, indeed. Uncompressed digital (even Red Book), promises wide dynamic range, excellent frequency response and low distortion. Subject to the limitations of stereo recordings. Stereo recording has pronounced inherent limitiations. No way does one capture enough information to recreate the original sound field. I disagree here, somewhat. We all know that absolute perfection, in either recording or playback is impossible. However the detours from good recording practices to which I'm referring are not due to limitations in either the recording or the playback technology. These less than satisfying recordings are the result of deliberate choices made at the time the recordings are captured. It should be possible to make recordings so good that, given a halfway decent playback system, the musicians are in the room with you. How, given all of the audible information that is lost during recording with microphones? Obviously, I'm talking about an illusion here. Of course an absolute virtual presence is impossible. But your comment seems to indicate that you don't really understand what I'm talking about. As a sometimes "recordist" yourself, if you've never made a recording that produced the illusion that, on playback, the musicians are there in the room with you (obviously we're talking small ensembles here - you aren't going to get an entire symphony orchestra in the room with you, virtually, or otherwise), then your recording practices are part of the problem, not part of the solution. It is technically possible and surprisingly easy to do this, but it rarely happens with commercial recordings. I don't know about that. I could play for you recordings that I have made, using simple (but high-quality) equipment which you would find very convincing. Why is it that still, in this digital age, audiophiles cling to performances recorded more than fifty years ago as the pinnacle of the recording arts? Recordings made in the late 1950's and early 1960's by such people as Mercury Record's C. Robert Fine, or RCA Victor's Lewis Leyton in the classical recording world, and Rudy Van Gelder of Riverside, and Impulse Records fame in the world of jazz are held in such high esteem, that even CD and SACD re-releases of their recordings still sell very well today. Sentimentality. Musical works and performances that can't possibly be re-recorded today. EHHHHH! Thanks for playing, Arnie. But you are wrong. Nostalgia and sentimentality are probably the LEAST significant factors here. The reason why audiophiles still purchase these 50-year-old-plus recordings is because they were simply recorded with straightforward gear directly to two or at most three channels with as simple a signal path as possible and they were properly miked using some minimalist miking technique and, as a result, they still SOUND GREAT. It's as if no progress has been made in the art and science of recording in the last 55 years or so. 2013 - 55 =3D 1958 Since 1958 there has essentially been one signfiicant upgrade to stereo recording which is digital media. Nothing wrong with digital. Properly applied, even old Red Book CD format can yield jaw-dropping results. The media's not the problem. Magnetic tape could do that too, even 55 years ago! I have found, in building my stereo system, that this has become a dog chasing his tail endeavor. My playback equipment gets better and better and yet the recordings to which I listen, ranging from terrible to OK seldom get any better than just OK. Even so-called audiophile recordings from labels such as Telarc and Reference and Naxos, to name a few, never sound quite as good as I think they should. Any ideas, other opinions, criticisms or nasty comments? All of the above would be welcomed. We need a better recording technology than stereo. I think that we are talking at cross purposes, here. My point is that we already have adequate tools to produce stupendous recordings of lasting audio merit. They just aren't being employed to give us what they are capable of. These tools have been perverted to other tasks, such as making recordings as loud as possible, all the time, or to make the recordist's job "easier" by allowing him or her to capture a performance of each instrument, separate form the others so that the musicians can be dismissed and the producers and engineers can vacillate over the balances and effects until the cows come home. We have a ready alternative called multitrack, 5.1, 7.1, whatever, but while it adds something, it still doesn't provide us with that something that only exists at the live recording. So-called surround systems might have something to add to pure music recordings, but if so, I've yet to hear it. They way I see it, most recording engineers working today haven't even mastered simple stereo, so I don't see how complicating the situation with more channels of information is going to improve that situation any. |
#16
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz, here
On Thursday, April 4, 2013 7:14:12 PM UTC-7, Barkingspyder wrote:
On Tuesday, April 2, 2013 7:20:55 PM UTC-7, Audio_Empire wrote: I don't know if any of my fellow audio enthusiasts out there have noticed this, but even the best recordings always seem to "lack" something. Uncompressed digital (even Red Book), promises wide dynamic range, excellent frequency response and low distortion. It should be possible to make recordings so good that, given a halfway decent playback system, the musicians are in the room with you. It is technically possible and surprisingly easy to do this, but it rarely happens with commercial recordings. Why is it that still, in this digital age, audiophiles cling to performances recorded more than fifty years ago as the pinnacle of the recording arts? Recordings made in the late 1950's and early 1960's by such people as Mercury Record's C. Robert Fine, or RCA Victor's Lewis Leyton in the classical recording world, and Rudy Van Gelder of Riverside, and Impulse Records fame in the world of jazz are held in such high esteem, that even CD and SACD re-releases of their recordings still sell very well today. It's as if no progress has been made in the art and science of recording in the last 55 years or so. I have found, in building my stereo system, that this has become a dog chasing his tail endeavor. My playback equipment gets better and better and yet the recordings to which I listen, ranging from terrible to OK seldom get any better than just OK. Even so-called audiophile recordings from labels such as Telarc and Reference and Naxos, to name a few, never sound quite as good as I think they should. Any ideas, other opinions, criticisms or nasty comments? All of the above would be welcomed. Audio_Empire My experience is that you are there feeling only comes from recording of a very few instruments playing, the fewer the better. Orchestral works are just impossible IME. There are a few audiophile recording companies that use purist techniques that give very good results, but it is always with things like quartets at most. I've had god luck with Sheffield Labs, Telarc, and Bainbridge. Your mileage may vary. I have recordings made by me and others of large ensembles which will challenge your assertion to the point of proving it wrong. I have recordings that one can, while listening to them, point to the location of every instrument in the group and even tell if some of the players are behind others or whether some are on risers or not! Image specificity, depth and height are all parameters that exist in the sound field and can be accurately picked-up and recorded by the use of correct microphone usage and placement. |
#17
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
On Thursday, April 4, 2013 4:27:06 PM UTC-7, Dick Pierce wrote:
Gary Eickmeier wrote: Arny Krueger wrote: "Audio_Empire" wrote in message ... It should be possible to make recordings so good that, given a halfway decent playback system, the musicians are in the room with you. How, given all of the audible information that is lost during recording with microphones? Of course we have heard this said a few times before, but how about pinning that down a little better Arn? Exactly what "information" are you saying gets lost during recording? Perhaps you could approach this by dividing up the total sound "picture" that we experience into its component parts and analyzing which parts are missing (?). You're kidding, right? This stuff is hardly new, and wwas known back in the 1930's and before. ALL directional information iis lost in any single microphone. The output of the microphone is a simply two-dimensional record of instantaneous pressure or velocity amplitude vs time. That's it. There is no information in that electrical signal as to where the sound that caused it came from. None. As I said, even in a directional microphone, that information is irretrievably lost. Say a directional microphone is down 20 db 120 degrees relative to the principle axis. There's nothing in the resulting electrical stream that unambiguously (or even vaguely) provides a clue as to whether that signal was due to an 80 dB SPL sound on the principal axis or a 100 dB SPL sound 120 degrees off axis. And when you start to talk about recording in a complex sound field, the electrical output has NO indication AT ALL whether a direct sound came from there, while the reverberent sound came form over there. Now, take a stereo pair. The situation is really not any better It is geometrically impossible to disambiguate, for example, by any property in the elctrical signals, whether a source of a sound is anywhere on a circle whose center is defined by the line between the two microphones and whose plane is at right angles to that circle. Two omnis some distance apart will generate the SAME electrical signals whether the source is 20 feet ahead, 20 feet above, 20 feet behind or anywhere else on the circle. The same is true of any other mike position. The only position that can be unambiguously recorded is somewhere EXACTLY in between the two, which is arguably not very useful. Are you talking about omnidirectional microphones here? Because they don't work as a stereo pair unless you take extraordinary precautions, such as placing a big sound baffle between them as Ray Kimber does for his IsoMike recordings. Consider also the reciprocity principle as a gedanken (and, as a real-world excercise, if your want). Record something from a complex sound field with a microphone of your choosing. Now, play it back through the same microphone. While you're thinking about it, go study up on the reciprocity principle. If you did do that, say, through a magnetic microphone, it wouldn't sound very good I'm afraid. It would likely sound much worse, even, than a telephone. And I don't see what this has to do with the subject at hand. Microphones are designed, to capture sound and turn it into an electronic analog of that sound, it is not designed to be a reproducer. Now, if your assertion were correct, recording with a single mike of several sound sources in different direction, should result, if you insist there is no loss of information, in the sound that emenates from that same microphone finding their way back to the original location they were emitted from. That's not possible. The microphone is not designed to reproduce anything. it would make a more than lousy speaker. Also recoding with a single mike will result in NO spatial information being captured (it's called "monaural sound"). One needs two mikes and the spatial information results from the difference between the two mike signals and THAT takes place in the listeners' ears. We hear in stereo due differences in phase, time delay, and spatial separation of signals reaching our ears. if done right, those cues can provide a very satisfactory soundstage on a good stereo system. |
#18
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
Audio_Empire wrote:
That's not possible. The microphone is not designed to reproduce anything. it would make a more than lousy speaker. Also recoding with a single mike will result in NO spatial information being captured (it's called "monaural sound"). One needs two mikes and the spatial information results from the difference between the two mike signals and THAT takes place in the listeners' ears. We hear in stereo due differences in phase, time delay, and spatial separation of signals reaching our ears. if done right, those cues can provide a very satisfactory soundstage on a good stereo system. OK, let's take another run at this. I'm sure Mr. Pierce didn't mean to imply that microphones can be reproducers; he was making a philosophical point. Nor did he catch my meaning in my prickly post. Audio Empire's last paragraph above gets a little off the path. Most textbooks describe how stereo works much like he did, with what is happening at the ears. I say that is a red herring, a mislead that confuses stereo with binaural. Stereo is not a head-related, ear input system, it is a field-type system in which two or more transducers make sound in a room. We then experience that sound with our natural hearing mechanism, our own HRTF, freq response anomolies, everything the same way we hear live sound at a concert. The key to improvement of reproduction systems is how closely those sound fields that are made by the speakers and room come to a typical live sound in a good hall. Binaural, on the other hand, if recorded with the classic binaural head placed in a good seat, requires only those two microphones, experiences the whole original sound field, and no information is lost, any more than if your were sitting there. Stereo is just a different system, but with both systems all of the sounds arrive at the microphones and no "information" is lost. It's all there, waiting to be reproduced. But once you take the headphones off and work in the stereophonic system, the sound will be on speakers in a room, and those speakers become another sound source, and your ears are free to hear them and their interraction with the room. What you hear about that source is mainly its (their) frequency response and radiation pattern - but not radiation pattern per se, but rather how those patterns interract with the surfaces around them. Doesn't matter whether you are playing pink noise, test clicks, or Pink Floyd, you ears can hear those characteristics of your speakers and room, and ain't nuthin you can do but try and make it sound as much like a live field as possible by studying this model of reproduction and working with it. The main difference between the live sound and the stereo repro is this summing localization that can place an auditory event anywhere along the line between the speakers (and beyond if you take advantage of the reflecting surfaces of the right and left side walls). However, as I said, this effect works and works well to place instruments on a soundstage where they belong in your room, even if there are only two mikes and two speakers, it can work quite amazingly. Surround and center are definite enhancements that I support fully as correct techniques in reconstructing a realistic sounding field in your room. Summary so I stop rambling, stereo recording can be described as "close miking the soundstage" in a way that will result in a realistic sound field in your room when played on speakers which have been placed in a position which is geometrically similar to the positions of the microphones that captured the sound, in front of you at a distance from you, making sound patterns that hopefully mimic those that were recorded, so that you and your natural hearing can then experience sound that has most of the spatial patterns and frequencies and timings, except for the simple fact that the time between reflections in your smaller space will be superimposed on those timings. In this way, the real sound sources within your room are once again creating all of those "lost" pieces of information that Arnie and Dick are talking about. But you should not think of this recreation as something "fake" or artificial in any way, it is just an acoustical fact of life in the field-type game called stereophonic - or multichannel, 5.1, 7.1, or whateve clever commercial name will be attached next. What we are doing with all of those channels is reconstructing sound fields within a room. If done right, all of the sounds that were there during recording will still be there, and will come from all of those complex angles and locations that you thought were lost, because this sound in your room is REAL and not a trick of the two ears being assaulted by those two "lost" signals. This way of thinking about the reproduction problem is very, very, very different from what most of us would pick up in the magazines or even the classic texts on stereo. Study not ear input signals and lost information, but rather making sound in rooms. Gary Eickmeier |
#19
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
On 4/7/2013 5:47 AM, Gary Eickmeier wrote:
Audio_Empire wrote: That's not possible. The microphone is not designed to reproduce anything. it would make a more than lousy speaker. Also recoding with a single mike will result in NO spatial information being captured (it's called "monaural sound"). One needs two mikes and the spatial information results from the difference between the two mike signals and THAT takes place in the listeners' ears. We hear in stereo due differences in phase, time delay, and spatial separation of signals reaching our ears. if done right, those cues can provide a very satisfactory soundstage on a good stereo system. OK, let's take another run at this. I'm sure Mr. Pierce didn't mean to imply that microphones can be reproducers; he was making a philosophical point. Nor did he catch my meaning in my prickly post. Seems to me he caught the meaning perfectly well. As I read it, the point was not philosophical at all, and the assumption in his gedanken was clearly that the microphones were as accurate as radiators as they are microphones. The point is that, even in that situation, where you're in the venue, you record and replay from the exact same points, and the mikes are perfectly omnidirectional, all directional information is lost. The signal is two dimensional, so the reverberant field input from, say 120 degrees, will replayed omnidirectionally. The directional, spacial, information is simply lost forever. Take the thought one step further. Suppose you could playback and record coincidentally, with the same assumptions above. What would you hear when you played back that second recording? A second reverberant field that included reflections of the original reverberant field, as well as the direct sound. Would this sound *more* accurate than the first playback? Of course not, it would compound the problem. This, in essence, is the method you're espousing, except you're not holding the venue constant. snip In this way, the real sound sources within your room are once again creating all of those "lost" pieces of information that Arnie and Dick are talking about. But you should not think of this recreation as something "fake" or artificial in any way, Leaving aside arguments hashed out previously, this statement is what IMO most people would disagree with. It is artificial, and those lost pieces of information will Not be recovered by anything you can do during reproduction. No matter what you do in an effort to retrieve it, it will be a simulation only - i.e. fake. it is just an acoustical fact of life in the field-type game called stereophonic - or multichannel, 5.1, 7.1, or whateve clever commercial name will be attached next. What we are doing with all of those channels is reconstructing sound fields within a room. If done right, all of the sounds that were there during recording will still be there, and will come from all of those complex angles and locations that you thought were lost, because this sound in your room is REAL and not a trick of the two ears being assaulted by those two "lost" signals. *And* it is not the sound heard at the recording venue. All the sound in your room is "real", but in no wise does that imply accuracy to the original event. This way of thinking about the reproduction problem is very, very, very different from what most of us would pick up in the magazines or even the classic texts on stereo. Study not ear input signals and lost information, but rather making sound in rooms. That's the way to make an illusion that *you* like, or believe is more realistic. It's not the path to the most accurate reproduction of the live event IME. Keith |
#20
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
On Sunday, April 7, 2013 5:47:12 AM UTC-7, Gary Eickmeier wrote:
Audio_Empire wrote: =20 =20 =20 That's not possible. The microphone is not designed to reproduce =20 anything. it would make a more than lousy speaker. =20 =20 Also recoding with a single mike will result in NO spatial information =20 being captured (it's called "monaural sound"). One needs two mikes and =20 the spatial information results from the difference between the two =20 mike signals and THAT takes place in the listeners' ears. We hear in =20 stereo due differences in phase, time delay, and spatial separation of =20 signals reaching our ears. if done right, those cues can provide a =20 very satisfactory soundstage on a good stereo system. =20 =20 =20 OK, let's take another run at this. I'm sure Mr. Pierce didn't mean to im= ply=20 =20 that microphones can be reproducers; he was making a philosophical point.= =20 =20 Nor did he catch my meaning in my prickly post. =20 =20 =20 Audio Empire's last paragraph above gets a little off the path. Most=20 =20 textbooks describe how stereo works much like he did, with what is happen= ing=20 =20 at the ears. I say that is a red herring, a mislead that confuses stereo= =20 =20 with binaural. Stereo is not a head-related, ear input system, it is a=20 =20 field-type system in which two or more transducers make sound in a room. = We=20 =20 then experience that sound with our natural hearing mechanism, our own HR= TF,=20 =20 freq response anomolies, everything the same way we hear live sound at a= =20 =20 concert. The key to improvement of reproduction systems is how closely th= ose=20 =20 sound fields that are made by the speakers and room come to a typical liv= e=20 =20 sound in a good hall. Oh, I disagree with that. All due respect, Gary. The stereo effect very muc= h involves the head and the ears. The mechanism formed ny our heads and our= ears (down to the shape of the latter, is very much responsible for how we= perceive the space around us, aurally. This includes directionality of sou= nd sources as well as the sense of whether we're enclosed by a large space or = a small one.=20 =20 =20 Binaural, on the other hand, if recorded with the classic binaural head= =20 =20 placed in a good seat, requires only those two microphones, experiences t= he=20 =20 whole original sound field, and no information is lost, any more than if= =20 =20 your were sitting there. Stereo is just a different system, but with both= =20 =20 systems all of the sounds arrive at the microphones and no "information" = is=20 =20 lost. It's all there, waiting to be reproduced. That's something else entirely. Biauaral uses surrogate ears that are repla= ced on playback by two ear-phones. What the mikes are doing there is intercepting the sound= at the point where it interacts with our heads and recording it. when play= ed back, it merely re-inserts the sound into the ear-head interface at the point it was intercepted upon recording.= I can give a very=20 convincing "you-are-there" illusion, but because everyone's head and ears a= re a bit different, it=20 isn't perfect. For instance, binaural sound cannot produce an image that th= e brain can interpret as a sound coming from behind the listener. Also Binaural doesn't work very we= ll as a stereo source to be listened to on speakers. That's the difference = between stereophonic sound and binaural sound.=20 But once you take the headphones off and work in the stereophonic system,= =20 =20 the sound will be on speakers in a room, and those speakers become anothe= r=20 =20 sound source, and your ears are free to hear them and their interraction= =20 =20 with the room. What you hear about that source is mainly its (their)=20 =20 frequency response and radiation pattern - but not radiation pattern per = se,=20 =20 but rather how those patterns interract with the surfaces around them.=20 =20 Doesn't matter whether you are playing pink noise, test clicks, or Pink= =20 =20 Floyd, you ears can hear those characteristics of your speakers and room,= =20 =20 and ain't nuthin you can do but try and make it sound as much like a live= =20 =20 field as possible by studying this model of reproduction and working with= =20 =20 it. That's true but it still doesn't conflate Binaural with stereo. =20 The main difference between the live sound and the stereo repro is this= =20 =20 summing localization that can place an auditory event anywhere along the= =20 =20 line between the speakers (and beyond if you take advantage of the=20 =20 reflecting surfaces of the right and left side walls). However, as I said= ,=20 =20 this effect works and works well to place instruments on a soundstage whe= re=20 =20 they belong in your room, even if there are only two mikes and two speake= rs,=20 =20 it can work quite amazingly. Surround and center are definite enhancement= s=20 =20 that I support fully as correct techniques in reconstructing a realistic= =20 =20 sounding field in your room. =20 =20 =20 Summary so I stop rambling, stereo recording can be described as "close= =20 =20 miking the soundstage" Again I disagree with your wording. Stereo is the capture of the soundfield= associated with an acoustic event. Correctly done, it consists of two micr= ophones "viewing" that event from two different perspectives. Reproduced, t= hose two soundfields interact with one another (and the space around them) = in such a way as to fool the ear into reconstructing a stereo image from th= ose two perspectives. Since the normally functioning human ear EXPECTS to h= ear two perspectives of a sonic event happening at a distance, it hears ste= reo. It even hears directionality when the source is a multi-miked and mult= ichannel recording played back through two speakers and tries to make stere= o from that recording.=20 An Interesting diregression illustrates the point of even accidental stereo= recording can provide a=20 satisfactory illusion.=20 In 1947, Motion Picture composer Alfred Newman ('The Robe', 'How The West W= as Won', 'Airport', etc.) was scoring the Fox film 'Captain From Castile' in the Fox scoring studio. = He told the engineers that he wanted the orchestra recorded from the two si= des of the room so that he could choose which perspective suited the action= best. The perspective favoring the strings would emphasize the romance asp= ect of the film, that perspective favoring the brass would emphasize the mi= litaristic aspect of the action. Newman had (we assume) NO knowledge of ste= reo and never played the two optical sound tracks back simultaneously. Anyw= ay, the film finished and the score laid-in, the optical two track=20 sound track was forgotten. Then about seven years ago, a small "record" company specializing in film s= oundtracks stumbled upon the original session film for 'Captain From Castil= e'. looking at the optical tracks, the engineer, expecting to see music on = one of the tracks and dialog or sound effects on the other, noticed that bo= th tracks seemed to have music on them. Not just music but what looked to h= im like slightly different versions of the same performance. He hooked the = multitrack optical reader up to amplifiers and speakers and VOILA!, a stere= o performance of the music. After using a computer to clean it up some and = do some digital EQ, the performance was released on CD. It has to be the ea= rliest real stereo performance ever released as a commercial recording! The stereo, BTW, is quite good and the = two-disc CD sounds a lot=20 better than one might expect given that it was recorded optically, not magn= etically on equipment probably pre-war in origin. BUT THE KICKER IS THAT TH= E STEREO IS TOTALLY ACCIDENTAL! |
#21
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
On Sunday, April 7, 2013 10:04:56 AM UTC-7, KH wrote:
*And* it is not the sound heard at the recording venue. All the sound in your room is "real", but in no wise does that imply accuracy to the original event. You're right, and for a number of reasons, not all of which are readily apparent to the casual observer. First of course is that microphones are far from perfect. Not only are they not perfectly flat in frequency response from DC to daylight, but they aren't even all that accurate to their advertised patterns. For instance, Omnis are not really omnidirectional. Cardioids do not fully supress sounds from the sides and rear as they are said to do. Figure-of- eight microphones do not attenuate sounds from the side by very much at all. The experienced and savvy recordist understands this and uses this knowledge to their advantage. But I makes recording a far different proposition than it looks like on the face of it. The fact that omnis are only more or less omni-directional accounts somewhat for the way that Mercury Living Presence recordings image. It also explains why Bob Woods of Telarc Records was unsuccessful at copying Mercury's technique. Mercury used three Telefunken omnidirectional microphones that was only "semi-omnidirectional". It was the best that they could do in the early-to-middle fifties when these Mercuries were made. OTOH, Woods used Schoeps calibration mikes which were not on;t true omnis, they were also the flattest mikes, frequency response-wise available at the time. They were nothing like the mikes C.R. Fine at Mercury used, and thus Woods did not get the same results. |
#22
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
Audio_Empire wrote:
On Sunday, April 7, 2013 5:47:12 AM UTC-7, Gary Eickmeier wrote: Oh, I disagree with that. All due respect, Gary. The stereo effect very much involves the head and the ears. The mechanism formed ny our heads and our ears (down to the shape of the latter, is very much responsible for how we perceive the space around us, aurally. This includes directionality of sound sources as well as the sense of whether we're enclosed by a large space or a small one. That says nothing about the system itself. In both ideas about how stereo works, we are listening with our ears. But if you think that stereo is a system of creating, or recording, ear input signals, then tell me what you are doing in your recording to put the HRTF and appropriate ear spacing and head attenuation into your signals? That's something else entirely. Biauaral uses surrogate ears that are replaced on playback by two ear-phones. What the mikes are doing there is intercepting the sound at the point where it interacts with our heads and recording it. when played back, it merely re-inserts the sound into the ear-head interface at the point it was intercepted upon recording. I can give a very convincing "you-are-there" illusion, but because everyone's head and ears are a bit different, it isn't perfect. For instance, binaural sound cannot produce an image that the brain can interpret as a sound coming from behind the listener. Also Binaural doesn't work very well as a stereo source to be listened to on speakers. That's the difference between stereophonic sound and binaural sound. WHAT is the difference? What I said? That binaural is an ear input system and stereo is a field-type system? and ain't nuthin you can do but try and make it sound as much like a live field as possible by studying this model of reproduction and working with it. That's true but it still doesn't conflate Binaural with stereo. It does if you are holding up the Mercury recordings as a great example of the art. If you think they were recording signals that were intended to sub for your ears if you had been there, then let me know who here has ears that stretch what - some 12 to 18 feet across the front of the orchestra and are suspended above the conductor's head. Again I disagree with your wording. Stereo is the capture of the soundfield associated with an acoustic event. Very good! Correctly done, it consists of two microphones "viewing" that event from two different perspectives. Reproduced, those two soundfields interact with one another (and the space around them) in such a way as to fool the ear into reconstructing a stereo image from those two perspectives. Since the normally functioning human ear EXPECTS to hear two perspectives of a sonic event happening at a distance, it hears stereo. It even hears directionality when the source is a multi-miked and multichannel recording played back through two speakers and tries to make stereo from that recording. Did you say TWO microphones? And TWO perspectives? Then we all have a serious problem! Those terrific three spaced omni recordings, first of all, anything done without a proper dummy head second of all, many jazz and classical recordings made with more than two mikes, such as highlight mikes, drum kit mikes, piano mikes, and vocal mikes for the soloist. Now we are told that our ears will be stretched across the orchestra, stuffed into a drum kit, placed under the lid of a piano, and shoved into the face of the singer or suspended above the chorus. I just have a problem with that theory of what we are doing with recording. Now lets have some fun with Alfred Newman.... An Interesting diregression illustrates the point of even accidental stereo recording can provide a satisfactory illusion. In 1947, Motion Picture composer Alfred Newman ('The Robe', 'How The West Was Won', 'Airport', etc.) was scoring the Fox film 'Captain From Castile' in the Fox scoring studio. He told the engineers that he wanted the orchestra recorded from the two sides of the room so that he could choose which perspective suited the action best. The perspective favoring the strings would emphasize the romance aspect of the film, that perspective favoring the brass would emphasize the militaristic aspect of the action. Newman had (we assume) NO knowledge of stereo and never played the two optical sound tracks back simultaneously. Anyway, the film finished and the score laid-in, the optical two track sound track was forgotten. Then about seven years ago, a small "record" company specializing in film soundtracks stumbled upon the original session film for 'Captain From Castile'. looking at the optical tracks, the engineer, expecting to see music on one of the tracks and dialog or sound effects on the other, noticed that both tracks seemed to have music on them. Not just music but what looked to him like slightly different versions of the same performance. He hooked the multitrack optical reader up to amplifiers and speakers and VOILA!, a stereo performance of the music. After using a computer to clean it up some and do some digital EQ, the performance was released on CD. It has to be the earliest real stereo performance ever released as a commercial recording! The stereo, BTW, is quite good and the two-disc CD sounds a lot better than one might expect given that it was recorded optically, not magnetically on equipment probably pre-war in origin. BUT THE KICKER IS THAT THE STEREO IS TOTALLY ACCIDENTAL! My first thought on this was that if these were two separate optical tracks, it would be impossible to sychronize them in order to play them together in stereo. But if they wre recorded on the same piece of film, of the same performance, then they already had a stereo recording sound head which, if what you say about their ignorance of stereo, would be unlikely. Nor would they record the music on one track and the sound effects or narration on another, so that proposed guess would be silly. Anyway, I would be curious about the full story on this, being a film maker myself. All I can think is that these two tracks must have been recorded on separate optical heads of the same performance, which were running in sync with each other by means of whatever technology was available at that time. I know they used optical film recorders the same way I used fullcoat to sync up with the camera. I just can't imagine anyone having a multichannel optical recorder for original field work. Gary Eickmeier |
#23
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
KH wrote:
On 4/7/2013 5:47 AM, Gary Eickmeier wrote: OK, let's take another run at this. I'm sure Mr. Pierce didn't mean to imply that microphones can be reproducers; he was making a philosophical point. Nor did he catch my meaning in my prickly post. Seems to me he caught the meaning perfectly well. As I read it, the point was not philosophical at all, and the assumption in his gedanken was clearly that the microphones were as accurate as radiators as they are microphones. The point is that, even in that situation, where you're in the venue, you record and replay from the exact same points, and the mikes are perfectly omnidirectional, all directional information is lost. The signal is two dimensional, so the reverberant field input from, say 120 degrees, will replayed omnidirectionally. The directional, spacial, information is simply lost forever. Curious misunderstanding of stereo. If there are two or more mikes recording phase locked signals of the same performance, then the two signals are what gives us stereo perspective. Not each mike. Both acting together. NOW - where did the reverberant field go? It's true that stereo has a front/back problem, like when the audience applause is folded back behind the performance in live recordings, but the reverberant field is still recorded, just not as much as if it were a binaural recording with the head placed further back into the audience. With stereo recordings, we depend on the playback room to support the reverberant field spatially. Most people's music rooms do not have much of a reverberant field of their own, so the resultant playback takes on most of the "flavor" of the recorded field. But it would be a mistake to use all kinds of absorbant materials in the room to kill the reflected sound. Take the thought one step further. Suppose you could playback and record coincidentally, with the same assumptions above. What would you hear when you played back that second recording? A second reverberant field that included reflections of the original reverberant field, as well as the direct sound. Would this sound more accurate than the first playback? Of course not, it would compound the problem. This, in essence, is the method you're espousing, except you're not holding the venue constant. Not sure I get your example, what you are doing there. But we all know about the "central recording problem," that in order to do stereo we are running the sound through two acoustic spaces. I am just saying that if you understand the system, the two spaces will complement each other rather than compete with each other. snip In this way, the real sound sources within your room are once again creating all of those "lost" pieces of information that Arnie and Dick are talking about. But you should not think of this recreation as something "fake" or artificial in any way, Leaving aside arguments hashed out previously, this statement is what IMO most people would disagree with. It is artificial, and those lost pieces of information will Not be recovered by anything you can do during reproduction. No matter what you do in an effort to retrieve it, it will be a simulation only - i.e. fake. Oh dear! Recordings fake? Say it ain't so Joe! it is just an acoustical fact of life in the field-type game called stereophonic - or multichannel, 5.1, 7.1, or whateve clever commercial name will be attached next. What we are doing with all of those channels is reconstructing sound fields within a room. If done right, all of the sounds that were there during recording will still be there, and will come from all of those complex angles and locations that you thought were lost, because this sound in your room is REAL and not a trick of the two ears being assaulted by those two "lost" signals. And it is not the sound heard at the recording venue. All the sound in your room is "real", but in no wise does that imply accuracy to the original event. We are not doing accuracy. You cannot have accuracy, because of the central recording problem. This way of thinking about the reproduction problem is very, very, very different from what most of us would pick up in the magazines or even the classic texts on stereo. Study not ear input signals and lost information, but rather making sound in rooms. That's the way to make an illusion that you like, or believe is more realistic. It's not the path to the most accurate reproduction of the live event IME. Keith Let me propose a little thought experiment riddle to you, Keith. You want to do a modern live vs recorded demo for some commercial purpose, maybe to sell speakers. You will use a saxophone, drum, and maybe trumpet. Doesn't matter. So you close mike the sax and trumpet, and you use two or three mikes on the drum kit if it has some spatial extent in the room. You do this in a very anechoic space, maybe outdoors like the original experiment. You then play back these tracks on speakers that have substantially the same radiation patterns as the instruments - the sax and drums mainly omni with the trumpet more directional. You find that if you place the live instrument side by side with those speakers, their sound is indistinguishable. Success, so you take your act on the road and amaze all and sundry. So first, would you agree that this would work? Would be a terrific experiment and very realistic? If so then note that the recording technique had nothing whatsoever to do with the human hearing mechanism, the spacing between the ears, the HRTF, none of it. You are usning the acoustics of the playback space in the same way as the live instruments, so they sound the same and very realistic. You have recorded not ear input signals but the object itself, the sound of the instruments and their radiation pattern, to be played back in a real room to make sound fields in that room, not to cast ear input signals toward you. Furthermore, everyone in the room will hear the same sound, each with his own HRTF and total hearing mechanism. And the goal is not "accuracy" but realism! So there you have an example that gives a more useful understanding of the nature of the system. Actual realistic stereo recording (the subject of this thread, after all) is a point on a continuum between this example and a dummy head recording placed well out into the audience and reverberant field and reproduced with crosstalk elimination etc etc. A "dry" recording is more like the example, a "wet" recording more toward the farther back limit. I sometimes think of Ralph Glasgal's system up in New York, which uses speakers in a surround sound arrangement with crosstalk elimination but in a real room. That might be a very interesting cross between the "you are there" binaural technique and the "they are here" stereo technique. Anyway, I would love to get the opportunity some fine day. Gary Eickmeier |
#24
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
On Monday, April 8, 2013 6:24:44 AM UTC-7, Gary Eickmeier wrote:
Audio_Empire wrote: On Sunday, April 7, 2013 5:47:12 AM UTC-7, Gary Eickmeier wrote: Oh, I disagree with that. All due respect, Gary. The stereo effect very much involves the head and the ears. The mechanism formed ny our heads and our ears (down to the shape of the latter, is very much responsible for how we perceive the space around us, aurally. This includes directionality of sound sources as well as the sense of whether we're enclosed by a large space or a small one. That says nothing about the system itself. In both ideas about how stereo works, we are listening with our ears. But if you think that stereo is a system of creating, or recording, ear input signals, then tell me what you are doing in your recording to put the HRTF and appropriate ear spacing and head attenuation into your signals? Again, you seem to be conflating binaural sound with stereo. They aren't the same thing. Binaural sound is about capturing and playing back the sound field as the EAR receives it, while stereo is about capturing and transmitting (or playing back in the case of a recording) the sound field as the musical ensemble MAKES it. Totally different concept. In one, the spatial characteristics and the interplay of the listener's head are already part of the signal, having been created at the binaural microphones and surrogate head. The only thing that the listener does into interpret the sound of someone (or more likely, someTHING else's head. In Stereo the user uses his head and his ear shape to interpret sound. That's one reason why binaural can't parse the difference between sounds occurring behind the listener and sounds that are supposed be in front. They all seem to come from in front of the listener, even when they are clearly supposed to be in back. In contrast, so-called "surround" stereo has no problem=20 placing images anywhere in the sound field that it cares to, and the listener can correctly perceive it as being in it's correct spot. That's something else entirely. Biauaral uses surrogate ears that are replaced on playback by two ear-phones. What the mikes are doing there is intercepting the sound at the point where it interacts with our heads and recording it. when played back, it merely re-inserts the sound into the ear-head interface at the point it was intercepted upon recording. I can give a very convincing "you-are-there" illusion, but because everyone's head and ears are a bit different, it isn't perfect. For instance, binaural sound cannot produce an image that the brain can interpret as a sound coming from behind the listener. Also Binaural doesn't work very well as a stereo source to be listened to on speakers. That's the difference between stereophonic sound and binaural sound. WHAT is the difference? What I said? That binaural is an ear input system and stereo is a field-type system? and ain't nuthin you can do but try and make it sound as much like a live field as possible by studying this model of reproduction and working with it. That's true but it still doesn't conflate Binaural with stereo. It does if you are holding up the Mercury recordings as a great example of the art. If you think they were recording signals that were intended to sub for your ears if you had been there, then let me know who here has ears that stretch what - some 12 to 18 feet across the front of the orchestra and are suspended above the conductor's head. No one is saying that, least of all me. I am an adherent of closely placed cardioid or crossed figure-of-eight microphones and do no not believe that one generally gets very good stereo from spaced omnis. Bob Fine of Mercury is an exception. The mikes (Telefunken U-47s) he used were advertised as omnidirectional, but they really weren't. The long lobe in the front of the mike was very wide, but frequency response fell-off very quickly on the side lobes and was somewhat less attenuated at the back. Yes, it picked-up sound from those lobes, but the mikes were more of a semi-cardioide than they were omnidirectional. That's why the Mercury system worked, and why, when Bob Woods of Telarc tried the same three-mike spaced omni arrangement in the late '70's, it didn't really work. Woods was using modern omnidirectional mikes which were REAL omnis and the result was that most Telarcs image very poorly. Again I disagree with your wording. Stereo is the capture of the soundfield associated with an acoustic event. Very good! Correctly done, it consists of two microphones "viewing" that event from two different perspectives. Reproduced, those two soundfields interact with one another (and the space around them) in such a way as to fool the ear into reconstructing a stereo image from those two perspectives. Since the normally functioning human ear EXPECTS to hear two perspectives of a sonic event happening at a distance, it hears stereo. It even hears directionality when the source is a multi-miked and multichannel recording played back through two speakers and tries to make stereo from that recording. Did you say TWO microphones? And TWO perspectives? Then we all have a serious problem! Those terrific three spaced omni recordings, first of all, anything done without a proper dummy head second of all, many jazz and classical recordings made with more than two mikes, such as highlight mikes, drum kit mikes, piano mikes, and vocal mikes for the soloist. Now we are told that our ears will be stretched across the orchestra, stuffed into a drum kit, placed under the lid of a piano, and shoved into the face of the singer or suspended above the chorus. I just have a problem with that theory of what we are doing with recording. Why do you keep bringing up dummy heads, Gary? They have nothing to do with stereo. Have you ever tried to listen to a binaural recording through speakers? It sounds awful (from a soundstage point of view, anyway), pretty much like mono Now lets have some fun with Alfred Newman.... An Interesting diregression illustrates the point of even accidental stereo recording can provide a satisfactory illusion. In 1947, Motion Picture composer Alfred Newman ('The Robe', 'How The West Was Won', 'Airport', etc.) was scoring the Fox film 'Captain From Castile' in the Fox scoring studio. He told the engineers that he wanted the orchestra recorded from the two sides of the room so that he could choose which perspective suited the action best. The perspective favoring the strings would emphasize the romance aspect of the film, that perspective favoring the brass would emphasize the militaristic aspect of the action. Newman had (we assume) NO knowledge of stereo and never played the two optical sound tracks back simultaneously. Anyway, the film finished and the score laid-in, the optical two track sound track was forgotten. Then about seven years ago, a small "record" company specializing in film soundtracks stumbled upon the original session film for 'Captain From Castile'. looking at the optical tracks, the engineer, expecting to see music on one of the tracks and dialog or sound effects on the other, noticed that both tracks seemed to have music on them. Not just music but what looked to him like slightly different versions of the same performance. He hooked the multitrack optical reader up to amplifiers and speakers and VOILA!, a stereo performance of the music. After using a computer to clean it up some and do some digital EQ, the performance was released on CD. It has to be the earliest real stereo performance ever released as a commercial recording! The stereo, BTW, is quite good and the two-disc CD sounds a lot better than one might expect given that it was recorded optically, not magnetically on equipment probably pre-war in origin. BUT THE KICKER IS THAT THE STEREO IS TOTALLY ACCIDENTAL! My first thought on this was that if these were two separate optical tracks, it would be impossible to sychronize them in order to play them together in stereo. Again you're letting your unfamiliarity with this stuff write checks that your keyboard can't cash. There were several types of optical film sound recorders developed for the motion picture industry. There was a 16mm, single track recorder developed in the early thirties, then there was a two track 16mm recorder that had two parallel optical heads on it. Then there was a 35mm recorder that had FOUR parallel optical tracks on it. For Gone With The Wind, for instance, the music was on track one, the dialog on track two and the sound effects were on track three. Track four had a time code signal on it. From those three sound tracks, the final mono track (on the edge of the release print) was mixed. Post-production mixing was invented by Hollywood. So you see, the two tracks didn't need to be 'synchronized' as they were both recorded to the same length of film. But if they wre recorded on the same piece of film, of the same performance, then they already had a stereo recording sound head which, if what you say about their ignorance of stereo, would be unlikely. Nor would they record the music on one track and the sound effects or narration on another, so that proposed guess would be silly. Anyway, I would be curious about the full story on this, being a film maker myself. All I can think is that these two tracks must have been recorded on separate optical heads of the same performance, which were running in sync with each other by means of whatever technology was available at that time. I know they used optical film recorders the same way I used fullcoat to sync up with the camera. I just can't imagine anyone having a multichannel optical recorder for original field work. I suggest that you read-up on Hollywood production methods in the '30's and '40's before making wild assumptions. For instance, when you said that the optical recorders in question were used "in the field", I'm wondering from what sta tement of mine did he glean that statement? Because I don't remember mentioning field recording in any way, shape, or form! The Fox Scoring stage in Culver City is not in any way "in the field". You are right. They did use two optical recording heads, But rather than use two separate machines, the two heads were recording to the same 16mm piece of film simultaneously on the SAME machine producing two parallel tracks. And there's nothing "silly" about recording all of the sound elements of a film (dialog, music, sound effects and folly) separately to a single piece of sound film It allows them to be mixed and edited TOGETHER before being transferred to the final cut of the film. |
#25
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
On Monday, April 8, 2013 6:25:34 AM UTC-7, Gary Eickmeier wrote:
KH wrote: On 4/7/2013 5:47 AM, Gary Eickmeier wrote: SNIP We are not doing accuracy. You cannot have accuracy, because of the central recording problem. What central recording problem? This way of thinking about the reproduction problem is very, very, very different from what most of us would pick up in the magazines or even the classic texts on stereo. Study not ear input signals and lost information, but rather making sound in rooms. That's the way to make an illusion that you like, or believe is more realistic. It's not the path to the most accurate reproduction of the live event IME. Keith Let me propose a little thought experiment riddle to you, Keith. You want to do a modern live vs recorded demo for some commercial purpose, maybe to sell speakers. You will use a saxophone, drum, and maybe trumpet. Doesn't matter. So you close mike the sax and trumpet, and you use two or three mikes on the drum kit if it has some spatial extent in the room. You do this in a very anechoic space, maybe outdoors like the original experiment. You then play back these tracks on speakers that have substantially the same radiation patterns as the instruments - the sax and drums mainly omni with the trumpet more directional. You find that if you place the live instrument side by side with those speakers, their sound is indistinguishable. Success, so you take your act on the road and amaze all and sundry. So first, would you agree that this would work? Would be a terrific experiment and very realistic? No it would not because the capture of the live instruments was done stereophonically. Now, if you took those musicians outdoors (on a quiet day) or into an anechoic chamber and recorded them with an MS or coincident pair (like the original) then played back the recording through the speakers being demonstrated WITH the ensemble on stage in the exact formation that they occupied when the recording was made, THEN it would work. It would also work, I hasten to add, using a slight modification of your proposal, if each instrument were miked separately (not to mention anechoically) and recorded to a mutitrack recorder of some type and then each track were played back through a separate amp it's own speaker (in the case of your example with a sax, drums and trumpet, that would be THREE speakers of the type being auditioned) with the musician being either beside or behind his instrument's speaker. it would also work, but in that case, it would do nothing to show off the imaging and sound-staging characteristics of the speakers being demonstrated. That's why Edgar Vilchur had the string quartet recorded with a coincident mike technique for his original Acoustic Research "live vs recorded" demonstrations of his AR3 speakers. |
#26
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
On 4/8/2013 6:25 AM, Gary Eickmeier wrote:
KH wrote: On 4/7/2013 5:47 AM, Gary Eickmeier wrote: OK, let's take another run at this. I'm sure Mr. Pierce didn't mean to imply that microphones can be reproducers; he was making a philosophical point. Nor did he catch my meaning in my prickly post. Seems to me he caught the meaning perfectly well. As I read it, the point was not philosophical at all, and the assumption in his gedanken was clearly that the microphones were as accurate as radiators as they are microphones. The point is that, even in that situation, where you're in the venue, you record and replay from the exact same points, and the mikes are perfectly omnidirectional, all directional information is lost. The signal is two dimensional, so the reverberant field input from, say 120 degrees, will replayed omnidirectionally. The directional, spacial, information is simply lost forever. Curious misunderstanding of stereo. Really? How so? I would say, rather, that you fail to grasp the basic point of the thought experiment. If there are two or more mikes recording phase locked signals of the same performance, then the two signals are what gives us stereo perspective. Not each mike. Both acting together. Clearly you did miss the point. It is entirely irrelevant, for the purposes of the 'experiment' whether there are two microphones, or just one (or X #). Let's look at the situation with one microphone; assuming the microphone is 100% accurate and 100% omnidirectional, further assuming the microphone is a perfectly accurate and perfectly omnidirectional speaker. Will the playback sound like the performance did, in that precise venue? No, it won't, because the directional information is lost, and all sounds, irrespective of what direction the sound originally came from, will be radiated in ALL directions. NOW - where did the reverberant field go? It was transformed from a 3-D sound pressure field to a 2-D electrical signal. I.e., it is gone. It's true that stereo has a front/back problem, like when the audience applause is folded back behind the performance in live recordings, but the reverberant field is still recorded, just not as much as if it were a binaural recording with the head placed further back into the audience. Well, no, there's not "less of it" in our thought experiment. It's just a different perspective, with varying levels based on distance. At the back wall, the reflections are higher in level, relative to the direct component, than in the front row. But since our microphone is perfectly omnidirectional it is still recording an accurate translation (from 3 to 2 dimensions) of the reverberant field, wherever we locate it. snip Take the thought one step further. Suppose you could playback and record coincidentally, with the same assumptions above. What would you hear when you played back that second recording? A second reverberant field that included reflections of the original reverberant field, as well as the direct sound. Would this sound more accurate than the first playback? Of course not, it would compound the problem. This, in essence, is the method you're espousing, except you're not holding the venue constant. Not sure I get your example, what you are doing there. But we all know about the "central recording problem," that in order to do stereo we are running the sound through two acoustic spaces. I am just saying that if you understand the system, the two spaces will complement each other rather than compete with each other. And I'm disagreeing with your assertion. Take the one microphone instance discussed above. It has recorded the reverberant field during the performance, it then replays that information in a 2-D rendition (time/pressure) broadcasting in all directions. Clearly, the directional clues that a listener, in any location in the venue, heard during the performance are gone. The playback will also generate a reverberant field that consists of all of the direct radiated sound from the performance, and all of the reflected sound from the performance, and the reflections of all of these sounds from the venue. The resulting sound field will not resemble the original soundfield because the directional information is not in the recording, and you're simply adding more reflections to an already inaccurate recording. In this way, the real sound sources within your room are once again creating all of those "lost" pieces of information that Arnie and Dick are talking about. But you should not think of this recreation as something "fake" or artificial in any way, Leaving aside arguments hashed out previously, this statement is what IMO most people would disagree with. It is artificial, and those lost pieces of information will Not be recovered by anything you can do during reproduction. No matter what you do in an effort to retrieve it, it will be a simulation only - i.e. fake. Oh dear! Recordings fake? Say it ain't so Joe! You're now arguing with yourself Gary. Do you think it's fake or not? Above you say not fake "in any way". snip And it is not the sound heard at the recording venue. All the sound in your room is "real", but in no wise does that imply accuracy to the original event. We are not doing accuracy. You cannot have accuracy, because of the central recording problem. You might want to define "central recording problem". In my lexicon, it means the loss of the ability to record the directional information in a stereo recording. The kind of recordings we have to live with. And "we" appear to be trying for the best accuracy to the original event as possible. You may be trying for something else. snip Let me propose a little thought experiment riddle to you, Keith. You want to do a modern live vs recorded demo for some commercial purpose, maybe to sell speakers. You will use a saxophone, drum, and maybe trumpet. Doesn't matter. So you close mike the sax and trumpet, and you use two or three mikes on the drum kit if it has some spatial extent in the room. You do this in a very anechoic space, maybe outdoors like the original experiment. You then play back these tracks on speakers that have substantially the same radiation patterns as the instruments - the sax and drums mainly omni with the trumpet more directional. You find that if you place the live instrument side by side with those speakers, their sound is indistinguishable. Success, so you take your act on the road and amaze all and sundry. So first, would you agree that this would work? Would be a terrific experiment and very realistic? No I don't agree. If each instrument were recorded separately on it's own channel, then the three instruments were played back on their own dedicated speakers, in the same spatial orientation as the original, with speakers with the same FR and radiation pattern, it could work. If so then note that the recording technique had nothing whatsoever to do with the human hearing mechanism, the spacing between the ears, the HRTF, none of it. In *your* version of the experiment, making a stereo recording, no, the HRTF has nothing to do with it. Which is why it wouldn't work. As I suggested, using individual recordings, on individual channels, recorded anechoically, it very much recognizes the HRTF in the recording process. It's a recognition that reverberant field information has to be excluded from the recording for the listeners' HRTF to process the information the same as it would the live instrument. Any reverberant field information from, e.g. the "sax speaker" that is not the sax, will make the speaker easily distinguishable from the sax on replay. That's pretty straightforward. You are usning the acoustics of the playback space in the same way as the live instruments, so they sound the same and very realistic. You have recorded not ear input signals but the object itself, the sound of the instruments and their radiation pattern, to be played back in a real room to make sound fields in that room, not to cast ear input signals toward you. Furthermore, everyone in the room will hear the same sound, No they won't hear the same. That's like saying every seat in the hall is the same. each with his own HRTF and total hearing mechanism. Yep, so they'll know whether the sax is on the left or right side of the "stage", and which instrument is closer/farther. And the goal is not "accuracy" but realism! If the playback sounds the same as the real instrument, that *is* accuracy. This is not, however, possible in the real world, with real world recordings. Try this with an orchestra - gets out of control real quick. And if you could do it, it would sound like an orchestra stuffed into your living room, and sound nothing like a concert. That's not the kind of "realism" I'm looking for. So there you have an example that gives a more useful understanding of the nature of the system. Well no, not really, IMO. Keith |
#27
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
Audio_Empire wrote:
On Monday, April 8, 2013 6:25:34 AM UTC-7, Gary Eickmeier wrote: [ Extraneous attributions snipped. -- dsr ] We are not doing accuracy. You cannot have accuracy, because of the central recording problem. What central recording problem? The one I discussed above. Let me propose a little thought experiment riddle to you, Keith. You want to do a modern live vs recorded demo for some commercial purpose, maybe to sell speakers. You will use a saxophone, drum, and maybe trumpet. Doesn't matter. So you close mike the sax and trumpet, and you use two or three mikes on the drum kit if it has some spatial extent in the room. You do this in a very anechoic space, maybe outdoors like the original experiment. You then play back these tracks on speakers that have substantially the same radiation patterns as the instruments - the sax and drums mainly omni with the trumpet more directional. You find that if you place the live instrument side by side with those speakers, their sound is indistinguishable. Success, so you take your act on the road and amaze all and sundry. So first, would you agree that this would work? Would be a terrific experiment and very realistic? No it would not because the capture of the live instruments was done stereophonically. No, it was not. Each instrument was close-miked separately. Now, if you took those musicians outdoors (on a quiet day) or into an anechoic chamber and recorded them with an MS or coincident pair (like the original) then played back the recording through the speakers being demonstrated WITH the ensemble on stage in the exact formation that they occupied when the recording was made, THEN it would work. It would also work, I hasten to add, using a slight modification of your proposal, if each instrument were miked separately (not to mention anechoically) and recorded to a mutitrack recorder of some type and then each track were played back through a separate amp it's own speaker (in the case of your example with a sax, drums and trumpet, that would be THREE speakers of the type being auditioned) with the musician being either beside or behind his instrument's speaker. it would also work, but in that case, it would do nothing to show off the imaging and sound-staging characteristics of the speakers being demonstrated. That's why Edgar Vilchur had the string quartet recorded with a coincident mike technique for his original Acoustic Research "live vs recorded" demonstrations of his AR3 speakers. Very good AE - but I think you are reading my stuff and then thinking that they are your own thoughts. No complaints here, as long as we are getting closer. Now let's press on to Alfred Newman again... Gary Eickmeier |
#28
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
Audio_Empire wrote:
On Monday, April 8, 2013 6:24:44 AM UTC-7, Gary Eickmeier wrote: That says nothing about the system itself. In both ideas about how stereo works, we are listening with our ears. But if you think that stereo is a system of creating, or recording, ear input signals, then tell me what you are doing in your recording to put the HRTF and appropriate ear spacing and head attenuation into your signals? Again, you seem to be conflating binaural sound with stereo. They aren't the same thing. Binaural sound is about capturing and playing back the sound field as the EAR receives it, while stereo is about capturing and transmitting (or playing back in the case of a recording) the sound field as the musical ensemble MAKES it. Totally different concept. In one, the spatial characteristics and the interplay of the listener's head are already part of the signal, having been created at the binaural microphones and surrogate head. The only thing that the listener does into interpret the sound of someone (or more likely, someTHING else's head. In Stereo the user uses his head and his ear shape to interpret sound. That's one reason why binaural can't parse the difference between sounds occurring behind the listener and sounds that are supposed be in front. They all seem to come from in front of the listener, even when they are clearly supposed to be in back. In contrast, so-called "surround" stereo has no problem=20 placing images anywhere in the sound field that it cares to, and the listener can correctly perceive it as being in it's correct spot. Again, very good, but still reading my stuff and claiming it as your own thoughts. Welcome aboard! Now where is Pierce? Did you say TWO microphones? And TWO perspectives? Then we all have a serious problem! Those terrific three spaced omni recordings, first of all, anything done without a proper dummy head second of all, many jazz and classical recordings made with more than two mikes, such as highlight mikes, drum kit mikes, piano mikes, and vocal mikes for the soloist. Now we are told that our ears will be stretched across the orchestra, stuffed into a drum kit, placed under the lid of a piano, and shoved into the face of the singer or suspended above the chorus. I just have a problem with that theory of what we are doing with recording. Why do you keep bringing up dummy heads, Gary? They have nothing to do with stereo. Have you ever tried to listen to a binaural recording through speakers? It sounds awful (from a soundstage point of view, anyway), pretty much like mono Because you keep saying two perspectives with two microphones. But I think we are in agreement now that stereo and binaural are two totally separate systems. I point out that we must not confuse the two, because recording and playback techniques are very different for each, and audiophiles keep trying to play their LP collections on two speakers because they have two ears and no brain. (That's a joke, moderators). Now lets have some fun with Alfred Newman.... My first thought on this was that if these were two separate optical tracks, it would be impossible to sychronize them in order to play them together in stereo. Again you're letting your unfamiliarity with this stuff write checks that your keyboard can't cash. There were several types of optical film sound recorders developed for the motion picture industry. There was a 16mm, single track recorder developed in the early thirties, then there was a two track 16mm recorder that had two parallel optical heads on it. Then there was a 35mm recorder that had FOUR parallel optical tracks on it. For Gone With The Wind, for instance, the music was on track one, the dialog on track two and the sound effects were on track three. Track four had a time code signal on it. From those three sound tracks, the final mono track (on the edge of the release print) was mixed. Post-production mixing was invented by Hollywood. So you see, the two tracks didn't need to be 'synchronized' as they were both recorded to the same length of film. That's terrific knowledge, but you don't record original sound that way. Dialog would be recorded during filming, and even possibly looped afterward and mixed in later. Music would be recorded in a sound stage while the conductor watched a work print for timing. Foley effects would be each recorded separately and mixed in later. All of these tracks are then mixed down to a master and printed on an optical track to make the final release prints. They did not and could not record all of the sounds for Gone With the Wind simultaneously during filming. I know you realize that, but if you are saying that they put music, effects, and dialog on the same piece of 35mm for the mixdown, I hope you also realize that those tracks would not be recorded live, possibly not even simultaneously, but one at a time, running the film thru 3 or four times in sync with the editor's tracks. But if they wre recorded on the same piece of film, of the same performance, then they already had a stereo recording sound head which, if what you say about their ignorance of stereo, would be unlikely. Nor would they record the music on one track and the sound effects or narration on another, so that proposed guess would be silly. Anyway, I would be curious about the full story on this, being a film maker myself. All I can think is that these two tracks must have been recorded on separate optical heads of the same performance, which were running in sync with each other by means of whatever technology was available at that time. I know they used optical film recorders the same way I used fullcoat to sync up with the camera. I just can't imagine anyone having a multichannel optical recorder for original field work. I suggest that you read-up on Hollywood production methods in the '30's and '40's before making wild assumptions. For instance, when you said that the optical recorders in question were used "in the field", I'm wondering from what sta tement of mine did he glean that statement? Because I don't remember mentioning field recording in any way, shape, or form! The Fox Scoring stage in Culver City is not in any way "in the field". "The field" doesn't mean out in the country somewhere. It just means original recording. When they record the music that is original recording. I used cassette tape in sync with my camera, or sometimes fullcoat. I would then come home and sync up the fullcoat with the movie film, frame for frame on a vertical editing bench. Hollywood used Moviolas for a long time, then probably flatbeds, but the idea is always the same - each track is built separately and edited in by a film editor, laying in music, effects, and narration where he wants them and choosing sync sound and shots from the many takes they shoot in the field. Sync stereo recording was a quest of mine, and I finally produced several surround sound sync films running a cassette recorder in sycn with the projector which had its own mag tracks on it. You are right. They did use two optical recording heads, But rather than use two separate machines, the two heads were recording to the same 16mm piece of film simultaneously on the SAME machine producing two parallel tracks. And there's nothing "silly" about recording all of the sound elements of a film (dialog, music, sound effects and folly) separately to a single piece of sound film It allows them to be mixed and edited TOGETHER before being transferred to the final cut of the film. No, it's just not done that way. They may have used such a scheme for the final mixdown in the dubbing stage, but if they had multi track recorders for live recording, then they already had stereo recorders. There is no conceivable reason to have more than one track in a live sound recorder except to be able to record in stereo. I will try and Google up that story and see what was really happening and how they managed to record two sync tracks of the same performance before stereo. My guess is two separate recorders running in sync. Gary Eickmeier |
#29
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
Keith,
To be brief and not agonize this a lot more, let me just summarize your and my points and then I must get to work! In my thought experiment with the individually miked instruments, played back on speakers placed the same and with similar radiation patterns to the instruments, you first told me that I was wrong and then you repeated my version as how you would do it and agreed that it would work. I set this as an example of a very realistic reproduction of the instruments' sound. It would not be "accurate" to the sound of those instruments played in any particular hall or stage, because of the central recording problem, that you have to run the sound thru two acoustic spaces before it reaches you. Then I noted that we aren't worried about that, because we are "doing" accuracy in this example, we are doing realism. Those "lost" spatial patterns of the reverberant field would be present in their entirety in your playback room, but they are not the patterns of any recording room, they are those of your playback room. Agreed? Not fair, because your goal is to be able to record some original hall and play it so that you think you are there. Right? OK, NOW - all I am saying is you can imagine a continuum with the above example at one extreme, one in which there is NO acoustic recorded and 100% of what you hear for an acoustic is that of your room, and is real and taking on 100% of the duties of supporting the good sound that you are getting, that very realistic playback of someone's performance. That is one end of the continuum. The other end is a very wet recording with whatever recording mikes you want to use placed back in the audience where the good seats are in a mistaken attempt to record 100% the space of the live sound, and then you will try to play it back in a nearly anechoic very dry room, as some poor dumb *******s attempt all the time. We all know (heh) that that doesn't work. You will say that it is because that spatial info will be lost because we don't have enough microphones and speakers to make all of those patterns come back. OK, fine, great dodge, but in reality WE DON'T DO IT THAT WAY. We place the mikes much closer to the instruments than you would sit so listen in that hall, then we play it back in another room at a distance from you so that the two spaces complement each other and lead to greater realism even tho we cannot have "accuracy" of what was heard if you were there. Accuracy would mean you would hear a concert from 9 ft above the conductor's head, or from ears that are 18 ft apart, or some such reductio ad absurdum. I have described the idea as close-miking the soundstage, because we are recording not just the actual instruments but also the early reflected sound from the sidewalls, the most important reflections, and also a hint of the reverberant field. All of this info mixes with the playback acoustic to give you the realism of sitting there with them in front of you, and also the "flavor" of the live acoustic space. OK, so it wasn't very brief, but there is a lot more to it than even that. Bottom line, AE is correct that it is possible to record for greater realism, but the techniques and reasons may be surprising to both of you, and progress toward a goal of greater realism may take a path a little - well, a lot - different from what most of us assume. That path is NOT a search for greater and greater "accuracy," but rather trying to work with and understand what we are actually doing with a field-type system, which is making sound in rooms, not making sound in your ears. Gary Eickmeier |
#30
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
On Tuesday, April 9, 2013 4:45:05 AM UTC-7, Gary Eickmeier wrote:
Audio_Empire wrote: On Monday, April 8, 2013 6:24:44 AM UTC-7, Gary Eickmeier wrote: Now lets have some fun with Alfred Newman.... My first thought on this was that if these were two separate optical tracks, it would be impossible to sychronize them in order to play them together in stereo. Again you're letting your unfamiliarity with this stuff write checks that your keyboard can't cash. There were several types of optical film sound recorders developed for the motion picture industry. There was a 16mm, single track recorder developed in the early thirties, then there was a two track 16mm recorder that had two parallel optical heads on it. Then there was a 35mm recorder that had FOUR parallel optical tracks on it. For Gone With The Wind, for instance, the music was on track one, the dialog on track two and the sound effects were on track three. Track four had a time code signal on it. From those three sound tracks, the final mono track (on the edge of the release print) was mixed. Post-production mixing was invented by Hollywood. So you see, the two tracks didn't need to be 'synchronized' as they were both recorded to the same length of film. That's terrific knowledge, but you don't record original sound that way. Not any more, no. But if the original sound track, performed in a cinema sound recording studio was not recorded on film, how was it recorded. Before answering, keep in mind that the music, along with the dialog, and all sound effects MUST be editable. Remember, audio tape, even though it had been invented in 1947, was not being used in movie studios yet. Dialog would be recorded during filming, and even possibly looped afterward Very true, but that has nothing whatsoever with what I'm talking about. and mixed in later. Music would be recorded in a sound stage while the conductor watched a work print for timing. Absolutely, and THIS is the two-track optical recording that I'm talking about. Foley effects would be each recorded separately and mixed in later. All of these tracks are then mixed down to a master and printed on an optical track to make the final release prints. But they were consolidated FIRST onto a single, four track piece of 35 mm film into what is called a conformance print. They did not and could not record all of the sounds for Gone With the Wind simultaneously during filming. Who said that the did? I know you realize that, but if you are saying that they put music, effects, and dialog on the same piece of 35mm for the mixdown, I hope you also realize that those tracks would not be recorded live, possibly not even simultaneously, but one at a time, running the film thru 3 or four times in sync with the editor's tracks. Of course I know that. Why even bring it up? But if they wre recorded on the same piece of film, of the same performance, then they already had a stereo recording sound head It wasn't designed as a stereo record head. It was designed as a two-track sound head. Stereo in film hadn't been invented yet and wouldn't be until "This is Cinerama" 1952. which, if what you say about their ignorance of stereo, would be unlikely. Nor would they record the music on one track and the sound effects or narration on another, so that proposed guess would be silly. Anyway, I would be curious about the full story on this, being a film maker myself. All I can think is that these two tracks must have been recorded on separate optical heads of the same performance, which were running in sync with each other by means of whatever technology was available at that time. I know they used optical film recorders the same way I used fullcoat to sync up with the camera. I just can't imagine anyone having a multichannel optical recorder for original field work. I suggest that you read-up on Hollywood production methods in the '30's and '40's before making wild assumptions. For instance, when you said that the optical recorders in question were used "in the field", I'm wondering from what sta tement of mine did he glean that statement? Because I don't remember mentioning field recording in any way, shape, or form! The Fox Scoring stage in Culver City is not in any way "in the field". "The field" doesn't mean out in the country somewhere. It just means original recording. When they record the music that is original recording. I used cassette tape in sync with my camera, or sometimes fullcoat. I would then come home and sync up the fullcoat with the movie film, frame for frame on a vertical editing bench. Hollywood used Moviolas for a long time, then probably flatbeds, but the idea is always the same - each track is built separately and edited in by a film editor, laying in music, effects, and narration where he wants them and choosing sync sound and shots from the many takes they shoot in the field. Yes, of course. For years (before portable digital) Hollywood used a specially modified (with a SMPTE time code added) Sony WM-D6 servo capstaned cassette recorder to grab sound on location. But the equipment used to record a musical score doesn't have to move. Like any recording studio, it's fixed in the control room. Sync stereo recording was a quest of mine, and I finally produced several surround sound sync films running a cassette recorder in sycn with the projector which had its own mag tracks on it. Not necessary when all the tracks are recorded simultaneously on the same media. You are right. They did use two optical recording heads, But rather than use two separate machines, the two heads were recording to the same 16mm piece of film simultaneously on the SAME machine producing two parallel tracks. And there's nothing "silly" about recording all of the sound elements of a film (dialog, music, sound effects and folly) separately to a single piece of sound film It allows them to be mixed and edited TOGETHER before being transferred to the final cut of the film. No, it's just not done that way. You were there in 1939, perhaps? You've seen the sound master? Held it in your hands, perhaps? I have. It is BECAUSE this master existed that it was possible to replace the music track with a newly recorded stereo track for a 1970's re-release. It was easy to separate the music, dialog, and sound effects so that a stereo print could be made. They "panned" the dialog and the sound effects to follow the action. They also played with the aspect ratio of the picture to make it widescreen, but that's another story. They may have used such a scheme for the final mixdown in the dubbing stage, Isn't THAT what I said???!!! but if they had multi track recorders for live recording, then they already had stereo recorders. To use for WHAT? There was no stereo in film. Disney did play with multi-channel sound for Fantasia in 1941, but it wasn't stereo. You seem not to believe me on this. Why don't you look up a Westrex 1581A Photographic Film Recorder and get back to me. While you are at it, Check SAE-Records Catalogue Number CRS-0007, Complete ORIGINAL STEREO film soundtrack from the 1947 20th Century Fox production of "Captain From Castile". http://www.discogs.com/Alfred-Newman...elease/3594394 They had multitrack recorders what they used them for likely changed with the film being produced. There is no conceivable reason to have more than one track in a live sound recorder except to be able to record in stereo. Stereo hadn't been "invented" yet. I will try and Google up that story and see what was really happening and how they managed to record two sync tracks of the same performance before stereo. My guess is two separate recorders running in sync. There were so many multitrack photographic recorders made in those days. Westrex. Mauer RCA, etc. None had much response above about 7KHz. The producers of the "Captain from Castile" Two CD set, used some digital "enhancement" to autocorrelate the noise and to boost what High- frequencies present on the film, but the results aren't all that bad, especially for "accidental" stereo. |
#31
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
On 4/9/2013 8:58 AM, Gary Eickmeier wrote:
Keith, To be brief and not agonize this a lot more, let me just summarize your and my points and then I must get to work! If only you would "summarize" accurately, that might help. Alas... In my thought experiment with the individually miked instruments, played back on speakers placed the same and with similar radiation patterns to the instruments, you first told me that I was wrong Uhmm, no. You talked about close-miking the soundfield, as you do below. You used multiple mikes on the drum kit, for example, to get the "space". Basically, you described a partially close-mike stereo recording. and then you repeated my version as how you would do it and agreed that it would work. Only when recorded anechoically, which you certainly did not describe. They are two separate scenarios whether you realize it or not. I set this as an example of a very realistic reproduction of the instruments' sound. It would not be "accurate" to the sound of those instruments played in any particular hall or stage, because of the central recording problem, that you have to run the sound thru two acoustic spaces before it reaches you. OK, fine, at least you defined your term. Then I noted that we aren't worried about that, because we are "doing" accuracy in this example I assume you mean "...not doing..." , we are doing realism. Those "lost" spatial patterns of the reverberant field would be present in their entirety in your playback room, No, *different* patterns will be *generated* in the playback room. Again, two very disparate things indeed. but they are not the patterns of any recording room, they are those of your playback room. Agreed? Not fair, because your goal is to be able to record some original hall and play it so that you think you are there. Right? Yes. And it's clear that is NOT your goal, yet you seem to think any goal not your own is misguided. OK, NOW - all I am saying is you can imagine a continuum with the above example at one extreme, one in which there is NO acoustic recorded and 100% of what you hear for an acoustic is that of your room, and is real and taking on 100% of the duties of supporting the good sound that you are getting, that very realistic playback of someone's performance. That is one end of the continuum. The other end is a very wet recording with whatever recording mikes you want to use placed back in the audience where the good seats are in a mistaken attempt to record 100% the space of the live sound, and then you will try to play it back in a nearly anechoic very dry room, as some poor dumb *******s attempt all the time. Inability to converse without gratuitous ad hominem noted. Do you really wonder why you're not successful in furthering your arguments? We all know (heh) that that doesn't work. You will say that it is because that spatial info will be lost because we don't have enough microphones and speakers to make all of those patterns come back. No I won't. I'll repeat the focus of this entire discussion - the spatial information is LOST. Period. Even with myriad microphones and playback speakers, you would have to be able control the directivity precisely, recording and playback, on a very small scale, which isn't possible. You can devise schemes along those lines that can get you closer, for sure, but the electrical signal from each and every one of those mikes would be bereft of any directional information, being only ameliorated by simulating that information through placement and orientation of the playback devices. OK, fine, great dodge, but in reality WE DON'T DO IT THAT WAY. No, typically we record a great deal of the reverberant field as well. You feel adding more reflections on playback enhances realism. Most do not. We place the mikes much closer to the instruments than you would sit so listen in that hall, then we play it back in another room at a distance from you so that the two spaces complement each other and lead to greater realism even tho we cannot have "accuracy" of what was heard if you were there. Accuracy would mean you would hear a concert from 9 ft above the conductor's head, or from ears that are 18 ft apart, or some such reductio ad absurdum. A nonsensical argument. Accuracy would mean you'd hear, accurately, what an average listener heard at a predetermined point in the audience. The placement of the microphones are irrelevant - they do not define the point of accuracy, they are positioned as needed to provide the greatest accuracy, relative to the actual performance, at the predetermined location. Only in the thought experiment Dick provided, that I replied to, would the mike location and the "accuracy point" coincide, and then only if the method *worked*, the fallacy of which was the point of the example. I have described the idea as close-miking the soundstage, because we are recording not just the actual instruments but also the early reflected sound from the sidewalls, the most important reflections, and also a hint of the reverberant field. All of this info mixes with the playback acoustic to give you the realism of sitting there with them in front of you, and also the "flavor" of the live acoustic space. OK, so it wasn't very brief, but there is a lot more to it than even that. Bottom line, AE is correct that it is possible to record for greater realism, but the techniques and reasons may be surprising to both of you, and progress toward a goal of greater realism may take a path a little - well, a lot - different from what most of us assume. That path is NOT a search for greater and greater "accuracy," but rather trying to work with and understand what we are actually doing with a field-type system, which is making sound in rooms, not making sound in your ears. OK, I can summarize our arguments much more succinctly: You want playback that sounds "real" to you, irrespective of whether it resembles the actual recorded event or not. You don't think directional information is lost in recording, because you create your own, in your room, in a manner that pleases you, and sounds real - to you - and then say "see, it's there!", which it demonstrably is not. Clearly, once the playback is untethered from the actual event, then "realism" is strictly a matter of your preference. There is no reference, let alone definable quantitative or qualitative evaluation parameters. I, on the other hand, want to hear, to the extent possible, a playback that sounds more like the event, knowing that such accuracy is not truly attainable, but choosing equipment and setup parameters that get me as close as possible. I'm stuck with the same evaluation parameters challenge as you, but I at least have a reference that exists outside of my own very individual head. Keith |
#32
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
AE -
I don't doubt your knowledge and research on multitrack optical recorders, I am just pointing out that you do not record original sound for music, effects, and dialog or anything else simultaneously on one machine. Except maybe in post production, as you say, for ease in doing the mix. I don't know what all machines they had for doing the mix and don't really care that much - it changes too fast. I sat in on a mix one fine time at a big production house. The dubbing stage was like a little movie theater, only with an engineer at a huge board rather than an audience. Behind him there was a glass wall with a room behind it that had all of the sync film players. Each reel held 10 minutes, I think, and that is about all a mixer could concentrate anyway at one time. Controls may have been programmable as well, for when to come up and down. I did much the same thing in my amateur filmmaking, using a couple of fullcoat recorders and the sync cassette player running together. Except I had to do the whole film at one sitting, because I had no way of inserting cleanly after a take. The biggest change that came along for me was crystal sync, in which all of the recorders and cameras ran at a crystal controlled constant speed, and so didn't need a sync cord running between them. Today, with video and digital recording, all recorders automatically run at the same precise speed, so no worries about sync ever. I can shoot an entire wedding for over an hour and sync up the sound with the picture later at home and it stays in sync the whole time. And when you think about the digital projection revolution and what it means for distribution and costs and storage and back-breaking labor - whew! They used to deliver a movie - to every theater for every title - several heavy cans of 35 or 70 mm film that had to then be toted up to the projection room and set to go at showtime. I think they sometimes spliced it together for projection, but the main routine was to wait for the marks to come up and hit play on the next machine. Today they deliver the whole movie on a hard drive. No changing reels and no focus problems and no threading a projector. And footnote added - stereo was invented and well understood long before 1947. It wasn't done by the film people this time, but the film people are usually at the forefront of new technologies well before pure audio people. Example, stereo and surround sound. Got any more jazz recordings? I just finished my Concert Band recording yesterday. I got the singer's voice off the board, so I could use a clean feed. It still has all of the echo from the hall sound, but my dry track of the announcers and singers really adds a touch of clarity. Gary |
#33
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
KH wrote:
OK, I can summarize our arguments much more succinctly: You want playback that sounds "real" to you, irrespective of whether it resembles the actual recorded event or not. You don't think directional information is lost in recording, because you create your own, in your room, in a manner that pleases you, and sounds real - to you - and then say "see, it's there!", which it demonstrably is not. Clearly, once the playback is untethered from the actual event, then "realism" is strictly a matter of your preference. There is no reference, let alone definable quantitative or qualitative evaluation parameters. I, on the other hand, want to hear, to the extent possible, a playback that sounds more like the event, knowing that such accuracy is not truly attainable, but choosing equipment and setup parameters that get me as close as possible. I'm stuck with the same evaluation parameters challenge as you, but I at least have a reference that exists outside of my own very individual head. Keith Thanks for that, and may I respond with my argument against? This will be both theoretical and subjective results oriented, which is the best I can do. THEORETICAL: Assuming that you are saying that you want to record a sort of sound "picture" of a live performance, as if the playback will then cast that picture back to your ears and let you hear "into" another acoustic space - how am I doing - you will then play it back on these really accurate speakers which have been placed at an angle that will complement the recorded angles, in a space that is deadened down to some practical extent so that it doesn't dilute the recorded acoustic too much with its own sound. OK? So my objection on theoretical grounds is that this playback model with two point sources (or three, doesn't matter) will have a sound of its own despite your attempts to eliminate your room from the nuisance variables, and will therefore CHANGE the spatial characteristics of the original model - think of direct to reflected ratios, early reflected sound coming form different points in space than the direct sound, and the full reverberant field - to those of your playback system - two highly directional points in space surrounded by a void. Your problem on the theoretical level is how - by what theorem or scheme or mechanism - do you expect to get those spatial patterns of the original back again? If you say you want the sound from each side to enter your ears and fool you into hearing the much larger space, then you are confusing this system with binaural, which IS an ear input system. Complaining that the stereophonic system has "lost" those spatial characteristics is either a dodge or an admission that your theory, or paradigm does not work. Using the crosstalk cancellation idea is yet another confusion with binaural. It gets the sound out of the speakers as obvious sources, but it also spreads the sound artificially in a horseshoe pattern around you, where it doesn't belong. SUBJECTIVE OBSERVATION: I have observed over many years of evaluating imaging and speaker directivity that if we do it your way, with your theory of stereo, what results is that ALL of the recorded direct, early reflected, and reverberant sound are heard to come from those two points in space that are your speakers, plus of course the instruments in between the speakers, and any sounds that are extreme left or right on the stage are heard as coming right from those speakers, causing you to hear them as real sources, artificial sources in your room. If you would agree that a desirable goal would be to have the speakers disappear from the soundscape, then you are purposely running counter to that goal, for mistaken theoretical reasons, or mistaken ideas on how the system works. In my Image Model Theory with the more omnidirectional speakers in a large, good sounding room, with center and surround speakers to support the reverberant field reconstruction, what I get is the speakers completely disappearing - NO sound is heard as coming directly from any speaker - and a set of aerial images of instruments coming from a region behind and beside the actual speakers, magically placing them at acoustic points in my room and sounding very much as if they are right there in front of me. I also get the early reflected sound that was recorded sounding from the front and side walls, just like live, giving the impression that there is a decoding effect going on in which my model is placing all of the elements of the recorded sounds and acoustics coming once again from appropriate locations spatially within my room. This happens for reasons that are the same as why they happen live, because delayed extreme right or left sounds are actually bounced from similar angles on playback to those that were recorded. Obviously, this is talking about the frontal soundstage only, and not the full reverberant field, which condition is the same for both of us. We can both support all that with surround sound speakers and either discrete recording or processing. But the idea is the same - the object on playback is the reconstruction of a realistic model of the live fields, not casting the recorded sound back toward your ears! My playback will sound like a model of the live sound fields, the size and shape of my own room superimposed upon the recorded acoustic, the resultant sound being like a 30-70 mix of the live vs the local, leaning toward the live - as if you were sticking your head into a three dimensional model, or diurama of the concert hall. If it is a smaller group like a jazz trio, it will sound like they are right there in the room with you. Your playback (in my experience with similar) will sound like a rectangular hole cut in a wall separating you from the concert hall, but placed maybe halfway up, kind of like a widescreen movie that is limited right and left and is not 3D. You will be hearing more of the recorded acoustic space than I, but coming from this distorted set of incident angles represented by the separation of the speakers. Theoretical bull**** aside now, to get back to AE's OP, we can "retrieve" a very realistic sound on playback if the recording contains some decent proportions of the important images that we need for reconstruction, and we can reconstruct those images if we study the problem as a field-type system rather than an ear input system. It is not a system of "accuracy" because the recording does not contain this imagined "picture" of another acoustic space as if from a position that you would want to listen from in the concert hall. So if we continue to try to retrieve this non-existent accuracy by using directional speakers in a dead room the sound will all collapse to the speaker boxes in front of us, destroying the suspension of disbelief, and we will never reach the kind of realism that I am enjoying every day with my mistaken ideas of how stereo works. Ya pays your money and takes your choice. Gary Eickmeier |
#34
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
On 4/10/2013 8:04 AM, Gary Eickmeier wrote:
KH wrote: OK, I can summarize our arguments much more succinctly: You want playback that sounds "real" to you, irrespective of whether it resembles the actual recorded event or not. You don't think directional information is lost in recording, because you create your own, in your room, in a manner that pleases you, and sounds real - to you - and then say "see, it's there!", which it demonstrably is not. Clearly, once the playback is untethered from the actual event, then "realism" is strictly a matter of your preference. There is no reference, let alone definable quantitative or qualitative evaluation parameters. I, on the other hand, want to hear, to the extent possible, a playback that sounds more like the event, knowing that such accuracy is not truly attainable, but choosing equipment and setup parameters that get me as close as possible. I'm stuck with the same evaluation parameters challenge as you, but I at least have a reference that exists outside of my own very individual head. Keith Thanks for that, and may I respond with my argument against? This will be both theoretical and subjective results oriented, which is the best I can do. THEORETICAL: Assuming that you are saying that you want to record a sort of sound "picture" of a live performance, as if the playback will then cast that picture back to your ears and let you hear "into" another acoustic space - how am I doing - you will then play it back on these really accurate speakers which have been placed at an angle that will complement the recorded angles, in a space that is deadened down to some practical extent so that it doesn't dilute the recorded acoustic too much with its own sound. OK? Somewhat the general gist, dismissive phrasing aside. So my objection on theoretical grounds is that this playback model with two point sources (or three, doesn't matter) will have a sound of its own despite your attempts to eliminate your room from the nuisance variables, and will therefore CHANGE the spatial characteristics of the original model - Yes, and... think of direct to reflected ratios, early reflected sound coming form different points in space than the direct sound, and the full reverberant field - to those of your playback system - two highly directional points in space surrounded by a void. Well, no. They are not "highly" directional, nor are they surrounded by "a void". They have a great deal of information between them, in addition to the unavoidable reverberant field around them. So... Your problem on the theoretical level is how - by what theorem or scheme or mechanism - do you expect to get those spatial patterns of the original back again? Uhmm, you don't. Have you not paid any attention to what I'm saying? That directional information is GONE. By taking speakers that have a *specific*, not as you would claim, "highly", directional radiation pattern, and placing them in locations, at angles, to provide an illusion of the acoustic in the live event, one can achieve a pretty fair recreation. If you say you want the sound from each side to enter your ears and fool you into hearing the much larger space, then you are confusing this system with binaural, which IS an ear input system. Complaining that the stereophonic system has "lost" those spatial characteristics is either a dodge or an admission that your theory, or paradigm does not work. Using the crosstalk cancellation idea is yet another confusion with binaural. It gets the sound out of the speakers as obvious sources, but it also spreads the sound artificially in a horseshoe pattern around you, where it doesn't belong. Good grief! When will you get off the "binaural" dodge? No one here is conflating stereo and binaural except you. *NO* method, theory, or floobie dust will "work" to bring back information that is NOT in the recording. SUBJECTIVE OBSERVATION: I have observed over many years of evaluating imaging and speaker directivity that if we do it your way, with your theory of stereo, what results is that ALL of the recorded direct, early reflected, and reverberant sound are heard to come from those two points in space that are your speakers, plus of course the instruments in between the speakers, and any sounds that are extreme left or right on the stage are heard as coming right from those speakers, causing you to hear them as real sources, artificial sources in your room. If you would agree that a desirable goal would be to have the speakers disappear from the soundscape, then you are purposely running counter to that goal, for mistaken theoretical reasons, or mistaken ideas on how the system works. No, you are mistakenly assuming that your experience, and your interpretation of that experience is universal. It is not. You think no one in the history of audio, save for you, has experienced a setup where the speakers - on a good recording - disappear in space? Allow me to disabuse you of that misconception. In my Image Model Theory with the more omnidirectional speakers in a large, good sounding room, with center and surround speakers to support the reverberant field reconstruction, No - and this is where you are simply, and demonstrably WRONG. You are *constructing* a reverberant field, you are NOT REconstructing any 3-d field from a 2-d recording. Please explain the physics that would allow that. Any ideas at all? Tell me how the 3-d spatial information is encoded into a 2-d signal? Yes, with 2 or more channels, you can simulate 3-d to an extent, but it isn't accurate. Doesn't mean it isn't good, or realistic - it most definitely can be both. what I get is the speakers completely disappearing - NO sound is heard as coming directly from any speaker - and a set of aerial images of instruments coming from a region behind and beside the actual speakers, magically placing them at acoustic points in my room and sounding very much as if they are right there in front of me. I also get the early reflected sound that was recorded sounding from the front and side walls, just like live, giving the impression that there is a decoding effect going on in which my model is placing all of the elements of the recorded sounds and acoustics coming once again from appropriate locations spatially within my room. This happens for reasons that are the same as why they happen live, because delayed extreme right or left sounds are actually bounced from similar angles on playback to those that were recorded. That is simply impossible. The reverberant field in the recording has no directional information, and although you can bounce, or reflect, signals from all over the room, you are, for example, bouncing input that originally came from rear left in all directions - not from "the appropriate directions". All directions. Obviously, this is talking about the frontal soundstage only, and not the full reverberant field, which condition is the same for both of us. We can both support all that with surround sound speakers and either discrete recording or processing. But the idea is the same - the object on playback is the reconstruction of a realistic model of the live fields, not casting the recorded sound back toward your ears! In your opinion. My playback will sound like a model of the live sound fields, the size and shape of my own room superimposed upon the recorded acoustic, the resultant sound being like a 30-70 mix of the live vs the local, leaning toward the live - as if you were sticking your head into a three dimensional model, or diurama of the concert hall. If it is a smaller group like a jazz trio, it will sound like they are right there in the room with you. Your playback (in my experience with similar) will sound like a rectangular hole cut in a wall separating you from the concert hall, but placed maybe halfway up, kind of like a widescreen movie that is limited right and left and is not 3D. You will be hearing more of the recorded acoustic space than I, but coming from this distorted set of incident angles represented by the separation of the speakers. You are simply wrong. That is not what I hear, as you've been told many times. You assume, based on your tastes, and your guess as to what my system sounds like, that you know what *I* hear. Theoretical bull**** aside now, to get back to AE's OP, we can "retrieve" a very realistic sound on playback if the recording contains some decent proportions of the important images that we need for reconstruction, and we can reconstruct those images if we study the problem as a field-type system rather than an ear input system. Yes, although your method for creation is certainly not the only, nor, IMO, the better of the approaches that can be taken. And it is simulation, not reconstruction. It is not a system of "accuracy" Amen because the recording does not contain this imagined "picture" of another acoustic space as if from a position that you would want to listen from in the concert hall. So if we continue to try to retrieve this non-existent accuracy by using directional speakers in a dead room the sound will all collapse to the speaker boxes in front of us, No it doesn't. You either don't like the sound that most find realistic, or you haven't heard it. You likely are so used to the overblown fake "acoustic" you create by myriad reflections that you simply think anything else is just "dead". AE's original "sound splashed all over" is a very apt description of what many, if not most of us *HEAR* in the type of system you find to be the hallmark of realism. My soundfield sounds pretty good to me, and it certainly does not "collapse" to the speaker boxes. I'd be pretty unhappy if that were the case, and I'm actually quite satisfied. destroying the suspension of disbelief, and we will never reach the kind of realism that I am enjoying every day with my mistaken ideas of how stereo works. If you're enjoying it, more power to you. There may be many more just like you that would enjoy your system. But continuing to derogate all who disagree, or have other tastes, to the status of "poor dumb *******s", as you just did, will ensure the continuing irrelevance of your "theory", and ensure it's place in the dustbin of history. Keith |
#35
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
KH wrote:
On 4/10/2013 8:04 AM, Gary Eickmeier wrote: THEORETICAL: Assuming that you are saying that you want to record a sort of sound "picture" of a live performance, as if the playback will then cast that picture back to your ears and let you hear "into" another acoustic space - how am I doing - you will then play it back on these really accurate speakers which have been placed at an angle that will complement the recorded angles, in a space that is deadened down to some practical extent so that it doesn't dilute the recorded acoustic too much with its own sound. OK? Somewhat the general gist, dismissive phrasing aside. You've got me curious - what dismissive phrasing? Uhmm, you don't. Have you not paid any attention to what I'm saying? That directional information is GONE. By taking speakers that have a *specific*, not as you would claim, "highly", directional radiation pattern, and placing them in locations, at angles, to provide an illusion of the acoustic in the live event, one can achieve a pretty fair recreation. Curiouser and curiouser - I think I know what you are getting at Keith, but saying that directional information is lost in a stereo recording hits me the wrong way. It would be idiotic to state that a single microphone gives no directional information, so I'm sure that is not what you mean. Perhaps it has something to do with the well-known phenom that different miking techniques give different apparent perspectives on playback, and also that different speaker setups do somewhat the same thing. But all of that is why I say that we must reconstruct the stereo image on playback. It has to be contained in the recording and it has to be reconstructed on playback, by placing speakers in rooms and sending channels to them and experiencing the total result of speaker and room interface. No - and this is where you are simply, and demonstrably WRONG. You are *constructing* a reverberant field, you are NOT REconstructing any 3-d field from a 2-d recording. Please explain the physics that would allow that. Any ideas at all? Tell me how the 3-d spatial information is encoded into a 2-d signal? Yes, with 2 or more channels, you can simulate 3-d to an extent, but it isn't accurate. Doesn't mean it isn't good, or realistic - it most definitely can be both. Let's talk about that for a moment. Just indulge me. As you have so bravely - and courteously - up to this point. The three dimensions of which you speak are height, width, and depth. The width dimension I think we all agree is encoded in the stereo information from the summing localization from the multiple microphones used. We can perceive left, center, right, and anything in between - but only if we reconstruct that on playback by placing the speakers left and right with some separation. That used to go without saying, but now is worth pointing out. If your wife tries to make you stick one speaker behind the sofa and the other on the bookshelf above, you point out that you must reconstruct the width information contained in the recording by placing the speakers in a certain way, or it won't work. The height we aren't real concerned with, because the sources are all at the same level in front of us. There was a little interest in full periphony with Ambisonics, but didn't really take off. So we place the speakers situated at about the same level as the instruments would be, and press on - except for an intresting psychoacoustic effect wherein certain frequencies can seem to be higher up than others, like the horns seem to come from a little higher than the drums or others. We can also be fooled into hearing level sound from speakers that are hung from the ceiling. I guess this is mainly because we have no terrific height perception except from moving our heads, and if the installation is not done so poorly that the reflecting surfaces nearby call attention to the localization of the speakers themselves, our brain is satisfied that the orchestra must be level with us, where it belongs. Still, we usually just place the speakers there for obvious reasons. Now that bugaboo depth. We have no mechanism for detecting depth, either, except by moving our heads a little or moving around a little in a closed environment - sound in rooms gives more localization information than outdoors or anechoically. In experiments, if two sounds are played at different depths, but the farther one is compensated with increased gain, you can't tell the depth. Within a room, however, we can hear to a great extent a source's position w respect to the walls around it by the reflection pattern. Moving around even a little helps with that, but basically we can sense if something is right up against the wall or spaced out from it. Audiophile experience over a long period has led us to placing the speakers well out from the walls to give that impression of depth due to the simple observations above - that you have placed the reconstructed sound sources in your three dimensional space so that they have some degree of depth to the sound! Then if the recording contains some near and some far instruments, a learned response tells us the difference because "far" sounds different than "near" from loudness and reverberance etc etc. Oh, there he goes again, fakey fakey fakey! He wants to build a little model of the live soundscape by placing speakers like a pop-up book. Sorry, but ya, that's pretty much the way it is. That is simply impossible. The reverberant field in the recording has no directional information, and although you can bounce, or reflect, signals from all over the room, you are, for example, bouncing input that originally came from rear left in all directions - not from "the appropriate directions". All directions. NO - if the recording properly contains sound that was bouncing off the left side wall of the concert hall from those instruments on the left side of the orchestra, and if those sounds are played on a left channel speaker that has some output that bounces off the left wall of your room, you get that sound coming from the appropreate direction - from points in space that are different from the primary sound, whch is coming from the speaker first. Precedence effect, but that is getting too involved for now. Right now just imagine a light bulb on the left side of the room. It shines (bounces) more of its output from the left side wall than any others. NOT everywhere. Footnote, there is a lot more to speaker positioning than I can relate in this short essay. I understand your statement that info from the rear left (like, 120°) cannot be made to come from the rear left on playback, but that is not real important for reconstructing the frontal soundstage, and we can easily turn to surround sound if you think it is. Yes, although your method for creation is certainly not the only, nor, IMO, the better of the approaches that can be taken. And it is simulation, not reconstruction. . You either don't like the sound that most find realistic, or you haven't heard it. You likely are so used to the overblown fake "acoustic" you create by myriad reflections that you simply think anything else is just "dead". AE's original "sound splashed all over" is a very apt description of what many, if not most of us *HEAR* in the type of system you find to be the hallmark of realism. My soundfield sounds pretty good to me, and it certainly does not "collapse" to the speaker boxes. I'd be pretty unhappy if that were the case, and I'm actually quite satisfied. If you're enjoying it, more power to you. There may be many more just like you that would enjoy your system. But continuing to derogate all who disagree, or have other tastes, to the status of "poor dumb *******s", as you just did, will ensure the continuing irrelevance of your "theory", and ensure it's place in the dustbin of history. Keith Fair enough. Results and perception are what is important. We keep on truckin and make changes here and there as we go and try to figure out what causes either the improvements or unimprovements. One thing is certain: the Big Three, as I call them, of radiation pattern, speaker positioning, and room acoustics that Siegfried Linkwitz asked about in the Challenge to the AES, are difinitely audible, and are the main variables in the making of the sound that we both - all - percieve in our playback. It is those variables that must be studied to find out which ideas sound better than others in the playback of the field-type system called stereophonic sound. Gary Eickmeier |
#36
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
Gary Eickmeier wrote:
Audio_Empire wrote: That's not possible. The microphone is not designed to reproduce anything. it would make a more than lousy speaker. Also recoding with a single mike will result in NO spatial information being captured (it's called "monaural sound"). One needs two mikes and the spatial information results from the difference between the two mike signals and THAT takes place in the listeners' ears. We hear in stereo due differences in phase, time delay, and spatial separation of signals reaching our ears. if done right, those cues can provide a very satisfactory soundstage on a good stereo system. OK, let's take another run at this. I'm sure Mr. Pierce didn't mean to imply that microphones can be reproducers; he was making a philosophical point. No, I implied no such thing, Mr. Eickmeier, I stated it explicitly. Go do some research on the reciprocity principle and then come back and try to argue your point. -- +--------------------------------+ + Dick Pierce | + Professional Audio Development | +--------------------------------+ |
#37
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
Audio_Empire wrote:
On Thursday, April 4, 2013 4:27:06 PM UTC-7, Dick Pierce wrote: Now, take a stereo pair. The situation is really not any better It is geometrically impossible to disambiguate, for example, by any property in the elctrical signals, whether a source of a sound is anywhere on a circle whose center is defined by the line between the two microphones and whose plane is at right angles to that circle. Two omnis some distance apart will generate the SAME electrical signals whether the source is 20 feet ahead, 20 feet above, 20 feet behind or anywhere else on the circle. The same is true of any other mike position. The only position that can be unambiguously recorded is somewhere EXACTLY in between the two, which is arguably not very useful. Are you talking about omnidirectional microphones here? Because they don't work as a stereo pair unless you take extraordinary precautions, such as placing a big sound baffle between them as Ray Kimber does for his IsoMike recordings. Take ANY microphone you choose, ANY patter,n next to ANY other contrivance you want: the output of the microphone is an electrical signal which is simply the instantaneous magnitude at the diaphragm surface. NO directionality information CAn be encoded: it's a single dimension vs time. Consider also the reciprocity principle as a gedanken (and, as a real-world excercise, if your want). Record something from a complex sound field with a microphone of your choosing. Now, play it back through the same microphone. While you're thinking about it, go study up on the reciprocity principle. If you did do that, say, through a magnetic microphone, it wouldn't sound very good I'm afraid. It would likely sound much worse, even, than a telephone. And I don't see what this has to do with the subject at hand. Microphones are designed, to capture sound and turn it into an electronic analog of that sound, it is not designed to be a reproducer. Please, go read up on the reciprocity principle. It's a very well established acoustical principle with solid, hard science behind it. Try Balnkenstock, try Breanek, try Kinsler and Frey. If you succeed in convincing THEM they're wrong, then you get to some back here and continue your argument. Until then all due respect, you're arguing from a poit of technical ignorance. Now, if your assertion were correct, recording with a single mike of several sound sources in different direction, should result, if you insist there is no loss of information, in the sound that emenates from that same microphone finding their way back to the original location they were emitted from. That's not possible. The microphone is not designed to reproduce anything. it would make a more than lousy speaker. B*ll****! Within linear limits of the device, the reciprocity principle states quite clearly that an acoustical transducer will make just as good a speaker as it does a microphone. Again, go research the topic, because your argument simply is not based on solid, well-established physics. This same principle is the basis behind high-accuracy independent calibration of microphones. IN one example, you take two microphones: drive one as a speaker, and measure with the other. The resulting response of both measured together will have twice the aberrations of one. It's the basis behind precision independent microphone calibration systems by the likes of IET, Bruel & Kjaer, General Radio and others. Let's take a very simple experiment, practically realizable. Take a directional microphone of your choosing. Let's just say it's a cardiod (pick whatever you like). This is an experiment you can do yourself. Now, place a speaker 10' on its principle axis playing a 1 kHz tone sufficient to produce 80 dB at the position of the microphone. At the same time, place a speaker 120 degrees off axis and play a 2 kHz tone through it, also producing an SPL of 80 dB at the mike position. Record the signal. DO the same, only in the on-axis spaekr, play 2 kHz at 72 dB SPL and in the off-axis speaker, play 1 kHz at 86 dB. Record this signal. Now, use ANY means you want (other than your memory of the experiment), tell me unambiguously, which recording is which. If you cannot tell, then this tells me that all information about the sound field at THAT position was lost by the microphone. Okay, do the same thing with TWO microphones of your choosing. Without getting into the details, there are an infinite number of arrangements of the sources which CANNOT be unambiguously determined by the resulting pair of electrical signals. Also recoding with a single mike will result in NO spatial information being captured (it's called "monaural sound"). One needs two mikes and the spatial information results from the difference between the two mike signals and THAT takes place in the listeners' ears. But two mikes are utterly incapable of caturing unambiguously that information. Take a sound source that is placed such that it's 60 degrees off of both microphone. Okay, WHICH 60 degrees? There's an infinite number of position where the sound source geometrically satisfies that position. Which one is is? What in the electrical signal disambiguates which on it is. We hear in stereo due differences in phase, time delay, and spatial separation of signals reaching our ears. if done right, those cues can provide a very satisfactory soundstage on a good stereo system. No, the ears do not depend on that. A VERY IMPORTANT factor you have left out is the HRTF, the Head-Related Transfer Function: which is an aberration in the sound field created by our heads and our outer ear structures (and the rest of our body, for that matter) which is crucial in the disambiguation of the problem I cited above. ANY stereo pair of microphone cannot tell the difference between a sound source directly in front, directly overhead, directly behind, or directly below. Blindfolded, I'd bet you could tell such pretty quickly and you'd get it right the vast majority of the time (and, yet, I can still find some cases where you might get fooled). If your ancestors could NOT do this, well, you'd not be around to having this discussion. The shading effects along with the co,plex path differences and phase scrambling and all that is what give you the ability to further encode the phase and amplitude differences and turn them into real directional information. Extend the experiment one step further. DO the same thing blindfolded and with one ear plugged. According to your thesis, you'd lose ALL ability to sense direction, yet there is well over a century of hearing research that simply contradicts you. The ability and success to decode direction is less certain, to to be sure, but it does not vanish like you would assert. And it's all bnecause of YOUR HRTF, which YOU, inadvertantly, started training yourself to use from the moment you were born. Now, to your assertions about the reciprocity principle: by YOUR argument, if I then take the SAME signal recorded from the two microphone, and two microphones of YOUR choosing, feed them back through those two microphones (maintaining them in linear operation), then I should be able to recreate the same original sound field the captured. I should be able to trace back to the original sources the sound from which they originally eminated. That would mean that I should be able to go back and sample the resulting soundfield and find that the 2 kHz tone is being projected in the direction fron which it originally came, and the 1 kHz tone should be projected in the unique direction it came from. Will this happen? THAT'S what it means to "creat the original sound field." If you want to argue whether that's the point of music reproduction or the degree aof accuracy to which it is sufficient to satisfy the listener, or whether it's all aprlor trick, that's fine: that's one discussion. But if you want to make technical claims about the suitability of an acoustic transducer for such an experiment under the reciprocity principle, thus essentially denying the validity of the reciprocity principle, I might caution you that you're skating in VERY thin ice, technically, and you might not want to go there. -- +--------------------------------+ + Dick Pierce | + Professional Audio Development | +--------------------------------+ |
#38
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
On Friday, April 12, 2013 9:42:08 AM UTC-7, Dick Pierce wrote:
Audio_Empire wrote: On Thursday, April 4, 2013 4:27:06 PM UTC-7, Dick Pierce wrote: Now, take a stereo pair. The situation is really not any better It is geometrically impossible to disambiguate, for example, by any property in the elctrical signals, whether a source of a sound is anywhere on a circle whose center is defined by the line between the two microphones and whose plane is at right angles to that circle. Two omnis some distance apart will generate the SAME electrical signals whether the source is 20 feet ahead, 20 feet above, 20 feet behind or anywhere else on the circle. The same is true of any other mike position. The only position that can be unambiguously recorded is somewhere EXACTLY in between the two, which is arguably not very useful. Are you talking about omnidirectional microphones here? Because they don't work as a stereo pair unless you take extraordinary precautions, such as placing a big sound baffle between them as Ray Kimber does for his IsoMike recordings. Take ANY microphone you choose, ANY patter,n next to ANY other contrivance you want: the output of the microphone is an electrical signal which is simply the instantaneous magnitude at the diaphragm surface. NO directionality information CAn be encoded: it's a single dimension vs time. Well, of course it can't. That can be easily demonstrated by playing only one channel of a true stereo recording through both stereo speakers. If you are familiar enough with the recording you will notice that nothing is missing. The entire ensemble is there. Some instruments might be slightly attenuated (and depending upon how the recording was made, they might not), but they are present. There is no directional information without two microphones, each "viewing" the performance from a different perspective. Consider also the reciprocity principle as a gedanken (and, as a real-world excercise, if your want). Record something from a complex sound field with a microphone of your choosing. Now, play it back through the same microphone. While you're thinking about it, go study up on the reciprocity principle. If you did do that, say, through a magnetic microphone, it wouldn't sound very good I'm afraid. It would likely sound much worse, even, than a telephone. And I don't see what this has to do with the subject at hand. Microphones are designed, to capture sound and turn it into an electronic analog of that sound, it is not designed to be a reproducer. Please, go read up on the reciprocity principle. It's a very well established acoustical principle with solid, hard science behind it. Try Balnkenstock, try Breanek, try Kinsler and Frey. If you succeed in convincing THEM they're wrong, then you get to some back here and continue your argument. Until then all due respect, you're arguing from a poit of technical ignorance. I understand the reciprocity principle perfectly well, but the way you describe it sounded like you were talking literally. On the other hand It's moot point because I never said that one microphone picked up any actual directional information. In fact, in one post I stated that one microphone (I used and omni for clarity, but any mike will do) cannot pick up any spatial information. That's why it's called monaural sound. Now, if your assertion were correct, recording with a single mike of several sound sources in different direction, should result, if you insist there is no loss of information, in the sound that emenates from that same microphone finding their way back to the original location they were emitted from. I never made such an assertion, If you think I did, then you either misunderstood my point, or I didn't state it very well, which is possible. That's not possible. The microphone is not designed to reproduce anything. it would make a more than lousy speaker. B*ll****! Within linear limits of the device, the reciprocity principle states quite clearly that an acoustical transducer will make just as good a speaker as it does a microphone. No, because within the limits of physics, a microphone diaphragm cannot move enough air to make even a poor speaker and the reciprocity principle never says that it should. Most studio microphones can accept SPLs approaching 140dB before distorting, but under no circumstances could such a microphone reproduce anywhere near 140dB, except, that the diaphragm can MOVE (I.E. be displaced) as much in response to a driving signal as it would when intercepting a sound wave of 140 dB in intensity. Whether or not it produces any actual sound in the room in which it is energized is another story. The microphone diaphragm should move as much in response to the reciprocal of an electrical signal it generated as it did when converting the original acoustical signal to that electrical signal. Ideally, they would be identical. Again, go research the topic, because your argument simply is not based on solid, well-established physics. This same principle is the basis behind high-accuracy independent calibration of microphones. IN one example, you take two microphones: drive one as a speaker, and measure with the other. The resulting response of both measured together will have twice the aberrations of one. It's the basis behind precision independent microphone calibration systems by the likes of IET, Bruel & Kjaer, General Radio and others. I don't know what argument you are lumbering me with, but I think you are confusing me with Mr Eickmeier or perhaps someone else. Let's take a very simple experiment, practically realizable. Take a directional microphone of your choosing. Let's just say it's a cardiod (pick whatever you like). This is an experiment you can do yourself. Now, place a speaker 10' on its principle axis playing a 1 kHz tone sufficient to produce 80 dB at the position of the microphone. At the same time, place a speaker 120 degrees off axis and play a 2 kHz tone through it, also producing an SPL of 80 dB at the mike position. Record the signal. DO the same, only in the on-axis spaekr, play 2 kHz at 72 dB SPL and in the off-axis speaker, play 1 kHz at 86 dB. Record this signal. Now, use ANY means you want (other than your memory of the experiment), tell me unambiguously, which recording is which. If you cannot tell, then this tells me that all information about the sound field at THAT position was lost by the microphone. I agree and have never asserted anything else. The only thing that one MIGHT be able to tell is when the speaker gas been placed in the microphone's pattern shadow, and even that that would be largely frequency and distance dependent. Okay, do the same thing with TWO microphones of your choosing. Without getting into the details, there are an infinite number of arrangements of the sources which CANNOT be unambiguously determined by the resulting pair of electrical signals. Of course there are. Also recoding with a single mike will result in NO spatial information being captured (it's called "monaural sound"). One needs two mikes and the spatial information results from the difference between the two mike signals and THAT takes place in the listeners' ears. But two mikes are utterly incapable of caturing unambiguously that information. Take a sound source that is placed such that it's 60 degrees off of both microphone. Okay, WHICH 60 degrees? There's an infinite number of position where the sound source geometrically satisfies that position. Which one is is? What in the electrical signal disambiguates which on it is. Nothing is perfect, that's a given. But "listen" to yourself. You took a simple statement by me where I said that one microphone captures no spatial information, and that you need two mikes for real stereo and all that does is pick up any differences in phase or intensity that result from those two different perspectives. And you argue that I'm wrong simply because the process is far from perfect. I think we all know that, Dick. We hear in stereo due differences in phase, time delay, and spatial separation of signals reaching our ears. if done right, those cues can provide a very satisfactory soundstage on a good stereo system. No, the ears do not depend on that. A VERY IMPORTANT factor you have left out is the HRTF, the Head-Related Transfer Function: which is an aberration in the sound field created by our heads and our outer ear structures (and the rest of our body, for that matter) which is crucial in the disambiguation of the problem I cited above. I left out nothing. The Head Transfer Function is implied by my statement that the cues we use to pick up directionality occur in the air, as those cues reach our ears. My statement, though simplified from your lofty, technical standpoint, is correct in as far as it goes. ANY stereo pair of microphone cannot tell the difference between a sound source directly in front, directly overhead, directly behind, or directly below. Blindfolded, I'd bet you could tell such pretty quickly and you'd get it right the vast majority of the time (and, yet, I can still find some cases where you might get fooled). If your ancestors could NOT do this, well, you'd not be around to having this discussion. The shading effects along with the co,plex path differences and phase scrambling and all that is what give you the ability to further encode the phase and amplitude differences and turn them into real directional information. I have thirty years of location recording experience. I know exactly what microphones are capable of doing. I've probably miked more symphony concerts and jazz ensembles than most people have ever even heard. Extend the experiment one step further. DO the same thing blindfolded and with one ear plugged. According to your thesis, you'd lose ALL ability to sense direction, yet there is well over a century of hearing research that simply contradicts you. The ability and success to decode direction is less certain, to to be sure, but it does not vanish like you would assert. And it's all bnecause of YOUR HRTF, which YOU, inadvertantly, started training yourself to use from the moment you were born. Again. I don't recall ever saying anything that should cause you or anyone else to attribute the above attitude to me. There must be other windmills for you to joust at, Don Quixote. |
#39
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
Okay, so Mr. Pierce, Keith Howard, and Scott W are surprised that a single
microphone can't record directional information, Pierce thinks that recordings need to carry HRTF, and microphones can make great speakers. I think I am done here. Gary Eickmeier |
#40
Posted to rec.audio.high-end
|
|||
|
|||
Why do most commercial recordings (talking Classical and Jazz,
Gary Eickmeier wrote:
Okay, so Mr. Pierce, Keith Howard, and Scott W are surprised that a single microphone can't record directional information, Pierce thinks that recordings need to carry HRTF, and microphones can make great speakers. I never said any such thing, Mr. Eickmeier. I think I am done here. You are done if you insist in putting your fanatstic misconceptions into someone else's mouth. .. -- +--------------------------------+ + Dick Pierce | + Professional Audio Development | +--------------------------------+ |
Reply |
|
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Is reverb always used on the lead vocal in commercial/pop recordings? | Pro Audio | |||
$3 Sale of CDs - Jazz, Classical | Marketplace | |||
drum samples used in commercial recordings | Pro Audio | |||
Why don't classical piano recordings sound as good as pop recordings? | High End Audio | |||
Best Non-Classical SACD Recordings? | High End Audio |