Home |
Search |
Today's Posts |
#41
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
Gary Eickmeier wrote:
"Scott Dorsey" wrote in message ... It's because a parametric has whatever Q you want it to have, so you can match whatever it is that you're trying to deal with whether it is a narrowband problem or a wideband one. [...] I suppose I should look up a few of the parametric kind and see just what they can do, Unless they have adjustable centre frequencies and Qs, they can never equalise resonance-related problems correctly. Attempting to equalise a resonance with a peak or notch that is even slightly off-tune will produce horrible phasey-sounding effects on wideband sounds. I have spent countless hours trying to equalise badly-recorded 78s with parametric equalisers and I know just how difficult it is. That was with no more than three resonances and no delay comb effects. The chances of doing that properly with a conventional graphic effects unit are nil. -- ~ Adrian Tuddenham ~ (Remove the ".invalid"s and add ".co.uk" to reply) www.poppyrecords.co.uk |
#42
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
Scott Dorsey wrote:
Gary Eickmeier wrote: On equalizers, I wonder why they aren't all digital yet, with infinitely fine number of sliders that are set automatically by the RTA. I know that some receivers have these "room correction" modes, right? I would love to see one that measures the output of your system, then allows you to draw the curve that you want, and have it do it perfectly with no lumpies. Ah well, we'll see now it goes. These days, a lot of them _are_ digital. But in the studio world, people use parametrics instead of graphics so they don't have any of those problems to begin with. --scott The Behringer DEQ2496 is a good learning tool, and a good tool period in experienced hands. Two channels of multi-band, parametric, dynamic, auto, etc., EQ in one box. -- shut up and play your guitar * HankAlrich.Com HankandShaidriMusic.Com YouTube.Com/WalkinayMusic |
#43
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"hank alrich" wrote in message ... The Behringer DEQ2496 is a good learning tool, and a good tool period in experienced hands. Two channels of multi-band, parametric, dynamic, auto, etc., EQ in one box. Wow thanks Hank - this is serious. Their only critique was the reliability and support from the company. But it has an RTA! I already have a good calibration microphone from another unit that might work OK. There is no dedicated one for this unit anyway. Got to think long and hard on this one. Gary |
#44
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
On 5/9/2014 3:35 PM, Gary Eickmeier wrote:
"hank alrich" wrote in message ... The Behringer DEQ2496 is a good learning tool, and a good tool period in experienced hands. Two channels of multi-band, parametric, dynamic, auto, etc., EQ in one box. Wow thanks Hank - this is serious. Their only critique was the reliability and support from the company. But it has an RTA! I already have a good calibration microphone from another unit that might work OK. There is no dedicated one for this unit anyway. Got to think long and hard on this one. Gary This is an amazingly flexible unit for the price. I installed three of them in my old venue and I have two of my own. However, I did have three out of the box failures (DOA) and two that crashed after about a year of regular use. I traced the failures to overheating. [IMHO the things could use better ventilation or forced air cooling.] == Later... Ron Capik -- |
#45
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
On 5/9/2014 1:02 PM, Adrian Tuddenham wrote:
Gary Eickmeier wrote: "Scott Dorsey" wrote in message ... It's because a parametric has whatever Q you want it to have, so you can match whatever it is that you're trying to deal with whether it is a narrowband problem or a wideband one. [...] I suppose I should look up a few of the parametric kind and see just what they can do, Unless they have adjustable centre frequencies and Qs, they can never equalise resonance-related problems correctly. Attempting to equalise a resonance with a peak or notch that is even slightly off-tune will produce horrible phasey-sounding effects on wideband sounds. I have spent countless hours trying to equalise badly-recorded 78s with parametric equalisers and I know just how difficult it is. That was with no more than three resonances and no delay comb effects. The chances of doing that properly with a conventional graphic effects unit are nil. Adrian, Have you tried time reversal on those 78s? One can't un-ring a bell in real time but can get a lot closer in reverse time. Try reversing the track and applying the correction. Also note that some DSP EQs don't have the same Q for boost and cut. I did some experiments on this with test signals (broad band noise) in sound forge. I'd boost a frequency then apply the inverse, same Q and frequency. Forward boost, forward cut showed phase errors (time smearing) where forward boost, reverse cut was nearly error free. The higher the Q the greater the time smearing. Getting back to the topic at hand, the EQ in a room can only be applied in real time, so the time smearing will be additive even if the average power is flattened. Put another way: you can't suck the energy out of a ringing cavity, you can only reduce the excitation power. == Later... Ron Capik -- |
#46
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"Gary Eickmeier" writes:
"Frank Stearns" wrote in message ... Chuckle, no kidding. Freq response is only one thing - uneven or inappropriate reverb times and phase (time) irregularities being the others, often more significant. All you can do with graphic EQ is typically make it worse, even though it "measures" flat. Woo hoo. Parametric is "better" but is still only a partial solution for many underlying problems. But hey, whatever floats your boat or bumps your already peaky/dippy bass. (For that, traps are highly recommended.) Maybe I missed it. Tell me again, Gary, why you're looking for category of solution, particularly this one? I thought you were happy with your room. Frank Mobile Audio Hi Frank - Whew - I am getting some odd reactions about the whole subject of EQing a room or a speaker system. Basically, my system has no tone controls on the receiver. I have no big complaints, but some of my audio buddies are always wanting me to measure the FR in my system, so I did. Just a Radio Shack SLM, digital, and a B&K test CD with 30 bands of narrow pink noise, but it works and reveals some anomolies that I cannot correct with my system as is. I am using a Velodyne subwoofer and just setting it by ear. So in measuring at the listening position, I am getting a hump at 63 Hz that is about 5 dB higher than I would like, then fairly smooth thru the midrange from 100 Hz to 5k, then another hump at 8k of 5 or 6 dB too high, then falling off smoothly to 20k with the zero crossing at 12.5 and 8 dB down at 16k. -snips- Well, you've hit the nail on the head why so many of us are fanatical about room treatment. Choice 1: An EQ box to "correct" room issues Choice 2: Measurement gear, RealTraps (or DIY traps made of framing lumber, panel material, 703, caulking, and a range of tools), diffusors, absorbers, patience, skill. Choice 1 is easy, but seldom (if ever) works because it simply doesn't go after the most likely root causes of the problem(s), especially from 200 hz on down. Choice 2, when done right, produces incredible results. It's sound like you've never heard it, especially with reasonably good monitors. YMMV. Frank Mobile Audio -- |
#47
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"Frank Stearns" wrote in message acquisition... "Gary Eickmeier" writes: Hi Frank - Whew - I am getting some odd reactions about the whole subject of EQing a room or a speaker system. Basically, my system has no tone controls on the receiver. I have no big complaints, but some of my audio buddies are always wanting me to measure the FR in my system, so I did. Just a Radio Shack SLM, digital, and a B&K test CD with 30 bands of narrow pink noise, but it works and reveals some anomolies that I cannot correct with my system as is. I am using a Velodyne subwoofer and just setting it by ear. So in measuring at the listening position, I am getting a hump at 63 Hz that is about 5 dB higher than I would like, then fairly smooth thru the midrange from 100 Hz to 5k, then another hump at 8k of 5 or 6 dB too high, then falling off smoothly to 20k with the zero crossing at 12.5 and 8 dB down at 16k. -snips- Well, you've hit the nail on the head why so many of us are fanatical about room treatment. Choice 1: An EQ box to "correct" room issues Choice 2: Measurement gear, RealTraps (or DIY traps made of framing lumber, panel material, 703, caulking, and a range of tools), diffusors, absorbers, patience, skill. Choice 1 is easy, but seldom (if ever) works because it simply doesn't go after the most likely root causes of the problem(s), especially from 200 hz on down. Choice 2, when done right, produces incredible results. It's sound like you've never heard it, especially with reasonably good monitors. YMMV. Frank Mobile Audio OK, I can believe you about the bass traps, but what would cause an 8k hump? I realize that I am not making super careful measurements yet, but I still see no reason for this. Gary |
#48
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
Gary Eickmeier wrote:
OK, I can believe you about the bass traps, but what would cause an 8k hump? Loudspeaker linearity issue comes to mind as - in my opinion, I didn't say experience - the primary and usual suspect, but do also consider that glass has the property of letting some frequency ranges through and reflecting others. Actually making valid measurements in a room is not at all simple, Speaker Workshop did it to my liking well when I tried it some years ago, but I liked smoothing 1/4 octave warble-tone measurements in a calcsheet even better. Note: when you want to make the room "behave" smoothing information about sharp peaks and dips away from the measurement may not be the best choice. Kind regards Peter Larsen |
#49
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
Ron C wrote:
On 5/9/2014 1:02 PM, Adrian Tuddenham wrote: Gary Eickmeier wrote: "Scott Dorsey" wrote in message ... It's because a parametric has whatever Q you want it to have, so you can match whatever it is that you're trying to deal with whether it is a narrowband problem or a wideband one. [...] I suppose I should look up a few of the parametric kind and see just what they can do, Unless they have adjustable centre frequencies and Qs, they can never equalise resonance-related problems correctly. Attempting to equalise a resonance with a peak or notch that is even slightly off-tune will produce horrible phasey-sounding effects on wideband sounds. I have spent countless hours trying to equalise badly-recorded 78s with parametric equalisers and I know just how difficult it is. That was with no more than three resonances and no delay comb effects. The chances of doing that properly with a conventional graphic effects unit are nil. Adrian, Have you tried time reversal on those 78s? One can't un-ring a bell in real time but can get a lot closer in reverse time. Try reversing the track and applying the correction. I haven't tried reversal, but the effect of a pure resonance (without time-delay components) is the same on the attack and the decay of a sound, so it shouldn't make any difference. There was once a theory that reversing a wax cylinder during playback would "improve the way the stylus impinged on the grooves" and overcome distortion due to wear, but this is a load of rubbish. The only detectable differences came from slew-rate problems in the original reproducer, which altered (but did not improve) with reverse playback. A good-quality playback sounded the same in both directions. The frequency distortions I have been trying to counteract are those caused by the discontinuity at the mouth of a conical recording horn and various other effects such as multiple reflections inside the horn and resonances in the recording diaphragm. Many recordings were made with multiple horns, which cannot be corrected, but some single-horn recordings can be successfully equalised for mouth effects. Once the mouth effect has been equalised, the other effects can be heard, but they are usually quite trivial compared with the acoustics of the (usually dreadful) recording room, so there is no point in attempting any further equalisation. ... Also note that some DSP EQs don't have the same Q for boost and cut. I don't get that problem because I use analogue filters; that way I can swing them through the signal in real-time and use the phasing effects as audible clues to the correct settings. Big black knobs on calibrated scale plates are the way forward! Getting back to the topic at hand, the EQ in a room can only be applied in real time, so the time smearing will be additive even if the average power is flattened. Put another way: you can't suck the energy out of a ringing cavity, you can only reduce the excitation power. I think most attempts at equalising a three-dimensional room with a one or two-dimensional system are doomed to failure. I have only once partially succeeded in dealing with a problem like this, when I had to recover an historic mono recording of an artist performing a live concert in a hall with the P.A. system on the verge of feedback. I ran it through a parametric filter which was accurately adjusted to get rid of only the one most prominent peak. Once that had been removed, I did a second pass to get rid of the next most obvious peak. After several passes like that, the recording sounded half-decent - but it was never going to sound particularly good. -- ~ Adrian Tuddenham ~ (Remove the ".invalid"s and add ".co.uk" to reply) www.poppyrecords.co.uk |
#52
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
Gary Eickmeier wrote:
OK, I can believe you about the bass traps, but what would cause an 8k hump? Hard reflecting surfaces. Move the microphone around and see how much the response changes as you change position. You may find it changes dramatically if you move a few inches, you may find it it doesn't change at all. You may also find it's a speaker problem, but there is a good chance that it is a room problem. And it may not be one hump either, it might be lots of little humps on top of one another. I realize that I am not making super careful measurements yet, but I still see no reason for this. This is pretty typical of untreated living rooms, and it's the reason why we have controlled playback rooms in studios. --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#53
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
Don Pearce wrote:
In some recordings, if you are very fortunate, you may be able to isolate a drum beat or other transient sound. You can then use that to make a fair approximation to the impulse response of the room. That can then be reverse-convolved to remove the room response from the remainder of the recording. In practice, I usually find I prefer the sound of the original, complete with its shortcomings. The "improvement" artifacts are usually more objectionable. Folks have attempted to do this with acoustical 78s for years and years. Probably the first example was Stockham's attempt to identify the impulse response of the system used to record some Caruso recordings back in the day when DSP processing involved week-long jobs on a minicomputer. I think it is possible for a lot of recordings. Unfortunately there were quite a few recordings made with multiple horns, for example one horn on a singer and one on a piano, connected together with rubber tubing and a manifold. The resonances on the two horns are different and can never be separated out so it becomes a matter of compromise. --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#54
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
Scott Dorsey wrote:
Don Pearce wrote: In some recordings, if you are very fortunate, you may be able to isolate a drum beat or other transient sound. You can then use that to make a fair approximation to the impulse response of the room. That can then be reverse-convolved to remove the room response from the remainder of the recording. In practice, I usually find I prefer the sound of the original, complete with its shortcomings. The "improvement" artifacts are usually more objectionable. Folks have attempted to do this with acoustical 78s for years and years. Probably the first example was Stockham's attempt to identify the impulse response of the system used to record some Caruso recordings back in the day when DSP processing involved week-long jobs on a minicomputer. I'm doing it with analogue circuits in real time. Obviously, analogue is the way of the future. I think it is possible for a lot of recordings. Unfortunately there were quite a few recordings made with multiple horns, for example one horn on a singer and one on a piano, connected together with rubber tubing and a manifold. The resonances on the two horns are different and can never be separated out so it becomes a matter of compromise. It seems as though most 'ensemble' recordings were multi-horn jobs, including solos if they had orchestral accompaniment. The exception was a vocal solo with piano accompaniment, which would often have been taken on one horn. There were far more 'talking' records in the catalogue than nowadays, so they are a useful source of single-horn test material for anyone experimenting with an equalisation system. It seems that 'mixers' weren't unknown, either. I have found one recording with two completley different frequency responses, which change over when the singer stops and the orchestra plays a short section. It sounds like a gas tap is being used to shut off one of the horns. I have found some recordings pre-1925 by U.K. Columbia which do not respond to the usual horn equalisation. In some cases it is because they used a flared horn, not a conical one; but in other cases the most likely explanation is that they are experimental electric recordings using the Holman system. -- ~ Adrian Tuddenham ~ (Remove the ".invalid"s and add ".co.uk" to reply) www.poppyrecords.co.uk |
#55
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"Gary Eickmeier" writes:
OK, I can believe you about the bass traps, but what would cause an 8k hump? I realize that I am not making super careful measurements yet, but I still see no reason for this. Are you using a calibrated test microphone? Much above 500 hz do NOT trust the Rat Shacks. The mid-expensive mics aren't much better, but they often do give you a good calibration curve that you can input into your test system. Tried taking multiple measurements with the mic moved slightly each time? Remember, the wavelen at 8k is about 1 3/4" Any hard, reflective surfaces nearby? Might be getting some HF swarm building up, or an unfortunate strong phase-add at that hz based on some even multiple of the wavelen and distance between direct and bounce. It's going to vary because of delay between direct and bounce, but over time (say 2x the time of the bounce path to your measurement mic) things very well could be adding enough phase-add energy often enough to cause your bump. Because it's 8K, nothing fancy needed to test this idea. Temporarly hang some heavy blankets or comforters on the side walls. I realize you're screaming at this point, but do you want to find the cause of the bump, or not? Anything in the room resonating at that hz? Look for small glass, metal, or hardwood things, knick knacks, whatever, 1-2" in size. Frank Mobile Audio -- |
#56
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"Scott Dorsey" wrote in message ...
Probably the first example was Stockham's attempt to identify the impulse response of the system used to record some Caruso recordings back in the day when DSP processing involved week- long jobs on a minicomputer. This is theoretically possible. (I was advocating such things 40 years ago. It's unfortunate it hasn't been done with classic tape recordings.) Was was /actually/ done was a spectral analysis of modern recordings, which were compared with the spectral analyses of Caruso's recordings. This made it possible to apply more-or-less rational corrective EQ. At the time these recordings were released, people who'd heard Caruso live swore that the EQ'd version sounded more like the live Caruso. I'm not so sure. I was accustomed to hearing Caruso's voice unprocessed, and the EQ'd version didn't sound as pleasing. |
#57
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"Gary Eickmeier" writes:
"Frank Stearns" wrote in message news "Gary Eickmeier" writes: OK, I can believe you about the bass traps, but what would cause an 8k hump? I realize that I am not making super careful measurements yet, but I still see no reason for this. Are you using a calibrated test microphone? Much above 500 hz do NOT trust the Rat Shacks. The mid-expensive mics aren't much better, but they often do give you a good calibration curve that you can input into your test system. Tried taking multiple measurements with the mic moved slightly each time? Remember, the wavelen at 8k is about 1 3/4" Any hard, reflective surfaces nearby? Might be getting some HF swarm building up, or an unfortunate strong phase-add at that hz based on some even multiple of the wavelen and distance between direct and bounce. It's going to vary because of delay between direct and bounce, but over time (say 2x the time of the bounce path to your measurement mic) things very well could be adding enough phase-add energy often enough to cause your bump. Because it's 8K, nothing fancy needed to test this idea. Temporarly hang some heavy blankets or comforters on the side walls. I realize you're screaming at this point, but do you want to find the cause of the bump, or not? Anything in the room resonating at that hz? Look for small glass, metal, or hardwood things, knick knacks, whatever, 1-2" in size. Frank Mobile Audio Oh believe me I am not rejecting your suggestions. I have some hard reflective surfaces around the speakers and a glass coffee table right in front of the listening sofa. I would be very interested in experimenting with removing some of that, in both the measurements and the listening. Getting the equalizer Monday. Will try the room treatments first. Well, do be careful in assumptions. I put these things more or less in order of strongest probability. That is, above 500 hz and before anything else, I'd suspect the test microphone (unless it was a higher-end brand and a calibration curve had been applied). And, I'd want additional tests at slightly different positions. Frank Mobile Audio -- |
#58
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
Frank Stearns writes:
"Gary Eickmeier" writes: snips, and followup Well, do be careful in assumptions. I put these things more or less in order of strongest probability. That is, above 500 hz and before anything else, I'd suspect the test microphone (unless it was a higher-end brand and a calibration curve had been applied). Forgot to mention: Remember, 8KHz is in that range of your typical condensor mic resonance bump. Even the little electrets used in a lot of test mics have that, but it's either flattened in the mic electronics or addressed in the calibration curve. Frank -- |
#59
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"Gary Eickmeier" wrote in message
... I predict it will be very useful to have the ability to shape the response of the things with a 31 band. I will be using the Velodyne with those new speakers too, and this Behringer FBQ6200 equalizer has a subwoofer output and adjustable crossover freq that is almost designed for my situation. I will predict with great confidence that you would get a much better EQ from a Behringer DEQ2496. It has 31 bands that can eith respond like a traditional 31 band (with little peaks or dips between faders) or give you a 'smoothed' response simply following the curve you set. But the best part is the dual five band parametric EQ. Even if your goal is to not use any more EQ than you need it's still nice for 'what if' scenarios: e.g. what if you notch out that 8K bump or fill bring up that dip at 600 Hz- how will it sound at different point in the room? Yes it has an RTA screen you can measure the response of pink noise, and let it auto-set the graphic, but I like watching the RTA screen while moving the mic or speakers around. For what you're trying to do with your speakers and room, I highly recommend it. You would find a lot of uses for it. Sean |
#60
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"Frank Stearns" wrote in message
acquisition... Frank Stearns writes: "Gary Eickmeier" writes: snips, and followup Well, do be careful in assumptions. I put these things more or less in order of strongest probability. That is, above 500 hz and before anything else, I'd suspect the test microphone (unless it was a higher-end brand and a calibration curve had been applied). Forgot to mention: Remember, 8KHz is in that range of your typical condensor mic resonance bump. Even the little electrets used in a lot of test mics have that, but it's either flattened in the mic electronics or addressed in the calibration curve. See, I always thought that it wasn't that hard to get to get an acceptably flat response from a very small diaphagm omni just from the physics of how it works. Obviously it can be hurt by diffraction around the edge of the element or in the preamp electronics. Acceptable to me is say +- 1 dB for muso RTA purposes - not for critical measurements of say mics or speakers - which would need an anechoic chamber anyway... Sean |
#61
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"Sean Conolly" wrote in message ... "Gary Eickmeier" wrote in message ... I predict it will be very useful to have the ability to shape the response of the things with a 31 band. I will be using the Velodyne with those new speakers too, and this Behringer FBQ6200 equalizer has a subwoofer output and adjustable crossover freq that is almost designed for my situation. I will predict with great confidence that you would get a much better EQ from a Behringer DEQ2496. It has 31 bands that can eith respond like a traditional 31 band (with little peaks or dips between faders) or give you a 'smoothed' response simply following the curve you set. But the best part is the dual five band parametric EQ. Even if your goal is to not use any more EQ than you need it's still nice for 'what if' scenarios: e.g. what if you notch out that 8K bump or fill bring up that dip at 600 Hz- how will it sound at different point in the room? Yes it has an RTA screen you can measure the response of pink noise, and let it auto-set the graphic, but I like watching the RTA screen while moving the mic or speakers around. For what you're trying to do with your speakers and room, I highly recommend it. You would find a lot of uses for it. Sean Well, you highlight my quandary. I want to high pass my main speakers with the subwoofer crossover, either from the receiver or from the Behringer because that relieves the main speakers from trying to do the lowest freqs and gives them more power in their range. But how do I do that with the 2496? Maybe Frank is right though and it is more a microphone problem with the RS SLM. Frank - I tried covering the coffee table with a quilt and doing the measurement all over again, but got the same result. Trying now to use my calibration microphone that came with an Audiocontrol equalizer/analyzer. But the mike doesn't want to operate on its own without the analyzer that it came with. I stuck it into my Zoom H6 recorder and nothing, either with or withour power applied from teh recorder. So OK, maybe I have to stick it into the analyzer and then take a line signal out of there and into the recorder. Then I do a whole series all over again and take the recording into the edit room and analyze with Adobe Audition. Anyway, I think I am using the SLM correctly - it is set to fast, c weighted - is that right? But you may be right, because my AT 2050 mikes have a problem with a peak at about 10k as well. Gary |
#62
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"Mike Rivers" wrote in message ... What I'm surprised that I can't find, given how common computers with sound cards are, is a modern computerized version of the clunky General Radio chain-driven synchronized oscillator and plotter that we had in our college lab in 1960. Connect the device you want to test between the audio output and input of a computer, use a program to generate a slow sine wave sweep, and generate a plot of what comes back into the computer's audio input. There are a number of FFT programs but it's just not the same thing. Why not? Many programs can generate a sweep output on one channel and do a FFT meaurement on the other, and can apply tracking filters if necessary. Programs like SpectraLab were doing it a couple of decades ago. Trevor. |
#63
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"Gary Eickmeier" wrote in message news "hank alrich" wrote in message ... The Behringer DEQ2496 is a good learning tool, and a good tool period in experienced hands. Two channels of multi-band, parametric, dynamic, auto, etc., EQ in one box. Wow thanks Hank - this is serious. Their only critique was the reliability and support from the company. But it has an RTA! I already have a good calibration microphone from another unit that might work OK. There is no dedicated one for this unit anyway. Not so, I have both the DEQ2496 and the Behringer measurement mic. What it doesn't have is calibration data, not that any generic data you get with similar cheap measurement mics is of much use anyway. Actual calibration will always cost more than the mic unfortunately. But if you already have one you think is a "good one" it should work just fine. Trevor. |
#64
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"Ron C" wrote in message ... On 5/9/2014 3:35 PM, Gary Eickmeier wrote: "hank alrich" wrote in message ... The Behringer DEQ2496 is a good learning tool, and a good tool period in experienced hands. Two channels of multi-band, parametric, dynamic, auto, etc., EQ in one box. Wow thanks Hank - this is serious. Their only critique was the reliability and support from the company. But it has an RTA! I already have a good calibration microphone from another unit that might work OK. There is no dedicated one for this unit anyway. Got to think long and hard on this one. This is an amazingly flexible unit for the price. I installed three of them in my old venue and I have two of my own. However, I did have three out of the box failures (DOA) and two that crashed after about a year of regular use. I traced the failures to overheating. [IMHO the things could use better ventilation or forced air cooling.] Yep, a small computer fan helps, or you can build an external analog power supply if you want to go that route. (It is multi-voltage though) Trevor. |
#65
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
Sean Conolly wrote:
Frank writes: Forgot to mention: Remember, 8KHz is in that range of your typical condensor mic resonance bump. Even the little electrets used in a lot of test mics have that, but it's either flattened in the mic electronics or addressed in the calibration curve. See, I always thought that it wasn't that hard to get to get an acceptably flat response from a very small diaphagm omni just from the physics of how it works. Obviously it can be hurt by diffraction around the edge of the element or in the preamp electronics. It's not all that hard. As you point out, the two big issues are the diffraction and the resonance (not of the diaphragm but of the resonator placed in front of the diaphragm to extend the response). The thing is... the market right now is awash in "measurement microphones" that really aren't measurement microphones at all. For room work, you don't need a real IEC Type I measurement mike, and you can get away with one of the inexpensive electret mikes. But what makes the inexpensive electret mikes useful is the calibration curve that comes with it. You pay around $200 for a cheap but usable measurement mike, and my great suspicion is that about $25 of that is for the mike and $175 is for the calibration curve. Acceptable to me is say +- 1 dB for muso RTA purposes - not for critical measurements of say mics or speakers - which would need an anechoic chamber anyway... These days we can do speakers without an anechoic chamber using gated methods. Technology like MLS has made field measurement a thousand times easier than it was when all we had were swept sines. On the other hand, these methods bring new sources of errors to deal with. --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#66
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
Gary Eickmeier wrote:
Well, you highlight my quandary. I want to high pass my main speakers with the subwoofer crossover, either from the receiver or from the Behringer because that relieves the main speakers from trying to do the lowest freqs and gives them more power in their range. But how do I do that with the 2496? Maybe Frank is right though and it is more a microphone problem with the RS SLM. My inclination would be just to do it with a second-order passive filter rather than deal with active filtration. But it is probably easier to do with an active filter for the first cut approximation. You should have no problem doing that with any parametric that has a shelving filter option. --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#67
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"Gary Eickmeier" writes:
snips Frank - I tried covering the coffee table with a quilt and doing the measurement all over again, but got the same result. Trying now to use my No, not the table in front of you (though that's a good place to start) but the sides walls as well. This topic is huge. You're putting your toe into a large ocean, where there's a lot of stuff going on. If you're serious about this, you've got to first get tools you can trust. Those cost money, though if spend wisely and with knowledge going in you won't break the bank. Then you have to know how to interpret what you measure as it's not as cut and dried as we might hope. This takes experience, and careful thought and understanding. Then it's also good to experiment and re-measure -- get that first-hand experience. I'm all for active cross-overs for subs. I wasn't at first, and it took a lot of work to modify an active xover so that it was completely transparent, but totally worth the effort in the long run. There is so much more control, as a better active crossover will give you some extra controls to deal with what happens at the crossover freq. Frank Mobile Audio -- |
#68
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
Frank Stearns wrote:
I'm all for active cross-overs for subs. I wasn't at first, and it took a lot of work to modify an active xover so that it was completely transparent I'm quite happy with the Ashly I use with my videoputer. Frank Kind regards Peter Larsen |
#69
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"Gary Eickmeier" wrote in message
... snip extended diversion into denial I feel in a hurry to have something to demonstrate and tell the world. Then they can take it from there and refine all of the things you are talking about. That last sentence is one of the most hilarious things you've written here. As a classic text-book example of a "crank", you're usually not quite so entertaining. Thanks for your interest and don't stop reading or give up on me. I need all of you guys. As people for you to ignore? |
#70
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
On Sunday, May 11, 2014 5:48:54 PM UTC-6, Gary Eickmeier wrote:
I basically want to check and make the frequency response "correct" as part of the theory and I want to know what that very simple aspect of the problem is doing when it sounds the best by ear. I want to eliminate that aspect of reproduction as one of the main differences among speakers so that I can study what I am after, the spatial characteristics. Speaker designers, engineers and perceptual psychologist have been working on that "very simple aspect" since the 1920s, with only marginal results. Good luck trying to find a quick and easy answer. Peace, Paul |
#71
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"Peter Larsen" writes:
Frank Stearns wrote: I'm all for active cross-overs for subs. I wasn't at first, and it took a lot of work to modify an active xover so that it was completely transparent I'm quite happy with the Ashly I use with my videoputer. Ashly stuff is interesting. Often good, basic designs, but sometimes suffer from cheap parts. I took the XR1001 crossover (is that what you have?) and did these mods to get the sound: 1. Beefed up PS; hexfred dioides, HF decoupling across newer, somewhat larger filter caps. 2. Added decoupling at power pins of each IC. Tied to new ground buss. 3. Upgraded ICs 4. Removed redundant coupling caps, upgraded and HF bypassed others. 5. Removed all but one level pot (LF); changed other gain points to unity with 0.1% resistors. 6. Biggest help was converting the contour, LF gain, and xover frequency select from crappy cheap pots to Elma rotary switches loaded up with precision resistors. After all that, the thing has no sonic fingerprint, whatsover, along with very stable controls. (Those cheap pots had a bad habit of drifting, which caused no ends of problems.) And, ultimately, with proper crossing, the whole system just sounds so much better. Credits to folks I met here years ago who gave me those mod suggestions -- Monte McGuire, Stephen Sank, Jim Williams. Others here had some on-off suggestions that were most useful, Scott Dorsey and others who I'm forgetting; apolgies. Frank Mobile Audio -- |
#72
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
On Sunday, May 11, 2014 5:48:54 PM UTC-6, Gary Eickmeier wrote:
I basically want to check and make the frequency response "correct" as part of the theory and I want to know what that very simple aspect of the problem is doing when it sounds the best by ear. I want to eliminate that aspect of reproduction as one of the main differences among speakers so that I can study what I am after, the spatial characteristics. Unfortunately, getting the frequency response correct is the hard part. Don't forget that it ALSO implies getting the dispersion correct, since frequency response and pattern are intimately tied together. You are correct that if the response is not flat, you can't accurately determine the imaging accuracy. The problem is that the response is never really flat. --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#73
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
I basically want to check and make the frequency response "correct" as part of the theory and I want to know what that very simple aspect of the problem is doing when it sounds the best by ear. I want to eliminate that aspect of reproduction as one of the main differences among speakers so that I can study what I am after, the spatial characteristics. in that case, here are two options for you 1) use a fixed set of speakers, don't change them, just move them around...maybe buy a variable delay box.. 2) concentrate on the mid-range... create source material that is only mid-range and use test your various speakers to match in the mid range, it will be almost impossible to match speakers at the low and high end. If you can figure out how the spatial effects work in the mid range then you can try to expand to the top and bottom ends. And for those looking for low cost waterfalls.. try this http://www.qsl.net/dl4yhf/spectra1.html Also N track studio has a very good user interface with of an RTA and "draw the response you want EQ". have fun Mark |
#74
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
Frank Stearns wrote:
"Peter Larsen" writes: Frank Stearns wrote: I'm all for active cross-overs for subs. I wasn't at first, and it took a lot of work to modify an active xover so that it was completely transparent I'm quite happy with the Ashly I use with my videoputer. Ashly stuff is interesting. Often good, basic designs, but sometimes suffer from cheap parts. Never opened it, bought it with a rack case and something else from a pa company that had discontinued their disco-rental. Taking the bundle was the best deal. I took the XR1001 crossover (is that what you have?) Yes. It is not obnoxious but also not totally transparent, it definitely has a signature. It sits in front of a NAD 906 6 x 25 watts amp powering a pair of small KEF "rear loudspeakers", the old version with silk dome and a pair of ATC 9" Studio in factory recommended boxes. and did these mods to get the sound: That's one for the archive, thanks, imo it has way too many front panel controls, but tweaking the cross-over Q did nice things in terms of making the cornerplaced face2face subs vanish and having extended downwards range for the videoputers audio is one of my better ideas. I'm not gonna rush into modding it, but it is an interesting idea. Thank you very much Frank. Frank Mobile Audio Kind regards Peter Larsen |
#75
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
Peter Larsen wrote:
Scott Dorsey wrote: Peter Larsen wrote: Two important points: the treble may not appear to be detached and a subwoofer, if available, may not be detectable on male vox. Detached? Yes, apparently elevated after a dip, something like: [perception] xxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxx xxxxxx x xxxxx Are you talking about ringing (meaning that was a time domain plot with overshoot) or are you talking about trying to equalize a dip with a filter that wasn't the same width as the dip (meaning that was a frequency domain one)? --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#76
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
I'm an fan of equalizing the speaker only and letting the room do it's thing. A single reflection causes a comb effect and trying to cancel the ripples with an EQ just makes the impulse response into a horrible mess.
I was the designer of the dbx 2020 back around 1980 that was the first (I think ...) consumer RTA/equalizer combo that would auto- EQ a room based on pink noise and an octave EQ. It often didn't sound so great when it was done. But in a consumer environment it usually added more bass which as we all know is what 9 out of 10 consumers prefer Bob |
#77
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
wrote in message ...
I'm a fan of equalizing the speaker only and letting the room do its thing. A single reflection causes a comb effect and trying to cancel the ripples with an EQ just makes the impulse response into a horrible mess. I was the designer of the dbx 2020 back around 1980 that was the first (I think ...) consumer RTA/equalizer combo that would auto- EQ a room based on pink noise and an octave EQ. It often didn't sound so great when it was done. But in a consumer environment it usually added more bass which as we all know is what 9 out of 10 consumers prefer I did about a half-dozen installations using the Crown EQ-2. If the room had reasonably good acoustics to begin with, the equalization /always/ improved the sound. By the way, I used an analyzer whose test signal was a mixture of swept-frequency square waves. There are new approaches to EQ that (supposedly) take into account time factors. Guess Im going to have to learn about them. |
#78
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
William Sommerwerck wrote:
wrote in message ... I'm a fan of equalizing the speaker only and letting the room do its thing. A single reflection causes a comb effect and trying to cancel the ripples with an EQ just makes the impulse response into a horrible mess. I was the designer of the dbx 2020 back around 1980 that was the first (I think ...) consumer RTA/equalizer combo that would auto- EQ a room based on pink noise and an octave EQ. It often didn't sound so great when it was done. But in a consumer environment it usually added more bass which as we all know is what 9 out of 10 consumers prefer I did about a half-dozen installations using the Crown EQ-2. If the room had reasonably good acoustics to begin with, the equalization /always/ improved the sound. And see, my experience going into studios was the first thing I always did was disable those things, and disabiling them always improved the sound. By the way, I used an analyzer whose test signal was a mixture of swept-frequency square waves. Was this an MLS thing or something else? There are new approaches to EQ that (supposedly) take into account time factors. Guess Im going to have to learn about them. Well, that's not really equalization.... --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#79
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
wrote in message ... in that case, here are two options for you 1) use a fixed set of speakers, don't change them, just move them around...maybe buy a variable delay box.. 2) concentrate on the mid-range... create source material that is only mid-range and use test your various speakers to match in the mid range, it will be almost impossible to match speakers at the low and high end. If you can figure out how the spatial effects work in the mid range then you can try to expand to the top and bottom ends. And for those looking for low cost waterfalls.. try this http://www.qsl.net/dl4yhf/spectra1.html Also N track studio has a very good user interface with of an RTA and "draw the response you want EQ". have fun Mark Sean and Mark - Sorry I am late getting back to you - has been a very long week with little time to do some of my own stuff. Those programs look interesting, might be able to try them next week. Right now, I have just received the equalizer (6200) and finally got it mounted in the rack but not wired in yet. Anyway, the thing slid right into an available opening in the rack and will be very easy to install, all the way down to the sub out with variable crossover point. The info I had before on the response problems was no correct. I finally got the calibration mike working by sticking it into the Audiocontrol C101, and the readings I got with that agree perfectly with those of a reviewer friend from a couple years ago, so I trust those more. No peak at 8 to 10k But there are still problems that I will now be able to correct, for these or any other test speakers. I still can't understand why the calibration microphone will not work in anything but the C101. I inserted it into my Zoom H6 recorder and tried it with Phantom power or without, with other settings of microphone power but nothing worked. Anyone know what's up with that? Thanks, Gary Eickmeier |
#80
Posted to rec.audio.pro
|
|||
|
|||
Freq Response Graph Paper
"Gary Eickmeier" wrote in message
news "hank alrich" wrote in message ... The Behringer DEQ2496 is a good learning tool, and a good tool period in experienced hands. Two channels of multi-band, parametric, dynamic, auto, etc., EQ in one box. Wow thanks Hank - this is serious. Their only critique was the reliability and support from the company. But it has an RTA! I already have a good calibration microphone from another unit that might work OK. There is no dedicated one for this unit anyway. Got to think long and hard on this one. Like I said - you'll find many ways to use it - it's like a swiss army knife. My only regret is that I don't have two, one for the live rack and one for the home rack. But they do suffer from overheating and shutting down on occasions, so try to leave some space around it in the rack. Sean |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Chicken Or The Egg? Freq Response Or Impedance? Headphone Help! | Pro Audio | |||
What do these freq response and xover specs mean? | Tech | |||
I need headphones which has a low freq. response of 4Hz. Any brands to recommend? | Audio Opinions | |||
Output stage freq response question | Vacuum Tubes | |||
Freq response ADS AL6 | Car Audio |