Home |
Search |
Today's Posts |
#1
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
Ok, so I understand that 44.1k is 44,100 samples per second and 48k is
48,000 samples per second. Obviously 48,000 is better. I'm not exactly sure what bit rate is though? CDs are 16 bit, DVDs are 24. What exactly does that mean though? Another question - if I'm recording a project to audio CD, is it better to just record at 16/44 since that's what the CD will be anyway, and I can save system resources? or should I do 24/48 and then dither it down, essentially changing what I originally heard? I read in the ProTools book by Berklee Press that its best to record on LE using 24/44 since you won't hear much difference between the 48k and 44.1k. Any insights into this? --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.541 / Virus Database: 335 - Release Date: 11/14/2003 |
#2
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
|
#3
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Ric Oliva" wrote in message
... Ok, so I understand that 44.1k is 44,100 samples per second and 48k is 48,000 samples per second. Obviously 48,000 is better. I'm not exactly sure what bit rate is though? CDs are 16 bit, DVDs are 24. What exactly does that mean though? The term bit "rate" does contribute to the confusion. I think bit "depth" would be better. The quick answer is sampling rate (rate makes sense here) is directly related to frequency response. Bit depth is directly related to dynamic range. According to the theory, your sampling rate needs to be twice the highest frequency you want to record. So theoretically, 44.1 gives you 22kHz response, which is beyond human hearing. In practice, the actual top end limit will be somewhat lower due to analog filtering required to keep the clock noise out of the audio. But still, anything over 44.1 is probably superfluous rather than "better." Each sample has to reflect the amplitude of the signal at that sample. That value is stored in a digital "word." We're talking about storing the value in either a 16-bit or 24-bit word. The more bits, the better the resolution, which in audio is refered to as "dynamic range." Another question - if I'm recording a project to audio CD, is it better to just record at 16/44 since that's what the CD will be anyway, and I can save system resources? or should I do 24/48 and then dither it down, essentially changing what I originally heard? I read in the ProTools book by Berklee Press that its best to record on LE using 24/44 since you won't hear much difference between the 48k and 44.1k. Any insights into this? 24/44. While your finished product can sound just fine to the vast majority of ears at 16-bit depth, 24 is still worthwhile for recording, applying effects (transforms) and mastering. The reason is that the greater dynamic range of the 24-bit depth manifests itself in a lower "noise floor." This extra "room" at the bottom of your dynamic range is valuable because each time you perform any kind of transform to your audio signal(s), you'll add a bit of noise due to rounding errors. A greater bit depth makes these errors smaller, and when you resample or dither your final, mastered recording to 16-bit, most of those rounding errors will hopefully live in those truncated bits. That's not to imply that you can't do a fair number of transforms on a 16-bit file without seriously degrading it. But there is at least a good argument for using greater bit depths for recording/editing. Moreso than for higher sampling rates, anyway. -------------------------------------------------- Denny Fohringer Itinerant guitarist -------------------------------------------------- Lessons and music: http://surf.to/dennyf Bands: http://bluepearlband.com http://doubletakeband.com -------------------------------------------------- |
#4
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
What he said.
;-) "Denny F" wrote in message ... "Ric Oliva" wrote in message ... Ok, so I understand that 44.1k is 44,100 samples per second and 48k is 48,000 samples per second. Obviously 48,000 is better. I'm not exactly sure what bit rate is though? CDs are 16 bit, DVDs are 24. What exactly does that mean though? The term bit "rate" does contribute to the confusion. I think bit "depth" would be better. The quick answer is sampling rate (rate makes sense here) is directly related to frequency response. Bit depth is directly related to dynamic range. According to the theory, your sampling rate needs to be twice the highest frequency you want to record. So theoretically, 44.1 gives you 22kHz response, which is beyond human hearing. In practice, the actual top end limit will be somewhat lower due to analog filtering required to keep the clock noise out of the audio. But still, anything over 44.1 is probably superfluous rather than "better." Each sample has to reflect the amplitude of the signal at that sample. That value is stored in a digital "word." We're talking about storing the value in either a 16-bit or 24-bit word. The more bits, the better the resolution, which in audio is refered to as "dynamic range." Another question - if I'm recording a project to audio CD, is it better to just record at 16/44 since that's what the CD will be anyway, and I can save system resources? or should I do 24/48 and then dither it down, essentially changing what I originally heard? I read in the ProTools book by Berklee Press that its best to record on LE using 24/44 since you won't hear much difference between the 48k and 44.1k. Any insights into this? 24/44. While your finished product can sound just fine to the vast majority of ears at 16-bit depth, 24 is still worthwhile for recording, applying effects (transforms) and mastering. The reason is that the greater dynamic range of the 24-bit depth manifests itself in a lower "noise floor." This extra "room" at the bottom of your dynamic range is valuable because each time you perform any kind of transform to your audio signal(s), you'll add a bit of noise due to rounding errors. A greater bit depth makes these errors smaller, and when you resample or dither your final, mastered recording to 16-bit, most of those rounding errors will hopefully live in those truncated bits. That's not to imply that you can't do a fair number of transforms on a 16-bit file without seriously degrading it. But there is at least a good argument for using greater bit depths for recording/editing. Moreso than for higher sampling rates, anyway. -------------------------------------------------- Denny Fohringer Itinerant guitarist -------------------------------------------------- Lessons and music: http://surf.to/dennyf Bands: http://bluepearlband.com http://doubletakeband.com -------------------------------------------------- |
#5
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Ric Oliva" wrote in message
Ok, so I understand that 44.1k is 44,100 samples per second and 48k is 48,000 samples per second. Obviously 48,000 is better. Why is this obvious? I'm not exactly sure what bit rate is though? CDs are 16 bit, DVDs are 24. What exactly does that mean though? That's not bit rate. It's sample size. Another question - if I'm recording a project to audio CD, is it better to just record at 16/44 since that's what the CD will be anyway, and I can save system resources? If you are going to do very much processing, it is wise to record with 24 bit samples to preserve dynamic range as you process the tracks. After you've mixed the channels you are going to distribute, dither them down to 16 bits. or should I do 24/48 and then dither it down, essentially changing what I originally heard? Dithering down is a fast operation with most software. A proper job of downsampling can involve quite a bit of processing time. If you're going to throw away all audio 22.05 KHz in the end, why bother ever recording it? I read in the ProTools book by Berklee Press that its best to record on LE using 24/44 since you won't hear much difference between the 48k and 44.1k. Any insights into this? Yes, I just made a post entitled "Why 24/96 sampling isn't necessarily better-sounding than 24/44 sampling" that addresses this question. You can also investigate this issue yourself by downloading and listening to files of the same musical sounds recorded in various sample formats, from http://www.pcabx.com/technical/sample_rates/index.htm . |
#6
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
If sample rate is viewed as resolution over time (horizontal axis),
bit rate is resolution of the amplitude (vertical axis). Each bit doubles the resolution, or in other words, the smallest increment of volume possible is hlaved with each additional bit. The difference between 16 bits and 24 bits is 2 to eighth power, or 256. That means between each volume increment in a 16 bit recording there are 256 intermediate steps added in a 24 bit recording. This means that volume changes can be portrayed far more accurately and smoothly. Also, when you manuipulate tracks with faders and plug-ins, you are essentially doing mathematical operations, so with much higher resolution the rounding errors are minimized. In practice, the result is increased dynamic range, better stereo imaging, smoother less grainy fades and reverb tails, and less worry about having to track "hot". There is no reason not to track atr 24 bits if you can. The only disadvantage is each sound file will be 150% bigger. The difference between recording at 44.1 and 48k, on the other hand, is pretty tiny. Many people (myself included) record at 44.1 so that you don't have to worry about doing a sample rate conversion somewhere down the line (to a 44.1 CD) which may do more harm than whatever tiny gain you are getting from the slightly higher sample rate. If you are concerned with using a higher sample rate, 88.2 seems to make more sense. If you are going to do all your mixing on an analog board, however, then you might as well use 48k. |
#7
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Ric Oliva" wrote in message m...
Ok, so I understand that 44.1k is 44,100 samples per second and 48k is 48,000 samples per second. Obviously 48,000 is better. I'm not exactly sure what bit rate is though? CDs are 16 bit, DVDs are 24. What exactly does that mean though? Another question - if I'm recording a project to audio CD, is it better to just record at 16/44 since that's what the CD will be anyway, and I can save system resources? or should I do 24/48 and then dither it down, essentially changing what I originally heard? I read in the ProTools book by Berklee Press that its best to record on LE using 24/44 since you won't hear much difference between the 48k and 44.1k. Any insights into this? --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.541 / Virus Database: 335 - Release Date: 11/14/2003 Basically speaking, the bit resolution determines the ability to describe the amplitude of a signal. Having 24 bits available gives you a safety cushion in digital recording, among other things. The same input signal that you are slamming to 0 dbfs in 16 bit format (not a good thing because of the possibility of "overs") can be recorded in 24 bit format with the same or better resolution while staying well below the red. In a MIX article years ago, Stephen St. Croix stated that, sound improvement-wise, he'd rather have 17 bits vs. 16 instead of 96 kHz sampling rate vs. 48 kHz, if he had to make a choice. Most pro's would rather work in the higher resolutions until the absolute last bounce or mix to 16/44.1. This is partly because, with digital processing (EQ, compression, etc.), the extra headroom yields real sonic benefits when recording, editing, etc. There are some that prefer to keep everything in 44.1 all the way through to avoid sample rate conversion at the end, but there is almost universal use of higher bit resolutions whenever possible. RP |
#8
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
In article , "Ric Oliva"
wrote: Ok, so I understand that 44.1k is 44,100 samples per second and 48k is 48,000 samples per second. Obviously 48,000 is better. I'm not exactly sure what bit rate is though? CDs are 16 bit, DVDs are 24. What exactly does that mean though? http://www.promastering.com/pages/techtalk.html Article 1 and 2 deal with bit depth and dither, article 3 with sampling rates. Recording at wordlegths higher than 16 bit is helpful. In practice, 20 is almost always as good as 24 for recording since A to D converters don't have the dynamic range to capture 24 bits and the lower bits just contain the self noise of the box. For digital processing, however, you want to use longer wordlengths like 48 bit, or at the very least, 32 bit floating point. Most simply stated, wordlength (or bit depth) is dynamic range. Bit rate actually means something a little different, but we won't get into that right now as you obviously are asking about bit depth. For every bit you get about 6 dB (just over actually) of dynamic range. 16 bit CD has 96dB while 24 bit has 144. Extra bits do not add headroom; they add footroom. 0 dB FS (full scale) represents the same value in both 16 bit and 24 bit audio. The extra bits come into play at the bottom of the range. You are able to record smaller events - sounds at a lower level. In addition to dynamic range, it also means noise. in 16 bit, there is a noise floor of -96dB while 24 bit has a noise floor of -144 dB. 24 bit offers no additional accuracy in the top 96db of the dynamic range. Actually, an 8 bit recording is just as accurate as a 24 bit recording from 0dBFS to -48 dB. The -48 dB noise floor is quite obtrusive and the 8 bit recording certainly sounds worse, but those top 48 dB are just as accurate as a 24 bit recording. If you took a 24 bit file and added 96 dB of noise, it would sound like an 8 bit file. Invariably any discussion of bit depths must eventually include dither. This, however, I'll leave to the tech talk articles I've pointed you to, or to a google search for the many posts that have appeared here in r.a.p. Be aware, however, that there are some common mistakes made quite often when discussing these subjects, so avoid the myths. Sometimes common sense tends to fail you until you understand how digital audio truly works, so some things that seem to make intuitive sense at first are actually technical rubbish. -- Jay Frigoletto Mastersuite Los Angeles promastering.com |
#9
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
Jay - atldigi wrote:
In article , "Ric Oliva" wrote: Ok, so I understand that 44.1k is 44,100 samples per second and 48k is 48,000 samples per second. Obviously 48,000 is better. I'm not exactly sure what bit rate is though? CDs are 16 bit, DVDs are 24. What exactly does that mean though? http://www.promastering.com/pages/techtalk.html Article 1 and 2 deal with bit depth and dither, article 3 with sampling rates. Recording at wordlegths higher than 16 bit is helpful. In practice, 20 is almost always as good as 24 for recording since A to D converters don't have the dynamic range to capture 24 bits and the lower bits just contain the self noise of the box. For digital processing, however, you want to use longer wordlengths like 48 bit, or at the very least, 32 bit floating point. Most simply stated, wordlength (or bit depth) is dynamic range. Bit rate actually means something a little different, but we won't get into that right now as you obviously are asking about bit depth. For every bit you get about 6 dB (just over actually) of dynamic range. 16 bit CD has 96dB while 24 bit has 144. Extra bits do not add headroom; they add footroom. 0 dB FS (full scale) represents the same value in both 16 bit and 24 bit audio. The extra bits come into play at the bottom of the range. You are able to record smaller events - sounds at a lower level. In addition to dynamic range, it also means noise. in 16 bit, there is a noise floor of -96dB while 24 bit has a noise floor of -144 dB. 24 bit offers no additional accuracy in the top 96db of the dynamic range. Actually, an 8 bit recording is just as accurate as a 24 bit recording from 0dBFS to -48 dB. The -48 dB noise floor is quite obtrusive and the 8 bit recording certainly sounds worse, but those top 48 dB are just as accurate as a 24 bit recording. If you took a 24 bit file and added 96 dB of noise, it would sound like an 8 bit file. Invariably any discussion of bit depths must eventually include dither. This, however, I'll leave to the tech talk articles I've pointed you to, or to a google search for the many posts that have appeared here in r.a.p. Be aware, however, that there are some common mistakes made quite often when discussing these subjects, so avoid the myths. Sometimes common sense tends to fail you until you understand how digital audio truly works, so some things that seem to make intuitive sense at first are actually technical rubbish. So, what answer is correct? Whiteswan, Rick Powell, and Jay have given three answers that sound good but are mutually exclusive. I've been at this a few years and I still don't know what is right. Does 24 bit give greater resolution than 16 bit or does it merely give a larger dynamic range without a finer resolution? Peter |
#10
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
Peter Gemmell wrote: So, what answer is correct? Whiteswan, Rick Powell, and Jay have given three answers that sound good but are mutually exclusive. I've been at this a few years and I still don't know what is right. Does 24 bit give greater resolution than 16 bit or does it merely give a larger dynamic range without a finer resolution? The way that finer resolution manifests is as a larger signal to noise ratio. The noise is due to quantization and the wider the sample, the lower the noise is relative to the maximum representable signal. The noise is an approximately random error of +-1/2 the value of the low order bit. It is inescapable. It is intimately related to the dynamic range because it determines how small the signal can be before it loses signifigance relative to that error noise limit. The ratio of how large a signal that can be represented to how small a signal can be represented is the dynamic range. In practice, I don't think that yet any front end to a 24 bit ADC is itself nearly as quiet as that quantization noise so that you will see specifications, if they are honest, that are signifigantly lower than the theoretical 144 dB SNR that can be achieved with 24 bits. Bob -- "Things should be described as simply as possible, but no simpler." A. Einstein |
#11
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
|
#12
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Jay - atldigi" wrote in message
White Sawn's satement seems to indicate that the extra bits are within the same dynamic range, thereby giving you greater detail within that range. You can't into the trap of viewing digital audio like it's digital imagery. Unfortunately, 24 bits leaves the top 96db range of 16 bit alone, but lowers the noise floor and allows the recording of audio events that are even smaller, at a lower level, i.e. below -96dB. 24 bits puts 16 extra levels between each pair of levels that exist with 16 bits. Thus, the resolution is increased at any level, not just the smallest one. The reduction of the noise floor due to 24 bits is a consequence of the extra resolution 24 bit coding provides between any of the two levels in a 16 bit representation. The two go together hand-in-hand because the coding is linear. The idea that adding bits does not increase resolution is yet another popular urban myth about digital. It's similar to the urban myth that analog has resolution below the noise floor. In an exactly linear system, whether digital or analog, the noise floor and resolution are exactly the same. In a nominally linear (i.e., real-world) system, whether digital or analog, the noise floor and resolution are nominally the same. |
#13
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
|
#14
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Arny Krueger" wrote in message ... "Jay - atldigi" wrote in message White Sawn's satement seems to indicate that the extra bits are within the same dynamic range, thereby giving you greater detail within that range. You can't into the trap of viewing digital audio like it's digital imagery. Unfortunately, 24 bits leaves the top 96db range of 16 bit alone, but lowers the noise floor and allows the recording of audio events that are even smaller, at a lower level, i.e. below -96dB. 24 bits puts 16 extra levels between each pair of levels that exist with 16 bits. Thus, the resolution is increased at any level, not just the smallest one. The reduction of the noise floor due to 24 bits is a consequence of the extra resolution 24 bit coding provides between any of the two levels in a 16 bit representation. The two go together hand-in-hand because the coding is linear. The idea that adding bits does not increase resolution is yet another popular urban myth about digital. It's similar to the urban myth that analog has resolution below the noise floor. So, if you're recording, say, someone's vocals at both 16 and 24 bits, and the peaks are at -6dB to 0dB FS, does the 24 bit recording represent more accurately the signal in that region than the 16-bit version? |
#15
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Tommi" wrote in message
"Arny Krueger" wrote in message ... "Jay - atldigi" wrote in message White Sawn's satement seems to indicate that the extra bits are within the same dynamic range, thereby giving you greater detail within that range. You can't into the trap of viewing digital audio like it's digital imagery. Unfortunately, 24 bits leaves the top 96db range of 16 bit alone, but lowers the noise floor and allows the recording of audio events that are even smaller, at a lower level, i.e. below -96dB. 24 bits puts 16 extra levels between each pair of levels that exist with 16 bits. Thus, the resolution is increased at any level, not just the smallest one. The reduction of the noise floor due to 24 bits is a consequence of the extra resolution 24 bit coding provides between any of the two levels in a 16 bit representation. The two go together hand-in-hand because the coding is linear. The idea that adding bits does not increase resolution is yet another popular urban myth about digital. It's similar to the urban myth that analog has resolution below the noise floor. So, if you're recording, say, someone's vocals at both 16 and 24 bits, and the peaks are at -6dB to 0dB FS, does the 24 bit recording represent more accurately the signal in that region than the 16-bit version? The 24 bit recording has the capability to represent the signal much more accurately in *any* range from zero to max, than the 16 bit recording. |
#16
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
|
#17
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Tommi" wrote in
: So, if you're recording, say, someone's vocals at both 16 and 24 bits, and the peaks are at -6dB to 0dB FS, does the 24 bit recording represent more accurately the signal in that region than the 16-bit version? The extra 8 bits give you 48 db more dynamic range between EVERY sample. Between sample value = 0 and sample value = 1 they give you an extra 48 db on the bottom end. On the loud end, 16 bit max value is 32767 (0x7FFF), second value is 32766 (0x7FFE). That equates to 24 bit values 8388352 (0x7FFF00) and 8388096 (0x7FFE00), a difference of 256 values, the equivalent of 48 dB dynamic range. |
#18
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
Arny and I obviously agree about the resolution issue. I believe Arny
is cvorrect, but I make no claim to infallibility, so I'm very interested in the opposing points of view as well. Having no tech expertise, I'm going more on intuitve logic. Let's use reductio ad absurdum (or whatever it's called!). Let's take a 1 bit system. Now we have only two volume values: full volume, and full silence. using the 6dB per bit formula, we have values of 0dB and 6 dB. (I'm not using full scale dB, obviously, for this purpose). Now let's assume a 2 bit system. Now we have inserted two intermediate steps between the on and off. And we've also increased the dynamic range by 6 dB. This would give us values of 0db, 4 dB, 8 dB, and 12 dB. (For the sake of argument, let's assume a linear system in terms of dB values. For all I know, it might not be.) So what has happened? Yes, we have increased out dynamic range by 6 dB between the loudest and softest signals the system can represent. But we have increased the resolution throughout the system: from a 6dB increment to a 4 dB increment. Continuing to a 3 bit system: dynamic range is 18 dB. Values are 0dB, 2.25 dB, 4.5 dB, etc. to 18 dB. Wth each additional bit the dynamic range is increasing, but ALSO the resolution is increasing everywhere in the system. Now, my logic may well be flawed, so I'm most interested in finding out where the flaw is. This is a great way to learn, and i thank everyone who is teaching me! |
#19
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
|
#20
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Arny Krueger" wrote in message
"Jay - atldigi" wrote in message White Sawn's satement seems to indicate that the extra bits are within the same dynamic range, thereby giving you greater detail within that range. You can't into the trap of viewing digital audio like it's digital imagery. Unfortunately, 24 bits leaves the top 96db range of 16 bit alone, but lowers the noise floor and allows the recording of audio events that are even smaller, at a lower level, i.e. below -96dB. 24 bits puts 16 extra levels between each pair of levels that exist with 16 bits. Thus, the resolution is increased at any level, not just the smallest one. Correction: 24 bits puts 256 (!!) additional levels between every pair of levels that exist with 16 bits. |
#21
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Carey Carlan" wrote in message
. 205 "Tommi" wrote in : So, if you're recording, say, someone's vocals at both 16 and 24 bits, and the peaks are at -6dB to 0dB FS, does the 24 bit recording represent more accurately the signal in that region than the 16-bit version? The extra 8 bits give you 48 db more dynamic range between EVERY sample. Between sample value = 0 and sample value = 1 they give you an extra 48 db on the bottom end. On the loud end, 16 bit max value is 32767 (0x7FFF), second value is 32766 (0x7FFE). That equates to 24 bit values 8388352 (0x7FFF00) and 8388096 (0x7FFE00), a difference of 256 values, the equivalent of 48 dB dynamic range. Agreed. I figured that out a few hours after I posted, but I was nowhere near a computer with internet access. I then started wondering who would be the first to catch my error. |
#22
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
|
#23
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
|
#24
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
|
#26
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Jay - atldigi" wrote in message ... The 24 bit number is more precise than the 16 bit. True enough. What that means in audio, however, is that the 24 bit word can describe smaller values than the 16 bit word, thus signals that are lower in level. The 16 bit number is already describing 96 dB of dynamic range just fine. If you want to carry the precision further and capture signals that are lower, say to -144 dB, then 24 bits is your ticket. The myth is the dynamic equivalent to the argument that 4 samples on a 20kHz sine wave will render it more accurately than 2, and 8 samples even more so. That's not true either. I may well be suffering the myth, but my understanding is that it matters whether you sample a sine wave 2 or 8 times. Tests have been made where subjects had to determine which sound came first from their headphones. The same signal was fed to both L and R channels, only the other one was delayed by 5-15 _micro_seconds. Some of the people were able to "localize" the sound source even when it was delayed only by 5 microseconds. This implies that a sampling rate of 192kHz(which results in 5.2 microsecond's sample intervals), for example, is not only pushing the nyquist rate to the ultrasonic range, but also presents better channel separation on multichannel systems. So, it doesn't necessarily matter if you sample a sine wave 2 or 8 times on a mono system, but on a multichannel system higher sample rates result in better localization. |
#27
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
In article , "Tommi"
wrote: I may well be suffering the myth, but my understanding is that it matters whether you sample a sine wave 2 or 8 times. Tests have been made where subjects had to determine which sound came first from their headphones. The same signal was fed to both L and R channels, only the other one was delayed by 5-15 _micro_seconds. Some of the people were able to "localize" the sound source even when it was delayed only by 5 microseconds. This implies that a sampling rate of 192kHz(which results in 5.2 microsecond's sample intervals), for example, is not only pushing the nyquist rate to the ultrasonic range, but also presents better channel separation on multichannel systems. So, it doesn't necessarily matter if you sample a sine wave 2 or 8 times on a mono system, but on a multichannel system higher sample rates result in better localization. You have to take it one step at a time and separate the issues. 2 samples is enough to reconstruct the wave plain and simple. Bob Stuart and Tom Holman have talked about the possibility of better time axis resolution as it pertains to differences between two or more channels, not to be confused with time axis resolution meaning a more detailed representation of the waveform, and we're not just talking about single sine waves here. These are two different issues. It may well be that imaging improves with higher smaple rates, unless of course you dither properly at the lower ones. It's a little known fact that dither can also have an effect in the time domain. Other things including filter issues certainly can make higher sample rates sound better. However, this has nothing to do with the waveform being reproduced more accurately within the bandwidth of the system (i.e below Nyquist for the particular sample rate). -- Jay Frigoletto Mastersuite Los Angeles promastering.com |
#28
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Jay - atldigi" wrote in message ... You have to take it one step at a time and separate the issues. 2 samples is enough to reconstruct the wave plain and simple. Bob Stuart and Tom Holman have talked about the possibility of better time axis resolution as it pertains to differences between two or more channels, not to be confused with time axis resolution meaning a more detailed representation of the waveform, and we're not just talking about single sine waves here. These are two different issues. It may well be that imaging improves with higher smaple rates, unless of course you dither properly at the lower ones. It's a little known fact that dither can also have an effect in the time domain. Other things including filter issues certainly can make higher sample rates sound better. However, this has nothing to do with the waveform being reproduced more accurately within the bandwidth of the system (i.e below Nyquist for the particular sample rate). I absolutely agree, multichannel imaging is a different matter and that wasn't the topic here. |
#29
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
Jay - atldigi wrote in message ...
Rick understands that bit depth relates to amplitude and that DSP is better with longer wordlengths. A small clarification is in order, however. He seems to consider there to be extra headroom while technically there is not, unless you change the zero reference. In other words, increase the voltage that zero is referenced to. Nevermind working at -10 or +4, you'll be using a new, nonstandard reference voltage, and what about the analog electronics that probably can't handle that voltage? You're asking for trouble for that reason and several others (increasing the noise floor of the analog gear, compatability, and more). Unless you want to do that, you really are gaining what should be thought of as "footroom" more than headroom. In practice some feel that you need to push a digital recording right up to 0dB FS to "use all the bits". This really isn't as big an issue as some would have you believe, as long as you use good gain stageing and reasonable recording levels, especially with todays converters which perform far better than much or the early crappy digital stuff. It doesn't hurt to assumne that 24 bits gives you a little room to play with, but unless you are recording a program with greater than average dynamic range in a very quiet environment with excellent equipment and minimal processing, you really aren't going to be able to take advantage of those extra bits. Then again, they certainly don't hurt, and they could help, so there's no reason not to. Still, it helps to understand technically what's going on and when extra effort will pay off and when if won't. Jay, I'm not suggesting changing the zero reference. Correct the following if I'm wrong, but as a mastering engineer, you would rather take in a 2-track digital mix that peaked at -2db than one that peaks at 0 dbfs and has a few "flat tops". Using a 24 bit format to record or mix down to allows less artifacts towards the noise floor, given 2 "identical" sources (one recorded at 16 bit and one at 24 bit) peaking at, say, -2db. And reduces the need (perceived or real) to "slam" the recording all the way to 0 dbfs to take "full advantage" of the bit depth. If this is "footroom' instead of "headroom", isn't it still a margin nonetheless? RP |
#30
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
Tommi wrote: "Jay - atldigi" wrote in message ... You have to take it one step at a time and separate the issues. 2 samples is enough to reconstruct the wave plain and simple. Bob Stuart and Tom Holman have talked about the possibility of better time axis resolution as it pertains to differences between two or more channels, not to be confused with time axis resolution meaning a more detailed representation of the waveform, and we're not just talking about single sine waves here. These are two different issues. It may well be that imaging improves with higher smaple rates, unless of course you dither properly at the lower ones. It's a little known fact that dither can also have an effect in the time domain. Other things including filter issues certainly can make higher sample rates sound better. However, this has nothing to do with the waveform being reproduced more accurately within the bandwidth of the system (i.e below Nyquist for the particular sample rate). I absolutely agree, multichannel imaging is a different matter and that wasn't the topic here. Within the Nyquist criterion a signal can be produced with any arbitary phase or delay until you consider the quantization of the samples. Then the achievable delays become quantized as well and wider sample widths will have a positive effect on the delay/phase resolution (which controls the imaging resolution.) I don't know for sure but I rather doubt that the ear is sensitive to the resolution constraint imposed by even 16 bit samples. Bob -- "Things should be described as simply as possible, but no simpler." A. Einstein |
#31
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
|
#32
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
Ric Oliva wrote:
Ok, so I understand that 44.1k is 44,100 samples per second and 48k is 48,000 samples per second. Obviously 48,000 is better. I'm not exactly sure what bit rate is though? CDs are 16 bit, DVDs are 24. What exactly does that mean though? Another question - if I'm recording a project to audio CD, is it better to just record at 16/44 since that's what the CD will be anyway, and I can save system resources? or should I do 24/48 and then dither it down, essentially changing what I originally heard? I read in the ProTools book by Berklee Press that its best to record on LE using 24/44 since you won't hear much difference between the 48k and 44.1k. Any insights into this? The minimum necessary sample rate and bit depth is determined by the sounds you want to reproduce. The dynamic range (basically signal-to-noise ratio) of your material determines the minimum necessary bit depth. In practice, you will never record a source with a dynamic range greater than can be represented in 16 bits. The bandwidth (frequency range) determines the necessary sample rate. The sample rate is required only to be more than double the highest frequency you want to reproduce. In practice, almost nobody owns reproduction equipment that is useful beyond 20kHz. So in theory you can record at 44.1/16 and your digital audio will hold all of the audio data necessary to reproduce anything you can put into and pull out of any equipment. The need for greater data rates comes when you plan to manipulate the recording one or more times between the record and reproduce moments. If you are going to re-sample your data (through sample rate conversion or a D/A-A/D process), then you might benefit from an increased sample rate. More importantly, if you plan to manipulate the VALUES of those existing samples (by DSP processes such as gain changes, EQ, or anything else), then you might benefit from an increased bit depth. Both of these considerations serve to push the limitations of the quantization processes beyond our ability to detect them. The idea is that stacking these processes can compound their inherent errors and eventually make them audible. Even this precaution represents a judicious level of overkill in almost all cases, but data storage is getting cheap so it doesn't hurt. In summary: If you're recording live to 2-track and won't be doing any processing at all, then 44.1/16 is more than adequate. If you will be doing any processing, then start with 24 bits. If you will be doing any resampling, then start with 96k. Whatever rate you choose for your initial recording, you should maintain that rate until the final stage of processing. |
#33
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Jay - atldigi" wrote in message
In article , "Arny Krueger" wrote: "Tommi" wrote in message "Arny Krueger" wrote in message The idea that adding bits does not increase resolution is yet another popular urban myth about digital. It's similar to the urban myth that analog has resolution below the noise floor. So, if you're recording, say, someone's vocals at both 16 and 24 bits, and the peaks are at -6dB to 0dB FS, does the 24 bit recording represent more accurately the signal in that region than the 16-bit version? The 24 bit recording has the capability to represent the signal much more accurately in *any* range from zero to max, than the 16 bit recording. I think you're suffering the myth, Arny. Let me quote from another thread where Scott Dorsey is trying to explain the same thing that I am, and I'll and try to explain it yet another way: In article , (Scott Dorsey) wrote: A 16 bit number is significantly smaller and therefore less precise than a 24 bit number. Right. So, in a nutshell. Moving from 16 bit to 24 bit, we have 8 extra bits per sample to represent the analog wave which is a massive gain. Not really. It gives you more dynamic range, which is often wasted anyway. 96 dB is an awful lot. Scott was clearly addressing the "massive gain" part of the comment, not the "8 extra bits per sample". That there are 8 extra bits per sample is an inarguable fact that we can all agree on. The 24 bit number is more precise than the 16 bit. True enough. What that means in audio, however, is that the 24 bit word can describe smaller values than the 16 bit word, thus signals that are lower in level. It means that, but it also means that a 24 bit word describes 255 additional levels between every pair of levels that can be described by a 16 bit word. It means both things. Again, this is inarguable and readily observable in the real world. The 16 bit number is already describing 96 dB of dynamic range just fine. Psychoacoustically 16 bits does do a fine job, however technically the 24 bit word codes 255 additional levels between each pair of levels described by a 16 bit word. If you want to carry the precision further and capture signals that are lower, say to -144 dB, then 24 bits is your ticket. 24 bits also adds resolution in any region between -144 dB and full scale. The myth is the dynamic equivalent to the argument that 4 samples on a 20kHz sine wave will render it more accurately than 2, and 8 samples even more so. That's not true either. That an accrete 24 bit representation of a signal has more resolution at any level than a 16 bit representation of signal is readily observable in the real world as soon as you have converters that are sufficiently accurate, which we now have quite commonly. Look at how this works with DC levels. In the following examples some numbers may be off by 1 which is obviously practically irrelevant. If you have a 16 bit converter with a 1 volt range, there are 65,535 levels that can be uniquely described between 0 and 1 volt. 1.0000 volts is represented by 65,535. 0.0000 volts is represented by 0. 0.50000 volts is represented as 32,767 and there are 32,767 unique levels between 0 and 0.5000 volts and 32,767 more unique levels between 0.5000 volts and 1.000 volts. The smallest voltage 0 volts that can be coded is 3.0518509475997192297128208258309e-5 volts. If you have a 24 bit converter with a 1 volt range, there are 16,776,960 levels that can be uniquely described between 0 and 1 volt. 1.0000 volts is represented by 16,776,960. 0.0000 volts is represented by 0. 0.50000 volts is represented as 8,388,480 and there are 8,388,480 unique levels between 0 and 0.5000 volts and 8388480 more unique levels between 0.5000 volts and 1.000 volts. The smallest voltage 0 volts that can be coded is 5.9605554283970397497520408941787e-8 volts. Thus the 24 bit representation of voltages between 0 and 1 volt has both more dynamic range and also more resolution than the 16 bit representation of the same voltages. The same concept relates to audio signals. |
#34
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Tommi" wrote in message
I may well be suffering the myth, but my understanding is that it matters whether you sample a sine wave 2 or 8 times. Tests have been made where subjects had to determine which sound came first from their headphones. The same signal was fed to both L and R channels, only the other one was delayed by 5-15 _micro_seconds. Some of the people were able to "localize" the sound source even when it was delayed only by 5 microseconds. This implies that a sampling rate of 192kHz(which results in 5.2 microsecond's sample intervals), for example, is not only pushing the nyquist rate to the ultrasonic range, but also presents better channel separation on multichannel systems. For the purpose of discussion, I'll stipulate that your facts are correct to this point. I really don't know that, but it would help me make an important point if we don't argue over that part of your comments. So, it doesn't necessarily matter if you sample a sine wave 2 or 8 times on a mono system, but on a multichannel system higher sample rates result in better localization. The myth here is that signals in a digital system can have interchannel timing differences that are only integer numbers of sample periods. IOW this myth as applied to 44,100 Hz sampling is that interchannel timing differences can only be multiples of 22.675736961451247165532879818594 microseconds. I agree that this seems to be intuitively clear. But it is also quite wrong. The myth comes from the idea that two signals in different channels that are displaced in time can only be expressed as the same set of sample values, but time-shifted. This is not the case. Two signals in different channels that are displaced in time can be expressed as different sample values. For example, if two slowly-increasing (ramp) signals are displaced in time, one signal might have a set of sample values that starts out 0, 10, 20, 30... This is a ramp that starts at t = 0. The time-delayed version of this signal in another channel could have a set of values that is 0 at t = 0, but is 5, 15, 25... for successive samples. If you looked at these two signals over time, you'd say that the second signal is time-shifted by an amount of time equal to half a sample period. And, that is how it would sound. The correct time resolution of sampled signals is the sample period divided by the number of distinct amplitude levels. In the case of 16/44 this would be 5.1418904674492623958124444033093e-10 seconds or 514.18904674492623958124444033093 picoseconds. This is a tiny, tiny number. In reality, it is lost in the noise. |
#35
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Bob Cain" wrote in message
Within the Nyquist criterion a signal can be produced with any arbitrary phase or delay until you consider the quantization of the samples. Agreed. Then the achievable delays become quantized as well and wider sample widths will have a positive effect on the delay/phase resolution (which controls the imaging resolution.) Agreed. I just gave examples of this in another post. I don't know for sure but I rather doubt that the ear is sensitive to the resolution constraint imposed by even 16 bit samples. I estimated the minimum time delay that is quantizable in 16/44 in the other post and found that it was 514.18904674492623958124444033093 picoseconds. Not only is this a very small amount of time in the audio domain, but in reality it is lost in the noise, due to Shannon's theories. Shannon and Nyquist really had their acts together back in the late 1920s and early 1930s. |
#36
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Mike Rivers" wrote in message
news:znr1069010813k@trad In article writes: So what has happened? Yes, we have increased out dynamic range by 6 dB between the loudest and softest signals the system can represent. But we have increased the resolution throughout the system: from a 6dB increment to a 4 dB increment. Continuing to a 3 bit system: dynamic range is 18 dB. Values are 0dB, 2.25 dB, 4.5 dB, etc. to 18 dB. Wth each additional bit the dynamic range is increasing, but ALSO the resolution is increasing everywhere in the system. So your next question should be: "What do I HEAR that's different?" A good way to answer that question is to listen to some very low bit rate recordings. I know this sounds like blasphemy, but once you get up to about 8 bits, you don't get the sense that you're increasing resolution, you get the sense that you're reducing the background noise level into which the signal disappears. This is clearly audible in a series of samples, originally recorded by Ethan Winer, that people can download from http://www.pcavtech.com/test_data His article is good background reading and can be found at http://www.ethanwiner.com/BitsTest.html . So yes, your ears are able to RESOLVE a lower level signal in the presence of noise because the resolution down there is better. However, on the practical side, since most of the music we listen to today has a dynamic range of less than 10 dB and is played back well above the system and ambient noise floor, you don't get much of a chance to take advantage of the added resolution. Of course it doesn't hurt to have it there (for the occasions where you actually can use it) but I'll bet you could sell 8-bit pop music CDs today and nobody would complain about the sound quality. Sad but true. Only I'd raise the bar to about 10 bits. People can listen to the files at http://www.pcavtech.com/test_data and reach their own conclusions. |
#37
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
"Jay - atldigi" wrote in message
In article , (White Swan) wrote: Let's take a 1 bit system. Now we have only two volume values: full volume, and full silence. using the 6dB per bit formula, we have values of 0dB and 6 dB. Something to ponder: Why does DSD (SACD) work? Because it's a different data stream and it's not PCM. |
#38
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
Jay - atldigi wrote in message ...
In article , (Rick Powell) wrote: Jay, I'm not suggesting changing the zero reference. I know; I was going off on a bit of a tangent there while making what was really just a small point. but as a mastering engineer, you would rather take in a 2-track digital mix that peaked at -2db than one that peaks at 0 dbfs and has a few "flat tops". Using a 24 bit format to record or mix down to allows less artifacts towards the noise floor, given 2 "identical" sources (one recorded at 16 bit and one at 24 bit) peaking at, say, -2db. And reduces the need (perceived or real) to "slam" the recording all the way to 0 dbfs to take "full advantage" of the bit depth. If this is "footroom' instead of "headroom", isn't it still a margin nonetheless? Sure, I'd prefer a 24 bit file that's not squashed against the ceiling for sure. It can't hurt. In a practical sense, there's nothing really wrong with your post, but from a technical standpoint there were a few details to clarify. I don't mean to be overly critical, but to balance some circulating misconceptions I think it's worth getting specific now and then. Thanks, Jay. I appreciate your presence here in RAP, and your knowledge of digital theory. RP |
#39
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
In article , "Arny Krueger"
wrote: "Jay - atldigi" wrote in message In article , (White Swan) wrote: Let's take a 1 bit system. Now we have only two volume values: full volume, and full silence. using the 6dB per bit formula, we have values of 0dB and 6 dB. Something to ponder: Why does DSD (SACD) work? Because it's a different data stream and it's not PCM. But how do they make 1 bit work if it can only be off or full 6dB? The point is that digital doesn't actually work that way. -- Jay Frigoletto Mastersuite Los Angeles promastering.com |
#40
|
|||
|
|||
16 bit vs 24 bit, 44.1khz vs 48 khz <-- please explain
In article , "Arny Krueger"
wrote: That there are 8 extra bits per sample is an inarguable fact that we can all agree on. That we can agree on, but how audio behaves because of them we don't. The 24 bit number is more precise than the 16 bit. True enough. What that means in audio, however, is that the 24 bit word can describe smaller values than the 16 bit word, thus signals that are lower in level. It means that, but it also means that a 24 bit word describes 255 additional levels between every pair of levels that can be described by a 16 bit word. It means both things. Again, this is inarguable and readily observable in the real world. The 16 bit number is already describing 96 dB of dynamic range just fine. Psychoacoustically 16 bits does do a fine job, however technically the 24 bit word codes 255 additional levels between each pair of levels described by a 16 bit word. Looking at the system as a whole, dither and all, you gain no advantage in the top 96dB. You get and extra 48dB below them. -- Jay Frigoletto Mastersuite Los Angeles promastering.com |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Explain me this | Audio Opinions | |||
TS/TRS balanced/unbalanced can someone explain | General | |||
Can you explain this 50Hz hum?? | Pro Audio | |||
Reverb & EQ and "damping" etc .. please explain .. | Pro Audio |