Spaceduck
Max Output Level: -50.5 dBFS
- Total Posts : 2499
- Joined: 2004/12/29 12:51:03
- Status: offline
Does anyone use 192kHz?
Maybe an old question... How many of you have done any projects using 192kHz sampling freq? Can you tell the difference, or will it just clutter up my hard drive w/ huge audio clips?
Also, what's the deal with cheap soundcards (SoundBlaster, etc) that handle 192kHz audio? Are they worthwhile?
|
ohhey
Max Output Level: 0 dBFS
- Total Posts : 11676
- Joined: 2003/11/06 16:24:07
- Location: Fort Worth Texas USA
- Status: offline
RE: Does anyone use 192kHz?
2005/09/15 10:44:21
(permalink)
Good question ! No, I have not. I need to do that testing with that. I'll tell you where I would use it. If I had a project that only needed a few tracks and the source was live mics recording very dynamic and detailed music. Or say, live to two track Jazz or Bluegrass.
|
daverich
Max Output Level: -41 dBFS
- Total Posts : 3418
- Joined: 2003/11/06 05:59:00
- Location: south west uk
- Status: offline
RE: Does anyone use 192kHz?
2005/09/15 11:01:51
(permalink)
I read an article the other day about how 192khz could actually sound WORSE than 96khz. Something to do with how the signal is reproduced. hmm, might've been on www.digido.com or http://recforums.prosoundweb.com Kind regards Dave Rich
|
wogg
Max Output Level: -57 dBFS
- Total Posts : 1819
- Joined: 2003/11/14 16:07:44
- Location: Columbus, OH
- Status: offline
RE: Does anyone use 192kHz?
2005/09/15 11:04:26
(permalink)
Can any microphones actually respond up to 96kHz?
Seriously, I'm waiting for my next CPU upgrade to really try out 96k. At this stage in the game 192k seems so over the top.
|
ohhey
Max Output Level: 0 dBFS
- Total Posts : 11676
- Joined: 2003/11/06 16:24:07
- Location: Fort Worth Texas USA
- Status: offline
RE: Does anyone use 192kHz?
2005/09/15 11:26:14
(permalink)
ORIGINAL: wogg
Can any microphones actually respond up to 96kHz?
Seriously, I'm waiting for my next CPU upgrade to really try out 96k. At this stage in the game 192k seems so over the top.
The sample rate is NOT the frequency responce, that doesn't mean it can record sound at 96K. You are still recording the band the converters are designed to capture and that should be about 20hz to 20,000hz. You are just taking more samples of the waveform every second, that's all. In fact when designing A to D converters one of the things you have to design in are the filters that chop off the frequencies that are not in the range you want to capture so they don't mess up the audible sound. In theory the yeild is about half of the highest frequency you want to capture so to get to over 20k with some room to spare 44.1 was selected as the sweet spot. So in theory you could design converters that would accuratly capture up to 96k when using 192k sampleing but there is no reason to do that for audio gear for humans, so it just isn't done.
post edited by ohhey - 2005/09/15 11:38:44
|
gullfo
Max Output Level: -86 dBFS
- Total Posts : 232
- Joined: 2004/10/15 01:48:08
- Location: Old Tappan, NJ
- Status: offline
RE: Does anyone use 192kHz?
2005/09/15 12:43:27
(permalink)
"Nyquist pointed out that the sampling rate needs only to exceed twice the signal bandwidth. What is the audio bandwidth? Research shows that musical instruments may produce energy above 20 KHz, but there is little sound energy at above 40KHz. Most microphones do not pick up sound at much over 20KHz. Human hearing rarely exceeds 20KHz, and certainly does not reach 40KHz. The above suggests that 88.2 or 96KHz would be overkill. In fact all the
objections regarding audio sampling at 44.1KHz, (including the arguments relating to pre
ringing of an FIR filter) are long gone by increasing sampling to about 60KHz. Sampling at 192KHz produces larger files requiring more storage space and slowing down the transmission. Sampling at 192KHz produces a huge burden on the computational processing speed requirements. There is also a tradeoff between speed and accuracy. Conversion at 100MHz yield around 8 bits, conversion at 1MHz may yield near 16 bits and as we approach 50-60Hz we get near 24 bits. Speed related inaccuracies are due to real circuit considerations,
such as charging capacitors, amplifier settling and more. Slowing down improves accuracy." http://www.lavryengineering.com/documents/Sampling_Theory.pdf based on Dan Lavry's work, it would appear around 60Khz would solve most issues in terms of bits and thus accuracy. going to 192khz based on the argument that the D/A and A/D processing takes a finite amount of time to complete would seem logical and thus may result is less quality rather than more...
post edited by gullfo - 2005/09/15 12:53:54
|
ohhey
Max Output Level: 0 dBFS
- Total Posts : 11676
- Joined: 2003/11/06 16:24:07
- Location: Fort Worth Texas USA
- Status: offline
RE: Does anyone use 192kHz?
2005/09/15 13:14:19
(permalink)
ORIGINAL: gullfo
"Nyquist pointed out that the sampling rate needs only to exceed twice the signal bandwidth.
What is the audio bandwidth? Research shows that musical instruments may produce energy
above 20 KHz, but there is little sound energy at above 40KHz. Most microphones do not pick
up sound at much over 20KHz. Human hearing rarely exceeds 20KHz, and certainly does not
reach 40KHz. The above suggests that 88.2 or 96KHz would be overkill. In fact all the
objections regarding audio sampling at 44.1KHz, (including the arguments relating to pre
ringing of an FIR filter) are long gone by increasing sampling to about 60KHz.
Sampling at 192KHz produces larger files requiring more storage space and slowing down the
transmission. Sampling at 192KHz produces a huge burden on the computational processing
speed requirements. There is also a tradeoff between speed and accuracy. Conversion at
100MHz yield around 8 bits, conversion at 1MHz may yield near 16 bits and as we approach
50-60Hz we get near 24 bits. Speed related inaccuracies are due to real circuit considerations,
such as charging capacitors, amplifier settling and more. Slowing down improves accuracy."
http://www.lavryengineering.com/documents/Sampling_Theory.pdf
based on Dan Lavry's work, it would appear around 60Khz would solve most issues in terms of bits and thus accuracy. going to 192khz based on the argument that the D/A and A/D processing takes a finite amount of time to complete would seem logical and thus may result is less quality rather than more...
Humm... it sounds like if your target media is CD then 88.2 would be perfect and for sound for DVD maybe 96K just so the downsample math will be easier. I think for future projects I'm going to try 88.2 and see how I like it.
|
Spaceduck
Max Output Level: -50.5 dBFS
- Total Posts : 2499
- Joined: 2004/12/29 12:51:03
- Status: offline
RE: Does anyone use 192kHz?
2005/09/15 13:17:10
(permalink)
Oh man this thread is making my head hurt. I'll have you know I failed out of one of the best audio engineering schools on the east coast.  So I sorta know what yall are saying, but I'm also dumb as a post. Anyway, it seems to confirm my gut feeling... 192 is not only overkill, it's downright wasteful. Ok, I can definitely hear the superiority of 96kHz over 44.1kHz hands down. But yeah, if our mics and instruments are all designed for human hearing, what's the point in adding all the supersonic overhead? Maybe to pick up fret buzz and crackles? Also, all my sampled instruments (piano, drums, loops, etc) are 44.1kHz. My audio FX plugins are 96k at best... so even if my mic could pick up dog whistles, wouldn't it be lost the minute I add digital FX to the track? Still, I'd love to hear a side-by-side test to be sure.
|
gullfo
Max Output Level: -86 dBFS
- Total Posts : 232
- Joined: 2004/10/15 01:48:08
- Location: Old Tappan, NJ
- Status: offline
RE: Does anyone use 192kHz?
2005/09/15 15:45:44
(permalink)
its not so much about the frequency response as much as # of samples made (usually the argument is based on 2 points but the reality that is only legit on simple waveforms, not complex ones IMHO) but pushing the aliasing artifacts above our hearing level so the audible sections are cleaner. As Dan points out, the 60K rate seems to be the point where this is optimal. but due to "standards" we have 88.2K and 96K as options.
|
Spaceduck
Max Output Level: -50.5 dBFS
- Total Posts : 2499
- Joined: 2004/12/29 12:51:03
- Status: offline
RE: Does anyone use 192kHz?
2005/09/15 16:20:37
(permalink)
ORIGINAL: gullfo
its not so much about the frequency response as much as # of samples made
Isn't that sorta the same thing, though? Suppose I blow a dog whistle at 192kHz (192,000 peaks per second) but my sampler only records 48k samples per sec... then my sampler will only record every 4th peak, and so it'll get aliased down from 192k to 48k (4 octaves down). Or something like that, right? I'm pretty sure no one plays the dog whistle, but aren't harmonics infinite? When you hit any single note, aren't there countless harmonics resonating from here to kingdom come? Maybe they can't be heard, but they somehow help shape the tone of the original note. *tries hard to remember all that Taylor Fourier series crap*
|
wogg
Max Output Level: -57 dBFS
- Total Posts : 1819
- Joined: 2003/11/14 16:07:44
- Location: Columbus, OH
- Status: offline
RE: Does anyone use 192kHz?
2005/09/15 16:58:46
(permalink)
The sample rate is NOT the frequency responce, that doesn't mean it can record sound at 96K. You are still recording the band the converters are designed to capture and that should be about 20hz to 20,000hz. You are just taking more samples of the waveform every second, that's all.
In fact when designing A to D converters one of the things you have to design in are the filters that chop off the frequencies that are not in the range you want to capture so they don't mess up the audible sound. In theory the yeild is about half of the highest frequency you want to capture so to get to over 20k with some room to spare 44.1 was selected as the sweet spot. So in theory you could design converters that would accuratly capture up to 96k when using 192k sampleing but there is no reason to do that for audio gear for humans, so it just isn't done.
That's all true... I was speaking from the perspective of frequency responce and Nyquist rate, perhaps not the proper place of contention. Many of the arguments I've heard for 96k were based on subconscience reactions to supersonic material (remembering vaguely). Of course the next argument becomes the artifacts from antialiasing filters. From my admittedly limited knowledge of DSP I thought those issues were largely eliminated at 88.2k or even less. What I was really getting at... I think the returns from going to 192k are far, far too small to outweigh the huge overhead in processing and storage.
|
gullfo
Max Output Level: -86 dBFS
- Total Posts : 232
- Joined: 2004/10/15 01:48:08
- Location: Old Tappan, NJ
- Status: offline
RE: Does anyone use 192kHz?
2005/09/15 19:46:02
(permalink)
most of the nyquist theory deals with steady state signals like a sine wave - when is the last time you saw a sine wave in your music? music is generally complex composite signals and is you hit the nyquist limit, you end up depending on the filters to convert the complex signal. the filters have real life limits in terms of quality vs. speed vs. depth. so by using a higher sampling rate, the number of sample points + the aliasing point of the filter are extended and you get more accurate reproduction at a decent quality (in theory...) [most] converters by their design "oversample" to begin with... standards "force" us to 88.2 or 96 if we want a rate higher than 44.1 or 48K.
frequency response and noise floor is still limited by the signal chain... so you're still likely to be restricted to 20-20K freq bandwidth [maybe] and probably 30-50db of noise all effectively limiting you to 10-18bits depending on whether you pick 16 or 24 bit.
|
yep
Max Output Level: -34.5 dBFS
- Total Posts : 4057
- Joined: 2004/01/26 15:21:41
- Location: Hub of the Universe
- Status: offline
RE: Does anyone use 192kHz?
2005/09/16 10:02:55
(permalink)
One reason to record at higher sample rates: Jitter is less objectionable at higher sample rates. Some jitter is always unavoidable, but using high-quality converters and a smart setup can keep it to a minimum. I'm still mostly working at 44.1, though.
Cheers.
|
Ressurrector
Max Output Level: -90 dBFS
- Total Posts : 20
- Joined: 2005/09/16 17:05:35
- Status: offline
RE: Does anyone use 192kHz?
2005/09/16 17:16:06
(permalink)
Someone help me out here. If I plan my destination to be just an audio cd then what benefits,if any would I gain by recording at 96khz vs 44khz? I do understand more samples a sec equals more pieces of the pie for a more accurate pic of a wave BUT wouldn't that just be lost in the conversion process? I do however record in 24bit btw. I seem to hear a difference there for sure.
post edited by Ressurrector - 2005/09/16 17:23:46
|
ohhey
Max Output Level: 0 dBFS
- Total Posts : 11676
- Joined: 2003/11/06 16:24:07
- Location: Fort Worth Texas USA
- Status: offline
RE: Does anyone use 192kHz?
2005/09/16 17:21:54
(permalink)
ORIGINAL: Ressurrector
Someone help me out here. If I plan my destination to be just an audio cd then what benefits,if any would I gain by recording at 96khz vs 44khz? I do understand more samples a sec equals more pieces of the pie for a more accurate pic of a wave BUT wouldn't that just be lost in the conversion process? I do however record in 24bit btw. I seem to hear a difference there for sure.
Yes, 24bit is more bang for the bits. If you are recording for CD the next boost would be to go to 88.2. It would be worth a try for one project just to see if you think it is worth the CPU usage and disk space. Downsampling goes a lot better from 88.2 to 44.1 then from 96k to 44.1 (ugly math).
|
Guest
Max Output Level: -25.5 dBFS
- Total Posts : 4951
- Joined: 2009/08/03 10:50:51
- Status: online
RE: Does anyone use 192kHz?
2005/09/16 18:26:04
(permalink)
could be a different question he's asking too ... 192k is not uncommon for multi-channel encodings ...
but i agree that for stereo it's such total overkill it's ridiculous.
jeff
|
j boy
Max Output Level: -48 dBFS
- Total Posts : 2729
- Joined: 2005/03/24 19:46:28
- Location: Sunny Southern California
- Status: offline
RE: Does anyone use 192kHz?
2005/09/16 19:42:46
(permalink)
I'll go to 192 kHZ... when I can find a 500 GB hard drive for my laptop.
|
Ressurrector
Max Output Level: -90 dBFS
- Total Posts : 20
- Joined: 2005/09/16 17:05:35
- Status: offline
RE: Does anyone use 192kHz?
2005/09/16 21:34:46
(permalink)
Ohhey u got me curious. I think I will fire up a 88.2k session tonight and see how I like it. I bet file size doubles doesn't it? If so that kinda sucks. But I kinda think maybe this can make the end product better the same way usin 24 bits does.
|