Quantcast
What sample rate do you record at? - Page 4 - Gearspace.com
The No.1 Website for Pro Audio
What sample rate do you record at?
View Poll Results: What sample rate do you record at?
Record at 44
542 Votes - 43.15%
Record at 48
344 Votes - 27.39%
Record at 88
141 Votes - 11.23%
Record at 96
191 Votes - 15.21%
Depending on the project, I would record at 44 or 96.
108 Votes - 8.60%
Multiple Choice Poll. Voters: 1256. You may not vote on this poll

Old 3rd July 2009 | Show parent
  #91
Lives for gear
 
🎧 15 years
And while most slutz won't admit it, here's there real reason WHY 44 wins by a long shot...

The Emperor’s New Sampling Rate -- Are CDs Actually Good Enough?
BAS Experiment Explanation page - Oct 2007

"Not one listener, under any circumstances, could consistently distinguish between high-resolution audio that was passed through the 44.1kHz/16-bit CD “bottleneck” and audio that wasn't."

Old 3rd July 2009 | Show parent
  #92
Lives for gear
 
🎧 10 years
Quote:
Originally Posted by mmcfarlane ➑️
Unfortunately hard disks fail after a few years of use or non-use. Apparently if you don't spin them regularly the lubricants coagulate. If your backups are something that will be used in the next 6 months you are probably fine but if you archive a few hundred drives and then test them after 5 years you'll likely find a fairly high failure rate.

Much like the old practice of retensioning tape periodically, your backup drives need to be used, not sitting on a shelf. I don't think anyone bakes harddrives (like we do old tapes) but it's an interesting thought...

I suspect analog (or digital) tape has a longer life span than harddisk.

FWIW, we have several petabytes of storage where I work (about half the size of a gymnasium with racks of disks). We discard the drives on a 3 year cycle (1/3 per year) and copy the data to new drives.
The disk longevity issue will I think soon be a moot point. Not because disks will last any longer, but because you can already do backups to online services. These services themselves use redundant arrays, regular upgrades, and further backups, they'll just get bigger and cheaper.
Old 3rd July 2009 | Show parent
  #93
Registered User
 
Rick Sutton's Avatar
 
🎧 15 years
Quote:
Originally Posted by dkelley ➑️
you're correct, I've worked at 88.2 before based on that theory too. the issue is that many converters don't run as well at 88.2 as they do at any of the other rates... so you have to use your ears also.

certainly dithering is the biggest annoyance working at 96 or 48, so like you I don't use those rates unless I must.
Don't you mean SRC (sample rate conversion). Dithering has to do with going to a lower bit rate, ie 24 to 16.
Old 3rd July 2009 | Show parent
  #94
Gear Maniac
 
Rabbit's Avatar
 
🎧 15 years
Pretty much 48 @ 24 bit here. I can hear a real difference at that rate versus 44.1 @ 24 but not so much of a difference between 48 @ 24 vs 88.2 @ 24.
Old 4th July 2009 | Show parent
  #95
Lives for gear
 
doug hazelrigg's Avatar
 
9 Reviews written
🎧 10 years
According to Nyquist-Shannon, one must have a sample rate at least twice the highest frequency one wants to reproduce. A SR of 44.1kHz will yield a frequency range up to 22.05kHz. The average range of human hearing extends to about 20kHz. However, practically speaking, at age 48 I can barely hear a tone @ 16kHz, let alone freq's up to 20kHz. And, given that most music has a very low percentage of energy at these higher frequnecies, I find HIGHLY dubious these claims that people can hear the difference between 44.1kHz and higher sample rates. Basic science debunks what is probably a psychological effect at best

Equal loudness contours and audiometry - Test your own hearing

Last edited by doug hazelrigg; 4th July 2009 at 12:04 AM.. Reason: add link
Old 4th July 2009 | Show parent
  #96
Gear Guru
 
2 Reviews written
🎧 10 years
The only argument that potenitally holds water, to me, is that you are allowing the processing of the data to include those higher frequencies, and the equipment can 'hear' it even if we can't, and that those higher frequencies interact with the hearable frequencies and affects them during that processing, as it would in analog gear if you were doing the processing in the purely analog realm during tracking.

In a digital world plugs are often used, post-conversion, where otherwise they would have been analog (and pre-tape or pre-conversion), so none of the limitations of the capture mechanism would have applied to the signal in those cases.

Whether that's actually the case of course would have to be proven. But it's at least an argument that you can't immediately dismiss out of hand as inherently unscientific. Though, OTOH, the people who believe it probably have no scientific basis for their beliefs either.
Old 4th July 2009 | Show parent
  #97
Lives for gear
 
🎧 15 years
Quote:
Originally Posted by doug hazelrigg ➑️
According to Nyquist-Shannon, one must have a sample rate at least twice the highest frequency one wants to reproduce. A SR of 44.1kHz will yield a frequency range up to 22.05kHz. The average range of human hearing extends to about 20kHz. However, practically speaking, at age 48 I can barely hear a tone @ 16kHz, let alone freq's up to 20kHz. And, given that most music has a very low percentage of energy at these higher frequnecies, I find HIGHLY dubious these claims that people can hear the difference between 44.1kHz and higher sample rates. Basic science debunks what is probably a psychological effect at best

Equal loudness contours and audiometry - Test your own hearing
So how about recording a track at 32kHz instead of 44.1?
If you haven't tried it I recommend you do.
In the same vein have a listen to a 1kHz tone recorded at 2,200 samples per second.

There may be a good scientific/technical reason why Nyquist theory does not translate in the real world to lower frequencies than 22kHz. I'd be interested if someone could explain?

My hearing tops out a little under 17kHz, yet I can normally hear the difference between 44.1 and 96
Strangely I sometimes prefer the sound of a lower sample rate.
Old 4th July 2009 | Show parent
  #98
Lives for gear
 
🎧 15 years
Quote:
Originally Posted by Dean Roddey ➑️
The only argument that potenitally holds water, to me, is that you are allowing the processing of the data to include those higher frequencies, and the equipment can 'hear' it even if we can't, and that those higher frequencies interact with the hearable frequencies and affects them during that processing
Exactly my thoughts. Sub harmonics.

I have no scientific basis for this, it's just my reasoning after many hours pondering why higher sample rates sound different (and sometimes/often better) to me.
Old 4th July 2009 | Show parent
  #99
Lives for gear
 
DarkSky Media's Avatar
 
🎧 10 years
Quote:
Originally Posted by MarkRB ➑️
...yet I can normally hear the difference between 44.1 and 96
Strangely I sometimes prefer the sound of a lower sample rate.
I've had this experience also. Varies with the converters. I assume because converters typically don't handle higher SRs as well (more dropped samples etc). Whereas, the technology for 44.1 is pretty good these days.
Old 4th July 2009 | Show parent
  #100
Lives for gear
 
doug hazelrigg's Avatar
 
9 Reviews written
🎧 10 years
Quote:
Originally Posted by MarkRB ➑️
So how about recording a track at 32kHz instead of 44.1?
If you haven't tried it I recommend you do.
Surfan Stevens recorded his highly acclaimed album Illinoise on a Rolnand V-series recorder set to 32kHz (because according to Surfan he wanted to save disc sapce and didn't know any better). I have that album and it sounds pretty good to me.

Maybe that was your point?
Old 4th July 2009 | Show parent
  #101
Lives for gear
 
mmcfarlane's Avatar
 
🎧 10 years
Quote:
Originally Posted by doug hazelrigg ➑️
According to Nyquist-Shannon, one must have a sample rate at least twice the highest frequency one wants to reproduce. A SR of 44.1kHz will yield a frequency range up to 22.05kHz. The average range of human hearing extends to about 20kHz. However, practically speaking, at age 48 I can barely hear a tone @ 16kHz, let alone freq's up to 20kHz. And, given that most music has a very low percentage of energy at these higher frequnecies, I find HIGHLY dubious these claims that people can hear the difference between 44.1kHz and higher sample rates. Basic science debunks what is probably a psychological effect at best

Equal loudness contours and audiometry - Test your own hearing
The part of the theory you are missing is that any signal above the nyquist frequency will fold back (alias) as frequencies below the nyquist. Therefore, an analog filter must be applied prior to the AD conversion to 'drastically attenuate' all frequencies at and above the nyquist. The design of the antialias filter is an important part of the AD conversion circuit, and the filter starts attenuating at frequencies far below the nyquist. The antialias filter is not an abrupt cutoff, which would introduce ringing in the time domain (audio) signal, it is a sloping shelf.

For example, if the nyquist fruequency is 20K, a recorded 25K signal would alias and appear as a 15K tone on the sampled data. The job of a good antialias filter is to ensure all frequencies above the Nyquist are attenuated (near the noise floor) before the AD conversion.

Sucks, huh?

So, 44K samples/second does not accurately sample 22KHz frequencies because the antialias (hicut shelf) filter is kicking in at much lower frequencies. 48K gives a little more breathing room for the antialias filter to do its work outside the audible spectrum, and allows the strongest aliased frequencies to foldback still outside the audible spectrum.

That being said, I work at 48K/24bit at the conversion and save as 32bit floating point in the recorded files. I suspect that some ITB algorithms will actually perform better at 96K due to software design tradeoffs made to improve performance, but my quad core Mac can't handle my typical Cubase projects at 96K.

Last edited by mmcfarlane; 4th July 2009 at 08:50 PM.. Reason: clarification an grammar
Old 4th July 2009
  #102
Gear Maniac
 
Dimes's Avatar
 
🎧 10 years
48 has been my choice lately.
Old 5th July 2009 | Show parent
  #103
Gear Nut
 
🎧 10 years
I have no empirical proof to offer. But I'd like to know what some of you think of this theory. So far from what I have read in this thread and countless others on the subject. Is that there are some sound theories and plenty of info that we all mostly agree with somewhat. Given marketing also comes into focus perhaps some of the math only holds true depending on what equipment and or system you are using. How well the the said DAW is coded and what converters it is matched to.

That said for me when I was a Cubase user I recorded at 44.1/24bit then switched to Nuendo. In Nuendo I mostly recorded at 48/24bit but mixed to a analog console and captured it all to a Masterlink at 88.2 for Mastering. I started getting more urban pop gigs and hip hop and electronica, I then switched to OSX and Logic recorded and mixed at 44.1 24 bit. Same RME converters but man to ME Logic sounded way better than the Steinberg DAWS at 44.1 so there was no need to go to 48k.

Now I am doing more Rock and folk, Jazz and you name it using Pro Tools LE and HD.

Urban pop stuff hip hop ITB @ 44.1
OTB 48k

Same with most rock and alternative bands


Jazz,folk, singer songwriter gigs 96k OTB
88.2 ITB
Old 5th July 2009 | Show parent
  #104
Gear Guru
 
AllAboutTone's Avatar
 
🎧 15 years
Still at 44.1
Old 5th July 2009 | Show parent
  #105
Lives for gear
 
doug hazelrigg's Avatar
 
9 Reviews written
🎧 10 years
Quote:
Originally Posted by mmcfarlane ➑️
The part of the theory you are missing is that any signal above the nyquist frequency will fold back (alias) as frequencies below the nyquist. Therefore, an analog filter must be applied prior to the AD conversion to 'drastically attenuate' all frequencies at and above the nyquist. The design of the antialias filter is an important part of the AD conversion circuit, and the filter starts attenuating at frequencies far below the nyquist. The antialias filter is not an abrupt cutoff, which would introduce ringing in the time domain (audio) signal, it is a sloping shelf.

For example, if the nyquist fruequency is 20K, a recorded 25K signal would alias and appear as a 15K tone on the sampled data. The job of a good antialias filter is to ensure all frequencies above the Nyquist are attenuated (near the noise floor) before the AD conversion.

Sucks, huh?

So, 44K samples/second does not accurately sample 22KHz frequencies because the antialias (hicut shelf) filter is kicking in at much lower frequencies. 48K gives a little more breathing room for the antialias filter to do its work outside the audible spectrum, and allows the strongest aliased frequencies to foldback still outside the audible spectrum.

That being said, I work at 48K/24bit at the conversion and save as 32bit floating point in the recorded files. I suspect that some ITB algorithms will actually perform better at 96K due to software design tradeoffs made to improve performance, but my quad core Mac can't handle my typical Cubase projects at 96K.
My understanding is that the anti-alias filter is placed after the information has been oversampled, typically 128x, so that any foldback is well outside the audible spectrum
Old 5th July 2009 | Show parent
  #106
IM_
Gear Head
 
IM_'s Avatar
 
4 Reviews written
🎧 10 years
96/24

I have a very nice sampling clock in Prism ADA-8XR
and I really hear the difference in my projects' top end..

As mentioned above brickwall low-pass filter is much smoother at 96k.
Old 5th July 2009 | Show parent
  #107
Lives for gear
 
mmcfarlane's Avatar
 
🎧 10 years
Quote:
Originally Posted by doug hazelrigg ➑️
My understanding is that the anti-alias filter is placed after the information has been oversampled, typically 128x, so that any foldback is well outside the audible spectrum
Frequency aliasing happens as an unavoidable by-product of sampling. To avoid aliasing an analog antialias filter must occur before the data is sampled the first time.

Using the 20K Nyquist example in my previous post, once the analog signal is sampled you have no idea whether the 15K tone was originally 15K or 25K so you don't know what to do with it. Its too late to fix the problem. You must cut off everything above the Nyquist frequency before you sample the data into the digital domain.

Additional filtering may happen during subsequent over/down-sampling but you have to filter the analog signal first.

Note that this is the theoretically pure view. In a given recording in a given room with given equipment,.. there may not be much energy captured above the nyquist and so if it aliases it may be close enough to the noise floor to not cause much harm.

Still, I'd be surprised if any A/D circuits don't include (or are preceeded by) an analog hi-cut filter.

I suspect the anitalias filter may be part of the 'characteristic sonic signature' of some converters. Just a suspicion, no data here.

I'm not a converter designer, just a middle-aged geophysicist. I'd be interested to hear some converter geeks chime in on the sonic impact of antialias filters in modern or vintage converter designs.
Old 5th July 2009 | Show parent
  #108
Lives for gear
 
audiogeek's Avatar
 
🎧 10 years
well actually...

Quote:
Originally Posted by doug hazelrigg ➑️
According to Nyquist-Shannon, one must have a sample rate at least twice the highest frequency one wants to reproduce. A SR of 44.1kHz will yield a frequency range up to 22.05kHz. The average range of human hearing extends to about 20kHz. However, practically speaking, at age 48 I can barely hear a tone @ 16kHz, let alone freq's up to 20kHz. And, given that most music has a very low percentage of energy at these higher frequnecies, I find HIGHLY dubious these claims that people can hear the difference between 44.1kHz and higher sample rates. Basic science debunks what is probably a psychological effect at best

Equal loudness contours and audiometry - Test your own hearing
i'm not a physicist or psychologist, but i think you might be a bit off base hear. just because you can't identify a tone higher than 16k doesn't mean you aren't hearing a tone higher than 16k. FYI, after years of playing a big amp in small clubs, my hearing range slopes off right in the same ballpark as yours.

use a tone generator to play a 20k or 25k tone (and crank the output)... you may not hear the tone, but rest assured within a short while you'll start to get a splitting headache. these inaudible tones are very much present in the real world, i.e. sounds as created in the studio, on the stage, or elsewhere. to ignore their significance is to shortchange yourself as a recordist...

that said, i record at 48k/24bit.
Old 5th July 2009 | Show parent
  #109
Lives for gear
 
mdjice's Avatar
 
🎧 10 years
44.1
Old 5th July 2009 | Show parent
  #110
Registered User
 
Rick Sutton's Avatar
 
🎧 15 years
Quote:
Originally Posted by audiogeek ➑️
use a tone generator to play a 20k or 25k tone (and crank the output)... you may not hear the tone, but rest assured within a short while you'll start to get a splitting headache.
The acrid smell of a tweeter burning up always gives me a headache too!
Old 5th July 2009 | Show parent
  #111
Lives for gear
 
doug hazelrigg's Avatar
 
9 Reviews written
🎧 10 years
Quote:
Originally Posted by audiogeek ➑️
i'm not a physicist or psychologist, but i think you might be a bit off base hear. just because you can't identify a tone higher than 16k doesn't mean you aren't hearing a tone higher than 16k. FYI, after years of playing a big amp in small clubs, my hearing range slopes off right in the same ballpark as yours.

use a tone generator to play a 20k or 25k tone (and crank the output)... you may not hear the tone, but rest assured within a short while you'll start to get a splitting headache. these inaudible tones are very much present in the real world, i.e. sounds as created in the studio, on the stage, or elsewhere. to ignore their significance is to shortchange yourself as a recordist...

that said, i record at 48k/24bit.
But see, you're telling me on the one hand these freq's are important not to be ignored, and then on the other hand that they are irritating (at higher volume).

Which is my point: most music contains very low energy at these higher freq's, even if they are audible in some way. Reproduction doesn't necessarily suffer if they aren't included
Old 5th July 2009 | Show parent
  #112
Lives for gear
 
DarkSky Media's Avatar
 
🎧 10 years
Quote:
Originally Posted by Rick Sutton ➑️
The acrid smell of a tweeter burning up always gives me a headache too!
Heh. My thought precisely.
Old 11th August 2009 | Show parent
  #113
Lives for gear
 
gizeh12's Avatar
 
🎧 15 years
The whole discussion on sample rate goes further than just recording and playing i think.

-Why does 96khz sound more clare and deep even if you role off everything over 20?
- How come softsynths and native reverbs sound way better at 96khz (try Logics space designer at 96khz vs 44.1), even when you bounce them @ 96khz and convert them to 44.1?

Obviously there is more going on than just frequencies, also the processing is enhanced someway.

Can anyone elaborate on this i'm in the dark here?
Old 11th August 2009 | Show parent
  #114
Gear Guru
 
2 Reviews written
🎧 10 years
Quote:
Originally Posted by gizeh12 ➑️
The whole discussion on sample rate goes further than just recording and playing i think.

-Why does 96khz sound more clare and deep even if you role off everything over 20?
- How come softsynths and native reverbs sound way better at 96khz (try Logics space designer at 96khz vs 44.1), even when you bounce them @ 96khz and convert them to 44.1?
I don't think a lot of folks would agree that those bullet points are necessarily true. Have you ever put yourself to a blind test and proven that you can tell the difference?
Old 11th August 2009
  #115
Lives for gear
 
🎧 10 years
48/24

Just because...

If I was doing Classical. 96/24. Again just because. I could try and explain why but would probably end up just looking dumb.
Old 11th August 2009 | Show parent
  #116
Gear Guru
 
2 Reviews written
🎧 10 years
Since I'd gotten a new sound card, I did my last song at 88.2K, just as a performance test more than anything else, to see how much load it could handle before getting wierd. It just barely survived the test and wasn't going to be able to maintain super low tracking-friendly latencies if more tracks were added (partly because of BFD of course which I used and is very heavy weight.)

Anyway, I can't say as I think thought it sounded particularly better in sonic terms than previous efforts at 48K, taking into account as much as possible considerable changes in the tools used. Certainly not enough to justify the significantly higher overhead.
Old 11th August 2009 | Show parent
  #117
Lives for gear
 
stardustmedia's Avatar
still at 44.1, although i could do 192... but the hassle of sample rate conversion at the end stops me even trying it

i think it's way enough for electronic dance music, which is anyway overcompressed and superloud
Old 11th August 2009 | Show parent
  #118
Lives for gear
 
gizeh12's Avatar
 
🎧 15 years
Quote:
Originally Posted by Dean Roddey ➑️
I don't think a lot of folks would agree that those bullet points are necessarily true. Have you ever put yourself to a blind test and proven that you can tell the difference?
Have you ever tried to use Space designer at 96khz vs 44.1, or some softsynth?
I really can;t understand how anyone overhears that difference.

I did a blind test on different systems and even my mother (61 yrs old) heard the difference, no kiddin.
Old 11th August 2009
  #119
Here for the gear
 
listentommy's Avatar
 
🎧 10 years
Why 96 are there that many film people out there?

I think it's funny how many 96khz votes there where. I record CD Destined audio at either 44.1 or 88.2 (depending on how much they're paying me to eat up Hard-drive space) But why so many 96khz? If you're going to CD and you have perform a SRC anyways why would you want to do it from 96 as opposed to 88.2? There doesn't seem to be much reason behind it. Any thoughts... Now if you're going to DVD 96khz makes sense but otherwise, I'm kind of surprised.

At least this post didn't include 192khz, or we might see some fights, haha.
Old 11th August 2009 | Show parent
  #120
Here for the gear
 
listentommy's Avatar
 
🎧 10 years
After Reading...

Okay after reading some of posts that have come previous to mine, I'm realizing why I haven't jumped in this whole forum community thing earlier. It's like the blind leading the blind. It's like a mass community of people leading others with ideas that are incorrect, while I think there are advantages to this type of education, I also have to consider how easy it is to get some real misinformation about audio, which has really hurt the professional audio community. I wonder sometimes if Bob was smart to create this forum or foolish for creating a giant website full of snake oil type advice rambling chains and misinformation.

You guys need to read an AES journal or a white paper before posting some of the non-sense I've heard.

No wonder the iPod is the main way people listen to music this day in age.

If I can see this and I'm only 23 than something is seriously wrong.

-Tommy C.
πŸ“ Reply

Similar Threads

Thread / Thread Starter Replies / Views Last Post
replies: 4616 views: 615184
Avatar for smoke
smoke 7th May 2021
replies: 55 views: 32247
Avatar for IM WHO YOU THINK
IM WHO YOU THINK 13th October 2020
replies: 98 views: 39103
Avatar for dfghdhr
dfghdhr 5th June 2021
replies: 1296 views: 181657
Avatar for heraldo_jones
heraldo_jones 1st February 2016
Post Reply

Welcome to the Gearspace Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…

Forum Jump
Forum Jump