Go Back   Cockos Incorporated Forums > REAPER Forums > Recording Technologies and Techniques

Reply
 
Thread Tools Display Modes
Old 04-16-2007, 03:58 PM   #1
manning1
Human being with feelings
 
Join Date: Apr 2006
Posts: 2,959
Default recording at high sampling rates eg..96k....people should read this.

http://www.gearslutz.com/board/high-...x-96k-not.html

a very interesting thread imho.
particularly note the comments on page 3 by the esteemed
Mr dan lavry.
i dont feel as bad now staying at 44.1.
manning1 is offline   Reply With Quote
Old 04-16-2007, 05:35 PM   #2
Diogenes
Human being with feelings
 
Diogenes's Avatar
 
Join Date: Aug 2006
Location: A place that allows me to protect myself...
Posts: 8,245
Default

Yep. There's a good white paper at Dan's website. I read it long ago and have been at 44.1 or 48K ever since. Never tried 88.2 but if my tests at 96 and 192 are any indication... I can't hear the difference anyway...

D
Diogenes is offline   Reply With Quote
Old 04-16-2007, 05:50 PM   #3
Amberience
Human being with feelings
 
Amberience's Avatar
 
Join Date: Oct 2006
Location: swing on the spiral of our divinity
Posts: 2,037
Default

To be honest, I've never really strayed from 44.1 - I don't see much point. I've tried other sampling rates, and I don't hear any perceptible difference.

However I switched a few years ago from 16bit audio to 24bit. I definitely do percieve a difference there.
__________________
Amberience is offline   Reply With Quote
Old 04-17-2007, 05:26 PM   #4
ten
Human being with feelings
 
Join Date: Feb 2007
Location: Windsor, England
Posts: 73
Default

I havent noticed much if any when recording at 88/96 from 44, but plugins and especially softsynths noticably sound better at 88/96. There seems to be a much more apparent higher end and bass seems tighter and more rounded also. It may be plug/synth dependent but its enough to make me work at 88 all the time.

Ones ive noticed a difference on are..

Imposcar
Vstation
Albino
Vanguard

Some reverb and delay plugs also.

tom
__________________
Carlsberg dont code daw software, but if they did... http://reaper.fm
ten is offline   Reply With Quote
Old 04-17-2007, 06:20 PM   #5
todd24
Human being with feelings
 
Join Date: Nov 2006
Posts: 916
Default

I am no less confused by it all now.

Thanks anyways!
todd24 is offline   Reply With Quote
Old 04-17-2007, 06:37 PM   #6
Billoon
Human being with feelings
 
Join Date: Sep 2006
Location: Arse end of the earth.
Posts: 2,980
Default

My understanding of what Mr Lavry is saying is...60-70kHz would be ideal...but soundcards dont run at 60-70kHz so the best option is 88.2 or 96kHz...

From here...

http://www.gearslutz.com/board/high-...ml#post1234224

There is a "speed accuracy tradeoff"...and "Nowdays there are a number of good designers and ear people that find 60-70KHz sample rate to be the optimal rate for the ear. It is fast enough to include what we can hear, yet slow enough to do it pretty accurately. "

RE: 44.1kHz...

"44.1KHz can be a somewhat tight squeeze, especially when we keep “piling” attenuation on that 20KHz range – most mics have 3dB loss at around 20KHz, then there is the AD with 3dB at 20K, then the speaker, the processor… Pretty soon the accumulated impact is such that there is not much 20KHz left… "


So he suggests..

"Moving the sampling a little higher (be it 60Kh, 88.2 or 96K) takes some of the “offenders” out of the picture (all you need is a few KHz extra and the problem is gone). "


Anyone think ive misinterpreted?
Billoon is offline   Reply With Quote
Old 04-17-2007, 10:10 PM   #7
Xenakios
Human being with feelings
 
Xenakios's Avatar
 
Join Date: Feb 2007
Location: Oulu, Finland
Posts: 7,672
Default

I used to hear 20-21khz about 5 years ago, not any more, I think. However, my enjoyment of music and sounds hasn't really dramatically diminished because of that...

-X
Xenakios is online now   Reply With Quote
Old 04-18-2007, 07:04 AM   #8
RedStone
Human being with feelings
 
RedStone's Avatar
 
Join Date: Jan 2006
Location: British Columbia, Canada
Posts: 410
Default

From what I have been reading, what you hear in a recording (from your AD converter) will largly depend on the type of Filter it uses - I've noticed that with my interface, recording at 88.2khz sounds markedly cleaner, and has more impact - and I also notice some aliasing happening at 44.1khz. boo.

But I'm told that when this happens, it means that something is not right. My best guess is that a good anti-alias filter will prevent ALL aliasing.

So from what I can understand, some of the better AD converters capture at a higher sample rate (say, 88.2khz), and then digitally apply a well designed anti-alias filter before down-sampling the data and sending it off to the HD. BUT, if the filter isn't clean enough, or 'dirty' - this, I am led to believe, is where trouble starts to brew. Maybe it is also why some interfaces sound trashy at 44.1khz, and amazing at higher sample rates. aliasing sucks.
__________________
click.listen.enjoy
www.levityproject.com
RedStone is offline   Reply With Quote
Old 04-18-2007, 01:57 PM   #9
Justin
Administrator
 
Justin's Avatar
 
Join Date: Jan 2005
Location: NYC
Posts: 12,088
Default

Quote:
Originally Posted by RedStone View Post

But I'm told that when this happens, it means that something is not right. My best guess is that a good anti-alias filter will prevent ALL aliasing.

No practical filter can prevent ALL aliasing.. only a "perfect" filter can.. but many filters can reduce aliasing to such small levels that they are effectively inaudible
Justin is online now   Reply With Quote
Old 04-18-2007, 02:31 PM   #10
griz lee
Human being with feelings
 
griz lee's Avatar
 
Join Date: Jul 2006
Location: in a hotel room near you
Posts: 1,175
Default

Quote:
Originally Posted by Billoon View Post
My understanding of what Mr Lavry is saying is...60-70kHz would be ideal...but soundcards dont run at 60-70kHz so the best option is 88.2 or 96kHz...
"


Anyone think ive misinterpreted?
Absolutely spot on. I have a film sound design system which can run at 60k, and it sounds absolutely spectacular at that rate. Problem for me is, nothing else will play with it.
griz lee is offline   Reply With Quote
Old 04-18-2007, 04:03 PM   #11
Panatrope
Human being with feelings
 
Join Date: Jun 2006
Location: Melbourne, Australia
Posts: 28
Default

Bob Stuart of Meridian Audio published a study in Audio magazine several years ago as to what sampling frequency and bit depth was need to ensure no apparent degradation in digital conversion. The outcome was that a phase linear frequency response extending to 26kHz and (true) 20 bit resolution was required. (The latter is only achieved in top of the line convertors such as Prism etc). The current CD standard was influenced by the technology of the late '70s - the infinitesimal cost of storage and 100MBit/s in transmission speeds were unheard of. The original CD was going to be 36kHz and 14 bit. In the end Sony demanded 16 bit, and the sampling frequency was chosen becuase of the then reliance on VCR technology as the recording medium, and therefore had to be compatible with 50Hz and 60Hz systems. 44100 is the product of the squares of the first four prime numbers (2,3,5,7) and gives enough factors to ensure divisibility by 50 or 60. With some heroic filtering, a 20kHz bandwith could be quoted. (The first filters in the ADC process introduced severe distortion, hence the rise of Apogee).

With the wisdom of hindsight, it is a pity they didn't follow the telecommunications industry 8kHz digital hierarchy. Given Bob Stuarts findings, a sampling rate of 64kHz seems ideal, which is why I'm delighted with the reports that people are finding an interesting "sweet spot" in sampling rates around 60-70kHz. Means you can preserve flatness to above 20 khz, and a nice phase-linear rolloff up to an extinction frequency of 32kHz.

Most convertors have an extinction frequency above half the sampling rate, and the behaviour around cutoff can be interesting. Also, as people like Stanley Lipshitz have shown, the one-bit oversampling convertor can get itself into limit cycles introducing 'birdies' which might be confused with alias components. Modern convertors now use a multi-bit technique to eliminate these problems.

SO, based on my experience, a good 24 bit 44.1 convertor will deliver excellent results if the resulting signal is used without any significant post processing (minimalist clasical recording). However, higher resolution recording will always deliver you benefits in post processing - reverb, noise reduction, etc, because of the great number of samples to work on and hence greater precision/lower distortion. The downside is greater processing time and expanded storage requirements. 18 months of Moore's law takes care of the former, and competition in the storage market takes care of the latter. Personally, I would always record in 24/96 but in the interests of productivity I will usually got straight to 44.1 especially if it is targetted for FM transmission.

I close with an extract from Flanders and Swann's "Song of Reproduction"

High Frequency gain
Is easy to obtain
With every note neither sharp nor flat ..
The ear can't hear as high as that
Still I ought to please any passing bat
With my High Fidelity
__________________
Pan: (Gr) all, everything
Tropos: (Gr) turn.
Loose translation: all rounder!
Panatrope is offline   Reply With Quote
Old 04-18-2007, 07:29 PM   #12
Diogenes
Human being with feelings
 
Diogenes's Avatar
 
Join Date: Aug 2006
Location: A place that allows me to protect myself...
Posts: 8,245
Default

Hi Panatrope,

Interesting read... one thing though, I thought the "settling" on 44.1khz was due to Nyquist's research?

This is interesting stuff.

D
Diogenes is offline   Reply With Quote
Old 04-19-2007, 06:08 AM   #13
Panatrope
Human being with feelings
 
Join Date: Jun 2006
Location: Melbourne, Australia
Posts: 28
Default

The 36kHz proposal was based on a 15kHz specification for upper frequency. When the 20kHz passband was dictated by the market, then Nyquist theory dictated a frequency above 40kHz. Practically, a further 10% increase was needed to allow for filtering. 44100 was chosen as indicated previously (though 44.05kHz was used briefly in the USA because of the different TV standards - remember mastering was undertaken on VCRs and edited using video-like processes).

Nowadays, the high-resolution crowd are focussing on the newly-proposed DXD 384kHz sampling frequency standard. We are definitely well within the territory where the law of diminishing returns reigns ....
__________________
Pan: (Gr) all, everything
Tropos: (Gr) turn.
Loose translation: all rounder!
Panatrope is offline   Reply With Quote
Old 04-19-2007, 08:24 AM   #14
RedStone
Human being with feelings
 
RedStone's Avatar
 
Join Date: Jan 2006
Location: British Columbia, Canada
Posts: 410
Default

Quote:
Originally Posted by Justin View Post
No practical filter can prevent ALL aliasing.. only a "perfect" filter can.. but many filters can reduce aliasing to such small levels that they are effectively inaudible
that's fair - so a good anti-alias filter will effectively stop all aliasing as far as our ears are concerned - but not "actually". hmmm ... I wonder if it really is the filter in my m-audio gear that is making me hear all the weirdness at 44.1khz - my microphone has a response from 20hz-20khz exactly - and then a sample rate of 44.1khz would be nearly exactly double that.

heeey (D'oh!) ... maybe that's the *problem* - when a microphone can capture sound higher up the spectrum, a higher sample rate for capturing is necessary - and double the limit of the mic's (or our ear's) response is the "bare minimum" samplerate I can use with my mic since it's response is all the way up to 20khz. So then right now, I really only have 2.1khz of buffer for sampling errors like aliasing. And if my interface doesn't do any 'tricks' like up-sampling and digital anti-alias filters etc to make it's 44.1khz sample rate sound the absolute best that it can, then I'm up the creek and have to go higher or else aliasing is inevitable. Does that sound about right?
__________________
click.listen.enjoy
www.levityproject.com

Last edited by RedStone; 04-19-2007 at 08:27 AM.
RedStone is offline   Reply With Quote
Old 04-19-2007, 11:51 AM   #15
Diogenes
Human being with feelings
 
Diogenes's Avatar
 
Join Date: Aug 2006
Location: A place that allows me to protect myself...
Posts: 8,245
Default

Quote:
We are definitely well within the territory where the law of diminishing returns reigns ....
I agree. Actually, I think we passed the territory on the way to 192khz!

As I said elsewhere, IF I REALLY, REALLY try, I think I can hear a noticeable difference at 96 or 192 on SOME stuff. Now in a full mix of a band? No way... but that may be just ME! I'm not willing to burn the CPU cycles and disk space on "maybe".

D
Diogenes is offline   Reply With Quote
Old 04-19-2007, 12:42 PM   #16
Alistair S
Human being with feelings
 
Alistair S's Avatar
 
Join Date: Sep 2006
Location: UK
Posts: 1,490
Default

I understand (not in detail, but in principle) the reasons behind 44.1 being able to capture all of the detail up to 20KHz, and have been recording in 44.1 and 24 bit.

So far so good. However, I decided to experiment.

I rendered the final mix to 96K (i.e. I up-sampled). The resulting output sounds different - better than the stuff I rendered at 44.1. It also sounds better even after dithering back to 16 bit (in fact, the psycho dither is so good it is hard to hear much difference at all).

I am assuming that this has something to do with some of the VSTs (Reacomp for compression, Reaverb with IRs for reverb) working better at a higher sample rate - maybe? Maybe it has to do with the new antialiasing. Maybe it's to do with increased headroom giving it all room to breathe. I don't know.

Now I'm not sure whether to continue to record at 44.1 and render at a higher rate to see if this happens all the time or whether to record at higher rates to begin with.

I think the answer is to experiment and see what works best with the kit I have. And maybe that's the be-all and end-all.

Anyone know for sure why it would sound better?
Alistair S is offline   Reply With Quote
Old 04-19-2007, 12:47 PM   #17
manning1
Human being with feelings
 
Join Date: Apr 2006
Posts: 2,959
Default

heres something i'm finding also with ogg format....
can someone explain this ???
sometimes i'll render an rpp song project to ogg....
for back up purposes, and play it back,./
and i swear the ogg playback sounds better !!
(than the wav playback)
ao i'm thinking of mixing to final stereo master useing the ogg version.
anyone lse found this ??
manning1 is offline   Reply With Quote
Old 04-19-2007, 02:58 PM   #18
magicchord
Human being with feelings
 
magicchord's Avatar
 
Join Date: May 2006
Location: L.A. Cahleefornia
Posts: 278
Default

I've tried rendering - and even just realtime listening - at twice the sample rate of my track files, with upsampling, and it does sound clearer.
I think that's a good thing to do, as long as you have the processor overhead to handle it.
magicchord is offline   Reply With Quote
Old 04-19-2007, 03:58 PM   #19
Panatrope
Human being with feelings
 
Join Date: Jun 2006
Location: Melbourne, Australia
Posts: 28
Default

The improvement by upsampling may sound different, possibly better, but this may have more to do with the DAC in your sound card, and its post conversion filtering. The pioneering Benchmark DAC uses this principle to operate at a final sampling rate of 217kHz - the sampling rate conversion process is also used to eliminate the impact of jitter from the incoming digital signal. The post conversion hardware filtering can then be much simpler and, particularly, phase linear out past 20kHz.

This upsampling idea is not new. The CD specification was originally 14-bit at 44.1kHz. Philips and Sony were partners in the development (actually Philips did most of the design, Sony mainly provided expertise in error correcting codes which made the thing reliable - and of course marketing, including Akito Moro's famous dictum that the CD's playing time had to permit Beethoven's 9th symphony on a single disc). Philips had just completed their design of suitable quality (14 bit) DACs, when Sony insisted on 16 bit. Philips had no time left to redesign their DACs (remember this was 1980) but found they could run much faster than 44100. So they came up with the 4x oversampling process which also involved rudimentary noise-shaping, which delivered the equivalent of 16 bit performance up to 20kHz, and which required much simpler filtering to reject the anti-alias components now located above 100kHz. They called the process Bitstream, and the early Philips players (including my still-functioning original Marantz CD-52) were renowned for a much less "harsh" sound than the Sony players which used some fairly extreme hardware filtering with fairly savage phase distortion in the top of the band).

So your upsampling experience is not new, but it is most likely a product of your hardware which will perform better in the audio band (ie., up to 20kHz) at 96kHz sampling.

I have this theory about upper frequency limits which I can expound on at a later time, and possibly in a more appropriate thread. But as far as digital is concerned, the performance is dictated solely by the ADC and DAC hardware ... all else is basically storage.
__________________
Pan: (Gr) all, everything
Tropos: (Gr) turn.
Loose translation: all rounder!
Panatrope is offline   Reply With Quote
Old 05-10-2007, 08:27 AM   #20
chip mcdonald
Human being with feelings
 
chip mcdonald's Avatar
 
Join Date: May 2006
Location: NA - North Augusta South Carolina
Posts: 3,571
Default

96k does make a pretty big difference in my experience, but it's not immediately obvious by listening to just one track in all situations.

If you A/B a stereo miced cymbal setup with a room mic and some tom mics open (so you hear the bleed), it's more noticeable. Or if you've got a lot of acoustic guitars, or whispy vocals.

It sounds nicer on my setup (MOTU 24 I/O); but that could be a byproduct of the chips in it designed around 24/96 operation instead of 44.1. Different chips may have different "sweet spots" in their design/filtering, maybe.

The problem is I'm not sure if it translates to 16/44.1. I haven't done any "tests", and I think that's where your dither really comes into play. But I definitely think it sounds "nicer"....
__________________
]]]>-guitar lessons - www.chipmcdonald.com-<[[[
Experiencing Guitar: Essays from Teaching by Chip McDonald https://www.amazon.com/dp/1521877823..._QZJxAbA4GVDC1
chip mcdonald is offline   Reply With Quote
Old 05-28-2007, 10:28 AM   #21
HomoNeophilus
Human being with feelings
 
Join Date: May 2007
Posts: 42
Default

Quote:
Originally Posted by Diogenes View Post
I agree. Actually, I think we passed the territory on the way to 192khz!

As I said elsewhere, IF I REALLY, REALLY try, I think I can hear a noticeable difference at 96 or 192 on SOME stuff. Now in a full mix of a band? No way... but that may be just ME! I'm not willing to burn the CPU cycles and disk space on "maybe".

D
What you here in that case are artifacts from the higher spectrum causing distortion due to interference (or hopefully imagination if it's a well-built system)... There's really nothing but trouble if you go that high.

As for DXD... This really don't capture any higher freqs than 20000 it just uses high sampling rate to compensate for low bitrate (I think it's 8 bits for DXD). However, at least DSD was said to have a lot of energy out of band which caused a lot of trouble. Don't know about DXD, it's supposed to be a lot better.

I'd be all for a 60khz, 20bit PCM standard... Not likely to happen though with this samplerate race going on. Thankfully there's people like Lavry who won't compete with the others...
HomoNeophilus is offline   Reply With Quote
Old 05-28-2007, 08:01 PM   #22
Diogenes
Human being with feelings
 
Diogenes's Avatar
 
Join Date: Aug 2006
Location: A place that allows me to protect myself...
Posts: 8,245
Default

Quote:
I'd be all for a 60khz, 20bit PCM standard... Not likely to happen though with this samplerate race going on. Thankfully there's people like Lavry who won't compete with the others...
Yep. I agree with you... (and Dan Lavry) I also think it will never happen. The powers that be are already preparing to force "newer and better" down our throats to boost the economy! (Theirs...)

D
Diogenes is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -7. The time now is 01:39 PM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2019, vBulletin Solutions Inc.