Jump to content
Sign in to follow this  
Lavry tech

The Optimal Sample Rate for Quality Audio

Recommended Posts

Interested in the facts?

 

One of the world’s top converter designers Dan Lavry has written a new paper in simple language to demystify the subject.

 

http://www.lavryengineering.com/pdfs/lavry-white-paper-the_optimal_sample_rate_for_quality_audio.pdf

 

See why many professional engineers still work at 96kHz years after 192kHz became available.

 

Find out why “more” is not always “better!”

Share this post


Link to post
Share on other sites

Thank you for posting this. If only people will read it and understand it then a world of good may come from it.


And always keep in mind: Cognitive biases, like seeing optical illusions are a sign of a normally functioning brain. We all have them, it’s nothing to be ashamed about, but it is something that affects our objective evaluation of reality. 

Share this post


Link to post
Share on other sites

Find out why “more” is not always “better!”

 

Lavry tech

I would be interested to know what connections (if any) you have with this commercial product.

That statement sounds to me, like something that could have come from their Advertising Department.

 

I also feel sure that many other highly talented DAC designers are likely to have very different views in this area.

 

Regards

SandyK


How a Digital Audio file sounds, or a Digital Video file looks, is governed to a large extent by the Power Supply area. All that Identical Checksums gives is the possibility of REGENERATING the file to close to that of the original file.

PROFILE UPDATED 28-06-2020

Share this post


Link to post
Share on other sites
Find out why “more” is not always “better!”

 

Lavry tech

I would be interested to know what connections (if any) you have with this commercial product.

That statement sounds to me, like something that could have come from their Advertising Department.

 

I also feel sure that many other highly talented DAC designers are likely to have very different views in this area.

 

Regards

SandyK

 

Agreed 100%.

I wonder why folks can't give this one a rest.

 

It is too dern bad that the opportunity for commerce has led to many converters with "192" on their spec sheets, which actually perform *worse* at this rate than they do at say, 96k. This, I attribute to clocking that is not up to the significantly increased demands of the 4x rates and to analog stages that are not delivering good performance at wide bandwidth. So, we get folks who either don't hear a difference or for some other reason, miss the magic that the 4x rates (176.4k and 192k) can achieve when they're done right. Happily, the work of some of the best designers today is there to be heard by folks who are interested.

 

There are designers who are very good at doing the clocking section of their converters as well as the analog section so that both can very clearly demonstrate what I take to be *huge* advantages at 4x rates. As I've said elsewhere, to my ears, a very, very important threshold is crossed and with properly done 24/192 (or 176.4), the recording sounds - for the very first time in my experience - indistinguishable from the mic feed. This has never happened to me before with analog or with the very best 24/96 I've heard.

 

So, now that I've got a recording format that is truly transparent, I see "articles" (from designers of lossy formats) and "white papers" telling me that is isn't as "good" or is a waste of space or some other silliness. Might as well argue that there are no colors in a rainbow. I won't argue back. I'm too busy enjoying marvelous 176.4k recordings from Keith Johnson and making 192k recordings of my own.

 

Best regards,

Barry

http://www.soundkeeperrecordings.com

http://www.barrydiamentaudio.com

Share this post


Link to post
Share on other sites
Find out why “more” is not always “better!”

 

Lavry tech

I would be interested to know what connections (if any) you have with this commercial product.

That statement sounds to me, like something that could have come from their Advertising Department.

 

I also feel sure that many other highly talented DAC designers are likely to have very different views in this area.

 

Regards

SandyK

 

I would like to read an article promoting 192 as being the best sample rate. Could you provide one?

Share this post


Link to post
Share on other sites

Find out why “more” is not always “better!”

 

Lavry tech

I would be interested to know what connections (if any) you have with this commercial product.

That statement sounds to me, like something that could have come from their Advertising Department.

 

"Lavry tech" works for Lavry. He's posted in the forums before on behalf of Lavry and as his nickname indicates, is completely forthcoming about his identity and affiliation. So nothing at all untoward about this, he's simply announcing to the forum the availability of Lavry's most recent paper.

 

The paper contains a nice explanation of how, as long as one satisfies the Shannon-Nyquist conditions (sample at over 2x the highest frequency of interest), one may thereby specify two musical signals with an arbitrary time differential between them. That is, the time between samples doesn't have to be equal to or shorter than the time differential between the musical signals. The signals are completely specified as long as Shannon-Nyquist requirements are satisfied, and faster sampling rates won't specify the time differential more exactly.

 

There are two areas related to issues raised by the paper on which I'd like to comment.

 

- An aspect of timing not discussed by the paper is transient response where the initial inharmonic attack of an instrument has a rise time faster than the wavefront of a 22.05kHz sine wave. If the steepness of that initial inharmonic attack transient is something that our ears and brains react to (as distinct from a continuous harmonic tone at 22.05kHz, which I think we can agree is inaudible), then it raises the possibility that systems based on reproduction of the audible harmonic range may not be adequate. Academic papers on the subject by Kunchur conclude people can discriminate such transients at about half the rise time of a 22.05kHz harmonic wave, indicating double the CD sample rate may be necessary to reproduce them *in the mathematically ideal case with perfect filtering*. Since filtering in the real world is not perfect, this would tend to support the use of sampling rates substantially above 2x CD rate.

 

- The Lavry paper first says there is no reason to retain any information above the audible range, then notes such information can have a substantial effect within the audible range. These two statements, it seems to me, are contradictory. If real world instruments produce ultrasonic information (and they do) which, as the Lavry paper admits (in fact emphasizes), affects the audible sound, then removing the ultrasonics alters the original audible sound, i.e., introduces distortion. So in order to faithfully reproduce the audible range, including the audible effects of the ultrasonics produced by the instruments, the goal should be to correctly reproduce the ultrasonics as well. Particularly with the limitations of real-world filtering referenced in the previous paragraph, this would seem again to point toward sampling rates above 2x the CD rate.


One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> wi-fi to router -> EtherREGEN -> microRendu -> USPCB -> ISO Regen (powered by LPS-1) -> USPCB -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Share this post


Link to post
Share on other sites

" I would like to read an article promoting 192 as being the best sample rate. Could you provide one? "

Buffalobill

I take anything that comes from any manufacturer with a healthy dose of scepticism.

You could try Google though, for further information, although most of it is likely to be commercially related, and promoting their own "discoveries" and products. You are unlikely to find much peer reviewed material, as many from the "establishment" have a decidely anti high resolution agenda.( including a few C.A. members) That also applies to 24/88 and 24/96 !

I don't doubt that for the DAC designer, that 24/192 is harder to implement for best results than 24/96.

Nevertheless, on the few good 24/192 albums that I have, especially "Equinox" from Barry Diament of SoundKeeper Recordings,

I can still appreciate the improvement over the best 24/96 material from DVD-A etc. that I have, despite being 73 years old.

Barry's "Equinox" for example ,is extremely pure sounding with a complete lack of distortion and an excellent depth of image.

Alex


How a Digital Audio file sounds, or a Digital Video file looks, is governed to a large extent by the Power Supply area. All that Identical Checksums gives is the possibility of REGENERATING the file to close to that of the original file.

PROFILE UPDATED 28-06-2020

Share this post


Link to post
Share on other sites

I remember when experts proclaimed that 3.1 megapixels were all you'd ever need from a digital camera. Then it was 6 megapixels. Then 12 was the magic number.

 

Now we have the 36-megapixel Nikon D800 taking the industry by storm. Does anyone need 36 megapixels? Probably only those that print billboards and giant posters. But it's available, (relatively) affordable, and actually works. Why not?

 

As a part time sound engineer, I never considered recording above 24bit/96KHz. I figured everything will be burnt to 44.1KHz redbook CD. These days, we have to keep in mind hi-res distribution which the audiophile market demands. If I was going to distribute 96Khz, I would probably record at 192KHz so sample-accurate editing is not compromised. If I was archiving, then probably go double-rate DSD.

 

Today's computers and audio interfaces easily handle recording and playback of multiple 192KHz channels simulateously. Storage is cheap (getting back to normal after Thailand flood last year).

 

I guess same argument could be made with 32bit DACs...isn't 24 bits plenty? We're talking about S/N ratio that's beyond many measuring equipment, let alone our ears.

 

I would agree that technically 96Khz is plenty. But if the technology is here, affordable, and allows higher res, then go for it. We have bandwidth to burn.

 

Cheers,

JR


iMac/PM1.8, RME FF800+Metrum Octave, Accuphase DP-700, C-2410, A-45 x2, Gallo Reference 3.5, Vovox Sonorus

Share this post


Link to post
Share on other sites

I would agree that technically 96Khz is plenty. But if the technology is here, affordable, and allows higher res, then go for it. We have bandwidth to burn.

 

Cheers,

JR

 

Yeah, helluva reason to do it huh? Because we can. Yep, professional advice and experience, do the most intensive and expensive thing because we can. Heck, I don't figure it takes a lot of knowledge to make that decision. Anyone's goofy brother in law could have told you that. So, do we end up in ten year's time doing everything at 1536 khz/48 bit just because? Hey, congratulations.

 

And yes, the super megapixel thing is comparable. I mean a well done with best current tech 6 mp sensor would have excellent low light capability. But 18 mp or more with still the same low light issues as always will still sell because educated consumers, nay connoisseurs, know 18 is more than 6. Again congratulations, such input and reasoning is so hard to come by. More is more and less is less.


And always keep in mind: Cognitive biases, like seeing optical illusions are a sign of a normally functioning brain. We all have them, it’s nothing to be ashamed about, but it is something that affects our objective evaluation of reality. 

Share this post


Link to post
Share on other sites

Nope, not because we can. Because there is a market for hi res audio. Not advocating 192Khz, no professional advice or experience offered...JMHO.

 

Ironically, the 36Mp D800 outperforms just about every other 3,6,10,12,16,21,24Mp cameras in low light/high ISO capabiity (except for the low light King Nikon D3s). Such is technology.


iMac/PM1.8, RME FF800+Metrum Octave, Accuphase DP-700, C-2410, A-45 x2, Gallo Reference 3.5, Vovox Sonorus

Share this post


Link to post
Share on other sites
Yeah, helluva reason to do it huh? Because we can.

 

Just like we need a 6-litre V8 in our car... :)

Share this post


Link to post
Share on other sites
Interested in the facts?

 

One of the world’s top converter designers Dan Lavry has written a new paper in simple language to demystify the subject.

 

http://www.lavryengineering.com/pdfs/lavry-white-paper-the_optimal_sample_rate_for_quality_audio.pdf

 

See why many professional engineers still work at 96kHz years after 192kHz became available.

 

Find out why “more” is not always “better!”

 

 

Hi Lavry Tech - While this paper may be interesting and very valuable from an engineering standpoint, your surrounding statements really hurt your credibility. If you post here on CA in an effort to educate there is no need to tout Dan as "One of the world’s top converter designers..." or to begin your post with "Interested in the facts?"

 

I can find several engineers and AES Fellows who contradict much of what Dan says. My point is there's not one set of facts.


Founder of Audiophile Style

Announcing Polestar | Quick Community Reviews and Ratings

Share this post


Link to post
Share on other sites
Academic papers on the subject by Kunchur conclude people can discriminate such transients at about half the rise time of a 22.05kHz harmonic wave, indicating double the CD sample rate may be necessary to reproduce them *in the mathematically ideal case with perfect filtering*.

 

A fair amount of academic researchers have expressed concerns about Kunchur's work. There has been a number of attempts at replicating the results, and they have concluded that *if you remove the intermodulation distortion that causes the ultrasound to produce intermodulation components in the audible spectrum*, there are no audible difference if the ultrasonic component is removed.

 

- The Lavry paper first says there is no reason to retain any information above the audible range, then notes such information can have a substantial effect within the audible range. These two statements, it seems to me, are contradictory. If real world instruments produce ultrasonic information (and they do) which, as the Lavry paper admits (in fact emphasizes), affects the audible sound, then removing the ultrasonics alters the original audible sound, i.e., introduces distortion.

 

It might be the other way around - the ultrasonic information might cause intermodulation distortion that is audible in the sonic frequency range.

Share this post


Link to post
Share on other sites

Chris,

 

I can find several engineers and AES Fellows who contradict much of what Dan says. My point is there's not one set of facts.

 

What part of what Dan says do you disagree with? The thing about the timing resolution is pretty much basic sampling theory and is well supported by people working in the field (a good example is the Meridian white paper by J. Robert Stuart often quoted here on CA). For basic, undithered signals, the time resolution of a sampled signal is the sample interval divided by the number of digital levels. Dithering improves it further. Thus the time resolution of a 44.1kHz / 16 bit signal is 1/(44100*2^16) s, or 0.34 ns.

Share this post


Link to post
Share on other sites
Chris,

 

 

 

What part of what Dan says do you disagree with? The thing about the timing resolution is pretty much basic sampling theory and is well supported by people working in the field (a good example is the Meridian white paper by J. Robert Stuart often quoted here on CA). For basic, undithered signals, the time resolution of a sampled signal is the sample interval divided by the number of digital levels. Dithering improves it further. Thus the time resolution of a 44.1kHz / 16 bit signal is 1/(44100*2^16) s, or 0.34 ns.

 

 

Hi Julf - My point is simply the original poster writes as if Dan's word is the only word on the subject. I think Keith Johnson has contradictory things to say about the value of 176.4 kHz etc... Much of this is beyond my level of expertise. I'm just suggesting Dan isn't the Minister of Information when it comes to digital. In fact nobody is.


Founder of Audiophile Style

Announcing Polestar | Quick Community Reviews and Ratings

Share this post


Link to post
Share on other sites
My point is simply the original poster writes as if Dan's word is the only word on the subject.

 

I am definitely with you on that part - hyperbole never helps.

Share this post


Link to post
Share on other sites
Hi Julf - My point is simply the original poster writes as if Dan's word is the only word on the subject. I think Keith Johnson has contradictory things to say about the value of 176.4 kHz etc... Much of this is beyond my level of expertise. I'm just suggesting Dan isn't the Minister of Information when it comes to digital. In fact nobody is.

 

Well, lets not forget Keith Johnson also has a commercial interest to push. Why pay extra for his hi-rez recordings unless they are better right? Plus in more guarded comments at times he appears to think even his hi-rez is only a little bit better than properly done CD. That being the case, seems the difference between 96 khz and 192 khz might be small indeed.

 

As you say, no one is the Minister of Information on digital.


And always keep in mind: Cognitive biases, like seeing optical illusions are a sign of a normally functioning brain. We all have them, it’s nothing to be ashamed about, but it is something that affects our objective evaluation of reality. 

Share this post


Link to post
Share on other sites
I remember when experts proclaimed that 3.1 megapixels were all you'd ever need from a digital camera. Then it was 6 megapixels. Then 12 was the magic number.

 

Now we have the 36-megapixel Nikon D800 taking the industry by storm. Does anyone need 36 megapixels? Probably only those that print billboards and giant posters. But it's available, (relatively) affordable, and actually works. Why not?

 

JR

 

Just as an aside, the high resolution of the Nikon allows you to essentially just use one lens to replace two or more. The high resolution allows you to zoom in and enlarge to such detail, who needs a zoom lens.. You still need to observe basic photographic skills, but it opens the door to a new level we could not see or go to before. So hires would take us where we haven't heard before, 96/24 gets us to the door, 192/24 opens it, 768/24 gives us electrons to hear? :)


AS Profile Equipment List        Say NO to MQA

Share this post


Link to post
Share on other sites
A fair amount of academic researchers have expressed concerns about Kunchur's work. There has been a number of attempts at replicating the results, and they have concluded that *if you remove the intermodulation distortion that causes the ultrasound to produce intermodulation components in the audible spectrum*, there are no audible difference if the ultrasonic component is removed.

 

I'd appreciate references if you have them handy. Anything that helps me better understand the issues is always welcome.

 

It might be the other way around - the ultrasonic information might cause intermodulation distortion that is audible in the sonic frequency range.

 

Yes, absolutely. My point was that the electrical impulse representing the ultrasonic and audible range waves cannot "know" whether their intermodulation creates distortion or a correct representation of an original event in which both ultrasonics and audible range waves were involved. Once one grants that ultrasonics affect audible frequencies, one cannot then simply declare that such interaction is always distortion and by declaring make it so.

 

Granting the premise that ultrasonics affect audible frequencies, one has two choices if there are ultrasonics in the original event: Keep the ultrasonics or attempt to remove them. If one keeps the ultrasonics, this is an argument for greater than 2x CD samplng rates in order to get to the Shannon-Nyquist frequency and then leave sufficient headroom for non "brick wall" filtering. If one attempts to remove the ultrasonics (which to my mind is in itself distortion, since one is removing something that was present in the original), either one uses "brick wall" filtering to try to remove them completely, which creates distortion; or one uses a more gently sloped filter, which leaves altered ultrasonics that will intermodulate with audible frequencies differently than the original, and which also may necessitate higher than 2x CD sampling rates. Thus in order to reproduce the original event properly, it seems best to try to reproduce the ultrasonics correctly.


One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> wi-fi to router -> EtherREGEN -> microRendu -> USPCB -> ISO Regen (powered by LPS-1) -> USPCB -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Share this post


Link to post
Share on other sites

Hi elsdude,

 

...Plus in more guarded comments at times he appears to think even his hi-rez is only a little bit better than properly done CD. That being the case, seems the difference between 96 khz and 192 khz might be small indeed...

 

I'm not so sure about this. In fact, I'm pretty sure it is mistaken. I remember reading (though I wish I remember where) with much interest, an interview with Keith where he used a phrase very similar to the one I've used since I first experimented with 192k done right: he said a threshold is crossed with the 4x rates (i.e. 176.4 and 192). A threshold being crossed is exactly what I've said many times about my preferred 192k sample rate - with converters that have commensurate clocking and analog stage performance to do it - which in my experience to date, is far, far fewer than the number of converters with the number "192" on their spec sheet.

 

I doubt very strongly that someone who thinks the 4x rates are "only a little bit better" than 2x rates, much less "properly done CD" would refer to a threshold being crossed with the 4x rates.

 

To my ears, properly done 192 (or 176.4), precisely because of that threshold being crossed, is a significantly larger leap up from 96k (or 88.2) than the latter are from CD.

 

It isn't about "a commercial interest to push". It is about a sincere desire to see just how good it can get - and having found how much better it can be. When I visit Reference Recordings' site, nothing is forcing me to purchase the 4x versions of Keith's recordings. The CDs are quite easily available. With some selections, I've gotten both: the CD shows how good CD can be, the 4x version shows how very much more of the performance (and the space in which it occurred) there is to be heard in the original recording.

 

I would agree that no one is the Minister of Information when it comes to digital - or anything else, particularly in audio. But as I see and hear it, folks like Keith Johnson and B.J. Buchalter might be members of the Ministry's *Committee* on Information. ;-}

 

Best regards,

Barry

http://www.soundkeeperrecordings.com

http://www.barrydiamentaudio.com

Share this post


Link to post
Share on other sites
What part of what Dan says do you disagree with? The thing about the timing resolution is pretty much basic sampling theory and is well supported by people working in the field

 

Due to filter ringing features, I've defined optimal sampling rate such where:

- you can fit anti-alias/imaging filter's impulse response into half-wave of 20 kHz sine

- frequency response in pass-band must be within 0.01 dB

- phase reponse in pass-band must be within 1 degree

- final re-constructed analog signal must also have all alias/images attenuated at least by 144 dBFS


Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Share this post


Link to post
Share on other sites
Due to filter ringing features, I've defined optimal sampling rate such where:

- you can fit anti-alias/imaging filter's impulse response into half-wave of 20 kHz sine

- frequency response in pass-band must be within 0.01 dB

- phase reponse in pass-band must be within 1 degree

- final re-constructed analog signal must also have all alias/images attenuated at least by 144 dBFS

 

So what figures does that give us in practice?


System (i): (Stack Audio Link/MoOde > 2Qute+MCRU psu; Gyrodec/SME V/Ortofon 2M Black/EAT E-Glo Petit/Magnum Dynalab FT101A) > Glow Amp One > Klipsch RP-600M

System (ii): iUSB3.0 Nano/Allo USB Signature/MoOde > Bel Canto uLink+AQVOX psu > Chord Hugo > (Tandy LX5; JBL LSR305 ; Audeze LCD-3)

Share this post


Link to post
Share on other sites
So what figures does that give us in practice?

 

Of course it depends on specifics of particular design. But I've concluded that I can reach it in practice at 352.8 ~ 384 kHz.


Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Share this post


Link to post
Share on other sites
I'd appreciate references if you have them handy.

 

Griesinger, Perception of mid-frequency and high-frequency intermodulation distortion in loudspeakers, and its relationship to high definition audio (http://www.davidgriesinger.com/intermod.ppt)

 

Yes, absolutely. My point was that the electrical impulse representing the ultrasonic and audible range waves cannot "know" whether their intermodulation creates distortion or a correct representation of an original event in which both ultrasonics and audible range waves were involved. Once one grants that ultrasonics affect audible frequencies, one cannot then simply declare that such interaction is always distortion and by declaring make it so.

 

The Griesinger experiment shows that the intermodulation distortion was only audible when driving the amplifier hard.

Share this post


Link to post
Share on other sites

 

Julf, I don't understand what you intend by this.

 

You said several academic researchers had expressed concern about Kunchur's work and were unable to reproduce that work. I was very interested to hear this, since I'd looked for such papers and been unable to find any, so I asked you to provide them. In return, you cite a single presentation by a non-academic (though a highly qualified person in certain areas of audio, mainly regarding reverberant vs. direct sound) that expresses no concerns whatever about Kunchur's work, one very good reason for which is because the presentation predates Kunchur's first paper on this subject by four years.

 

Furthermore, you stated the following with regard to the reason why "other researchers" were supposedly unable to duplicate Kunchur's work:

 

A fair amount of academic researchers have expressed concerns about Kunchur's work. There has been a number of attempts at replicating the results, and they have concluded that *if you remove the intermodulation distortion that causes the ultrasound to produce intermodulation components in the audible spectrum*, there are no audible difference if the ultrasonic component is removed.

 

But Kunchur's work on detecting time differences does not have to do with audibility of ultrasonic tones or their intermodulation products, so this cannot be a correct statement. The part of your statement above emphasized by asterisks is not a correct description of Kunchur's work but rather regards an experiment Griesinger ran for himself and reported in his 2003 presentation. In fact Griesinger is careful to state at slide 29 and again at slide 35 that audibility of time differences is an entirely distinct thing from the ultrasonic information he is concerned with, pointing out very small timing differences can be detected well within the audible band. This agrees completely with Kunchur 2007, where his experiments on timing differences were carried out using tones within the audible range.

 

So I am wondering if you have an explanation as to why you gave me this in response to my request for you to provide the references you claimed to have regarding concerns about Kunchur's work and repeatability of his experiments?


One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> wi-fi to router -> EtherREGEN -> microRendu -> USPCB -> ISO Regen (powered by LPS-1) -> USPCB -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  



×
×
  • Create New...