Why 24/192 is a bad idea?
- Thread starter audioguy
- Start date
I took a very quick skim. That's a pretty poor article from a technical standpoint. They misrepresent some facets of sampling and reconstruction, leave out other things, and slant other things to reinforce their conclusions. Since much of it is covered elsewhere ad nauseum and I have too much else to do I'll just say read up on sampling and such before reading that article. Our tech forum has some starter articles (threads).
I don't know about debunking it but here is a thread with more than you ever wanted to read: http://www.avsforum.com/avs-vb/showthread.php?t=1398397
And if interested, here is a paper by Julian Dunn which gives very solid research & findings for the possible reasons why a higher sampling rate than 44.1KHz should sound better.
"Anti-alias and anti-image filtering:
The benefits of 96kHz sampling rate formats for those who
cannot hear above 20kHz"
I don't know if there is any improvement beyond this sample rate?
Apologies if this has been referenced before!
"Anti-alias and anti-image filtering:
The benefits of 96kHz sampling rate formats for those who
cannot hear above 20kHz"
I don't know if there is any improvement beyond this sample rate?
Apologies if this has been referenced before!
Julian's paper is old (14 years!) as he passed away many years ago sadly. There has been much discovery with hirez digital since then. I record classical music at 24/176 and 24/192 and the improvements over 24/96, tested over many sessions and gear, is noticeable.
Julian once wrote a great paper suggesting that human hearing could hear at least nanosecond timing differences (jitter)...since that paper most scientists have found that single digit jitter is not only possible but required for the best possible sound.
Julian once wrote a great paper suggesting that human hearing could hear at least nanosecond timing differences (jitter)...since that paper most scientists have found that single digit jitter is not only possible but required for the best possible sound.
I agree it's old but still relevant I believe. Any references for the later high-res discoveries of which you speak?
I'm not disputing your reports & experience in using high-res recording but any ideas why higher res material sounds better?
High-res, in fact, can be shown to have a higher susceptibility to jitter!
Yeah, I guess. There are a lot of good reasons for higher sampling rates, but then again I accidentally switched the power cords on my BD player and AVR. Didn't notice a difference. Ears of clay. I have to go practice...
I don't know about debunking it but here is a thread with more than you ever wanted to read: http://www.avsforum.com/avs-vb/showthread.php?t=1398397
Strange thread to link to for a debunking. Not quite what happened there. It would probably be useful to know exactly who Chris Montgomery (Monty) is, and what his field is, to a greater extent than "someone who wrote a codec."
Some of his projects:
Ogg - a multimedia container format, a reference implementation, and the native file and stream format for the Xiph.org multimedia codecs
Vorbis - a lossy audio compression format and software
Theora - a lossy video compression format and software
FLAC - a lossless audio compression format and software
Speex - a lossy speech compression format and software
CELT - an ultra-low delay lossy audio compression format and software
Tremor - an integer-only implementation of the Vorbis audio decoder for embedded devices (software)
OggPCM - an encapsulation of PCM audio data inside the Ogg container format
OggUVS - an uncompressed video codec for Ogg (a work in progress and not a final proposal)
Skeleton[18] - a structuring information for multitrack Ogg files (a logical bitstream within an Ogg stream)
RTP-containers for Vorbis, Theora, Speex
CMML - an XML-based markup language for time-continuous data (a timed text codec)
Ogg Writ - a text phrase codec (no longer maintained)
Ogg Squish - a lossless audio compression format and software (discontinued)
Tarkin - an experimental lossy video compression; no stable release (discontinued)[19]
libao, an audio-output library that operates on different platforms[20]
OpenCodecs - a codec pack for Windows, by Xiph and the WebM Project
Annodex - an encapsulation format, which interleaves time-continuous data with CMML markup in a streamable manner
Icecast - an open source multi-platform streaming server (software)
Ices - a source client for broadcasting in Ogg Vorbis or MP3 format to an icecast2 server (software)
IceShare - an unfinished peercasting system for Ogg multimedia (no longer maintained)
I didn't say it was a debunking thread. It was however a deep discussion about his then version of the article. I see that he has embellished it a ton more, no doubt to plug the deficiencies in the first version.
As to who he is, I know him exceptionally well, perhaps longer and earlier than most people, dating back to shortly when ogg had come out (some 20 years ago). I also know that he is a good salesman and doesn't mind pushing the line of what the reality is. When he first came out with Ogg, he used to claim it was a superior codec to others. What he wouldn't say is that the default encoding mode was variable rate encoding, not fixed. So people would use it out of box that way, comparing it with other codecs at fixed data rate. And of course, his would sound better because it would be able to increase its bit rate as needed while the others were capped. The other codecs had variable mode also but people would not know to use them. I went on the forum where he hung out and try as I might, I could not get him to accept this unfair set of comparisons. He also had little understanding of patents -- at least at the time. He had this simplistic idea that if he had written something, and put it in open source, it could not be covered by anyone else's patents.
In this case, as I noted, he wrote what can at best be described as a much poorer and less authoritative version of Bob Stuart's AES paper as I explained in that thread. Unlike Bob's paper, his references are often forum threads, and anecdotal. The whole thing sounds a lot more technical than it really is. Much of the knowledge that is needed to describe this field requires hardware and hands on design with things like DACs. Writing software and compression algorithms gives you little to no expertise for that. Bob on the other hand, does have this expertise and industry credentials to do the topic justice. Hence the reason if you want to see the proper take on this, you want to read his version.
BTW, Monty did not create FLAC. His organization adopted it after it was developed: http://en.wikipedia.org/wiki/FLAC.
One thing he seriously missed (also see last sentence below) regarding the fallacy of the rough stair step and I think it is important as it comes up a lot in arguments and this side of the technical is never mentioned.
In the past and closer to the origins of PCM DAC the stairstep would had been the output from a DAC as the reconstruction filters would had been seperate, however with modern chips most of the functions are now embedded/integrated onto one chip with the DAC.
So in practical terms a DAC does output in stairstep, even though we never notice this due to the integrated reconstruction filters function that now operate before we can measure (although a fair few have bypass capability for extrernal bespoke-custom developed algorithms), academic really but thought I would mention this as there are many arguments about output of a DAC over at the other forum without anyone ever mentioning this.
Much better to read Don's stuff on this forum than that link IMO as that provided link was not good about mentioning nyquist and aliasing-image resolution relating to frequency rate, amongst other things.
Cheers
Orb
In the past and closer to the origins of PCM DAC the stairstep would had been the output from a DAC as the reconstruction filters would had been seperate, however with modern chips most of the functions are now embedded/integrated onto one chip with the DAC.
So in practical terms a DAC does output in stairstep, even though we never notice this due to the integrated reconstruction filters function that now operate before we can measure (although a fair few have bypass capability for extrernal bespoke-custom developed algorithms), academic really but thought I would mention this as there are many arguments about output of a DAC over at the other forum without anyone ever mentioning this.
Much better to read Don's stuff on this forum than that link IMO as that provided link was not good about mentioning nyquist and aliasing-image resolution relating to frequency rate, amongst other things.
Cheers
Orb
For oodles of information about "the stair step" check out the link Amir supplied above, where various debunkalunk happened.
@amir
Where do I write that Monty "created FLAC?" He has been quite instrumental in developing it to the standard we now enjoy.
The point is that he does know whereof he speaks, and you may try to denigrate him by your claim of him being "a salesman." But there are lots of those in audiophool fora, many of them constantly vigilant of their turf, apparently.
You listed it as "some of his projects." The implied that he developed it.
Honestly, he really doesn't know what he needs to know about this area. I used to live and breath the space he comes from. It used to be my full-time career to manage such including lots of hands on work. If this were a discussion of lossy codecs, then he would be an authoritative figure. But when it comes to fidelity of capturing and reproducting analog data into and out of digital domain, he simply has no expertise -- professionally or from academia. Sure, he knows enough to describe things for people who don't know anything at all. But nothing that rises to proper level. It is people like Julian Dunn, Professor Hawksford, Bob Stuart, etc. that bring true insight there and deep, deep level of knowledge. This is abundantly clear as we compare his paper and that of the others who have been published at major industry conferences and journals.
And let's be clear, despite my feelings about him, I post that I generally agree with his stance. I simply suggest that if you want to read about this, don't go by what he has written. Go by real experts that give you the proper analysis and explanation.
Perhaps "debunking" was a poor choice of phrasing; maybe "authors biases discussed" and "technical inaccuracies detailed" would have been better? Nevertheless, I think many discussion forums established that "24/192 downloads" are not harmful, nor are they a "solution to a problem that doesn't exist". LP's potentially have the same sort of ultrasonic information as 24/192 digital, but in neither case does it appear to result in amplifier instability or poorer sound.
Thanks, FYI I was trained and worked with IBM and also a large network carrier on this very specific subject, looong time ago so pretty rusty these days alas.
My point actually was specific to that site and another, thankfully would never join AVSF but read for a long time (wears me out tbh), however I have been a member of HA for many years.
Stair step waveform is the output of an audio PCM DAC, but what I want to emphasise is that many confuse this output with the complete reconstruction filter output, which is incorporated on many modern DAC chips.
As I mentioned it is academic anyway as we do not listen to this, but is a source of a lot of technical semantic arguments I notice on those forums.
I think 24/192 or anything close is a good idea as it pushes further anything relating to the digital filters-algorithms further out from the audio band, however I appreciate some have the POV that this cannot be heard even when it is just above 22khz.
Another benefit in theory is that a native 24/192 file could work well with a NOS DAC, and some modern NOS DACs are meant to support this or will, be interesting when some here get a chance to audition these.
All IMO so as usual take with a pinch of salt I say
Cheers
Orb
Last edited:
What are the problems that are associated with 24/192 compared to say 24/96? I would like a simple answer from those who know more.. I tend to like the idea but know its implementation be more difficult than it seems to be worth to me.
Dan Lavry has a fair bit to say on this - distilled its really that there's always a trade-off between accurate conversion and fast conversion. So all things being equal, there does need to be a sound engineering reason for going beyond 96k. I can't see one myself, despite having trawled around for several years.
One aspect that I've rarely seen discussed is the DAC side of things. Lavry talks (see here as one example: http://www.computeraudiophile.com/threads/10776-Lavry-Engineering-Paper-on-Hi-Res?p=133063&viewfull=1#post133063 about the accuracy of recording (i.e. ADCs)) but what he says applies equally well to DACs too. Faster is more difficult to get right.
An interesting exercise I once did was compare the advances in multibit DAC chips over time - it was a side-by-side comparison between PCM63-K and PCM1704-K. What I found was that although the PCM1704's headline (read 0dBFS) figure was better, it actually got 4dB worse than the PCM63 where it really matters, which is down at -20dBFS. I wondered if one reason for this was the higher sample rate its measured at (768k vs 384k).
I have a suspicion (by no means yet supported by hard evidence) that one reason NOS DACs get such good subjective reviews is down to them working less hard and hence giving more accurate results in practice.
One aspect that I've rarely seen discussed is the DAC side of things. Lavry talks (see here as one example: http://www.computeraudiophile.com/threads/10776-Lavry-Engineering-Paper-on-Hi-Res?p=133063&viewfull=1#post133063 about the accuracy of recording (i.e. ADCs)) but what he says applies equally well to DACs too. Faster is more difficult to get right.
An interesting exercise I once did was compare the advances in multibit DAC chips over time - it was a side-by-side comparison between PCM63-K and PCM1704-K. What I found was that although the PCM1704's headline (read 0dBFS) figure was better, it actually got 4dB worse than the PCM63 where it really matters, which is down at -20dBFS. I wondered if one reason for this was the higher sample rate its measured at (768k vs 384k).
I have a suspicion (by no means yet supported by hard evidence) that one reason NOS DACs get such good subjective reviews is down to them working less hard and hence giving more accurate results in practice.
Similar threads
 | Steve Williams Site Founder | Site Owner | Administrator |  | Ron Resnick Site Owner | Administrator |  | Julian (The Fixer) Website Build | Marketing Managersing |