A very interesting article at Digit-Life.com:

http://www.digit-life.com/articles2/sacd-dvd-a/index.html

Thanks. It was an interesting read. Especially the part about SACDs being currently mastered from a PCM source which makes the DSD (direct stream digital) format not very direct as opposed to the actual idea behind the DSD techonolgy.

In the end it's all the same. The conversion for the 1bit DACs from processed PCM material is done for SACDs before mastering and for DVD-As while playback .

Sebastian

Smells like BS

Yep, what about using the CD format properly before moving on to 'better' formats

WTF....Acheive maximum dynamic range??? They're removing all dynamic range, and regardless, this isn't a shortcoming of the format, just the engineers/masterers.

In my opinion, CDA still very usable and the main reason behind the industry pushing these new solutions, is because they offer better support to copy control and such.
Can you guys imagine in the near future, DVD-A disks having not only copy protection, but region control too(like DVD video), making any album bought overseas not playable at all?

Kind of ironic, when you realize that a number of industry heavyweights' first exposure to the music that ultimately hooked them was imported from some country other than their own....

- M.

I have no intrests of abandon normal CD until there is format what I can copy. I have so many original CD's lost or badly scrached because I had not CD-RW drive those days.

An analogue, or designed-to-match-analogue digital filter will have phase issues within the pass band - that's certainly true. However...




Rubbish! For one thing, on a digital system we can cut everything above (say) 20kHz by (say) 140dB while introducing exactly zero phase shift at frequencies below 20kHz. Also, you won't find a phono cartridge with anything like this kind of phase accuracy.


I found much of this article difficult to read. Half the time, it's not entirely clear what the author is trying to say. Though one of the main themes seems to be that many SACDs and DVD-As are mastered from ~20kHz band limited sources - that wouldn't surprise me - most of the normal audio world still uses 24-bit 48kHz.




Rubbish - there are three options:
1. Do everything in analogue except the final DSD encoding
2. Process in DSD wide - this is 8-bit, not 1-bit, hence it's just high sample rate PCM with a lot of ultrasonic noise.
3. Process in DSD - every operation generates a multi-bit signal which had to be dithered/noise-shaped back to 1-bit. This isn't liked because the ultrasonic noise accumulates with every operation.

Cheers,
David.

I'm more than happy with the quality of a well produced and mastered audio CD and see no reason to change.

It's the people who are using the CD format properly (not me) who want this new stuff. I consider these future technologies in any case - with the scarcity of releases in these formats. Will they take off before discs are made obsolete as Gates predicts? I don't know.

I agree with LIF. I won't be buying the new formats. I want my music media to be flexible. I cannot backup SACD/DVD-A. Also, I want to listen my music where/whenever I want to, not stuck in front of a home theater system. Copy protection is what interests the recording companies, not sound quality. Do you guys remember pre-recorded MDs?

NB that the author of the webpage runs a component modification business; which gives hima potential stake in potentially pointless tweaks; he does however post regularly and informatively on AVSforum.com forums (under the moniker 'Gordon MacGregor' for some reason).
I suspect the article was translated from Russian, hence the slight obscurity of some of the language.

Seems to me he glided too glibly over the actual audibility of all this >20 kHz information, which is a foundational issue for the rest of his claims.

I buy SACDs and DVDas to 1) hear a different mastering, often *claimed* to have been from better sources/done more carefully and 2) for multichannel mixes. I don't put much stock in claims that the playback higher resolution or bitdepth makes much difference.

You might be interested to know that the engineers at Deutsche Grammaphon, the big classical label, considered both PCM and DSD for their digital archiving, and decided to go with PCM, based (they say)
on their comparative technical merits. So all their SACDs are sourced from 'hi rez' PCM transfers/remixes of their original master tapes.

Hello everybody!

Interesting how my business is related to the article and why everyone wants to discuss it? It is not pointless if you will visit DIYAUDIO forum and ask people if there is any difference in op amps sounding ...

Yes, I do, and I used the same pseudonym Gordon McGregor in all forums around the world. The reason is - I used it online till 7 years ... starting from Ultima Online, as I had no idea to publish articles and confuse people ...
My friend is from Clan McGregor ...

You hit the nail on the head! Digit-life is owned by the best Russian hardware website - iXBT.com and they have translated my article, published in Russian three weeks ago. I don't like the translation and intend to make my own.

OK, here is the scientific confirmation, that we can recognize hypersonic images, not separated from the audible signal:
http://jn.physiology.org/cgi/content/abstract/83/6/3548 Full article is in pdf file at this link.

it really depends on the equipment used. From analog output of 47Ai I also don't hear a huge difference between the Hi-Res formats and CDs ... completely different story in multi-channel mode through i-link ...

Concerning CDs. We definitely don't have phase shift using FIR Brickwall, but I was talking in the article about analog anti-aliasing filters. I also stated clearly, that correctly mastered and played CDs can sound better, then non-correctly played Hi-Res formats. For example - CDs, created using perfect dithering algorithms as ExtraBit, A/D conversion at the high sampling rate and resolution, processing in the good professional software using FIR cut-off, etc. can sound very good, using the good CD player.
I had auditioned such systems - for example CD in dCS CD->DSD upsampling mode was sounding almost perfectly ... http://members.cox.net/alex_lat/images/dCS.jpg
But we have to take in mind that there are not so many perfectly mastered CDs available, even in our days some modern CDs are badly clipped - take a look at Diana Krall 2001 album, even the first song has 25 possibly clipped samples in one channel and 110 in another one ... How can we find out which one is good and which one is not good? Check them all? I prefer to use High-Res formats, they have no such problem and sound significantly better anyway ... http://members.cox.net/alex_lat/Tests/Statistics_DK.JPG it had been ripped by EAC to HDD ... here is the zoomed version:
http://members.cox.net/alex_lat/Tests/DK_CD_clipped.PNG
I am open to discussions and have nothing against being corrected by the knowledgeable people ... we all are studying ...

I may do that, but I can tell you in advance that anecdotal , sighted reports of difference from 'satisfied customers' hold no weight with me -- confirmation bias is simply too strong a possibility.
(And note that I wrote 'potential' both times ;>)



I still like my 'superhero' theory, though. ;>



And you should look up the discussions of Oohashi's work, here, and on other audio forums. For starters, it's work that used 'home made' equipment, hasn't ever been replicated, and has some *counterevidence* from other research, going against it. Nor does it really show that people *hear* ultrasonics.



Which only restates that *implementation matters*. The more fundamental question, though, is whether the formats are intrinsically different-sounding. If not then it's *all* about implementation, and there really is no need for SACD or DVD-A *playback*, except for multichannel. There is only a need for *properly done* CD.

The "high-res formats" are definetely susceptible to clipping, just like CD Audio is. If the behaviour of the recording industry in the past is any indicator then we will get badly clipped DVDA/SACD as soon as the formats are a bit more whidespread. It's all about being the loudest, "sound quality" is just something to lure early adopters with.

As far as I know this study is still considered highly controversial. I don't think you can convincingly push it as "scientific confirmation" until the results are confirmed by other research.

Guys I really don't like this forum services. I wrote the reply, being logged in, tried to send it, it says "you don't have the permisson" (it means it logged me out after some time?) and then the post disapeared ... Need to re-write everything again and it is at least 30 min ... grrr

I've been burned by that too! I now write every post longer than one paragraph in WordPad and copy+paste it in.

1st - this test did not confirm adubility -- the 'audibility' portion of this test has not much information published and is questionably valid 2nd - this test focused primarily on brain scan activity 3rd - the test could not produce any result of any type when only ultrasonic information was presented. -- only when both ultrasonic and sonic data was presented. 4th - NHK labs to to task this study, and produced a careful perceptual test:

Perceptual Discrimination between Musical Sounds with and without Very High Frequency Components
Toshiyuki Nishiguchi, Kimio Hamasaki, Masakazu Iwaki, and Akio Ando
AES Preprint: 5876

Since probably don't want to dish out $$ for the paper, it is lucky this is one of the few AES papers availabe online for free, since it was also published on the NHK webstie:

http://www.nhk.or.jp/strl/publica/labnote/lab486.html



Their is no evidence that properly contstructed anti-alias filter, even analogue, will have audible consequences. Refer to the landmark perceptual paper, that used such steep filters:

What Bandwidth Is Nescarry For Optimal Sound Transmision
Plenge, Jakubowski and Schone
JAES, 1980, Volume 26, Number 3, Pages 114-119

I would also refer you to a later paper on the audibility of anti-alias filters:

Perception of Phase In Anti-Alias Filters
Pries and Bloom
AES Preprint: 2008

-Chris

Thank you for your inputs, guys!

We have a lot of opinions of well known respected people in audio, for example Doug Self's ... and many others, the whole tweaking industry is based on this principle. "Potentially" is a nice word, potentially I hate you ... what does it mean? Please note that I have no reason to hate you in reality nor any intention to abuse you, just used your word
I tried to make a suggestion about the difference, it is here - has to replace the old measurements with my Audigy with the new ones through Lynx L22 though.
http://members.cox.net/alexhardware/opa.htm
In the DBT I can recognize the difference between PCM1704+OPA627+OPA2134 and PCM1704+JRC5534+JRC2068+JRC2068 with 17/19 probability in the best case, usually it is 14-15/19 ... You still don't think it is enough information to make the suggestion? Well, this is a point to write another article, it is "on the way"

I don't think that I am any kind of "superhero", though potentially you know better

I know that anything can be blamed - bad equipment, bad sources of information, etc., etc. Please supply those opposite sources, we can discuss them as well. I am not a scientist, but I hear the difference using DBT as stated in the article, with 19/19 probability, and it looks as the samples are different only in the ultrasonic area ... I also see the ultrasonic images, and I have no other idea how to explain the difference in sounding through the same path ... there is no difference in implementation, as for PCM exactly the same path is used.

In the same best conditions High-Res format is a winner (I am talking about presented in the article DVD-A Advanced resolution and SACD multi-channel), in other case they are simply useless (and we know that this is not true). I can hear the difference through my system easily - as stated before, the same path for PCM audio signals. Not only me - as the moderator of this forum pio2001 said, and I agree with him:

To confirm this statement I am going to make an experiment - using Adobe Audition 1.5, record the sample of High-Res audio signal in 24/96 mode, then convert it to 16/44.1 using two algorithms of resampling - SSRC or PPHS plus different kinds of dithering, including ExtraBit and compare the results through DBT. The files will be provided for downloading. I don't have any other idea how to prove it, may be you can recommend something? I know about this website http://64.41.69.21/technical/sample_rates/index.htm , but they did't use ExtraBit and different re-samplers AFAIK, though their samples could be useful for expariments.
Concerning IMD, I use M&K S-150 speakers, which were designed with lowering of IMD in mind, as stated at M&K website, using two tone (means standard) IMD measurements. My measurements of 49TXi audio signal path including A/D stage of Lynx L22 path also shows quite low level of IMD:
IMD, %: 0.0073 Excellent
I can't be 100% sure, but can make some suggestion that the level of IMD in my DBT is quite low. It is not a scientifical provement though, I need AP and a great mic to confirm that. Please note, that the article is not scientific, if you have any scientific information to show me where I was wrong, I will be glad to correct it.

I never met even 1 SACD or DVD-A with clipping. I will gladly buy the disk with such problem, if you can recommend me any. As as we can see, 2001 CD is not relative to "early adopters" ...

First of all Nelson Pass's statement in the article is based on Ph.D. Steven Dear researches in the same area, so we have some "other researches".
Highly controversial? It is published in the well-respected Journal of Neuropsychology, 10 people are co-authors in it, it is "scientifical enough" for me. Can you provide the links with the opposite information? If you are talking about this document http://www.nhk.or.jp/strl/publica/labnote/lab486.html I have some points to discuss. I did DBT and easily recognized that it is necessary to know what difference are you going to hear and train yourself for a while, then it becomes completely different. If you use just some participants who know nothing about the test, the probability that they can hear the difference is very low when signals are more or less equal. A very important point, which never had been discussed or researched - ultrasonic images in multi-channel audio, we have a significant difference in perception of it vs stereo or even mono signals, usually used in the tests.

The article also states:


That is incorrect/misleading! Though Microsoft does provide such capabilities in their WMEncoder, such audio is NOT part of HD-DVD specification ratifiend by the DVD forum and will not be found on future HD-DVD disks.

Dolby, DTS, LPAC (lossless), and mp4 (including HE and that new stuff Nero is working on to improve it, I'm bad with acronims) are the codecs. Only WMV9 Video has been approved.

This is a very important point as well. Dr. Griesinger's presentation shows us that we definitely can't hear the separated ultra-sonic images - he had a perfect experiment with the keys.

Of course, a valid DBT that can be accepted as credible requires some signficant effort and resources. Just review any of the peer-reviewed/scrutinized studies, to see the general level of effort and statistics required for a meaningful DBT. The ABX tests produced by this site are useful for thie purposes, but do to lack of actual controls(remote testing - no verification, etc.) and procedure, would not hold up under hard scrutiny -- though far more useful then a standard sighted subjective test.

Do you realize the 'audibility' chart on your site pertains to pure sinuisodal waveforms? The tolerances for audbility are far higher with music program. Even, with a purist piano recording, the JNDs are in the 1-2% average range of total THD of detectiablity thresholds for listeners:

Just Detectable Distortion Differences
Wireless World, February, 1981, pages 32, 33, 34 and 38
James Moir

-Chris

This doesn't have anything to do with the format itself! If the recoding industries set themselves a higher standard in the production of digital music than they have in the past, I will for sure only applaud that. But no new format such as DVDA or SACD is needed for that. See also krabapple's post.

WmAx and krabapple already explained why I consider the paper you mentioned rather "controversial" in this discussion.

In slight OT, I had a recent run-in with Telarc's Michael Bishop in e-mail. I was upset that a new cd from them had so much clipping in it and was set ast such high absolute levels, and only something like -40db min, levels of the 96dB avaialble. In this exchange, he told me something suprising: That the RBCD version had been more compressed then the SACD version. Apparently to 'compete' with the other RBCDs. No such warning were found on or inside the CD case!

Take a look at this excerpt from the reply email stating this:



In this case, we literally have a CRAP PERPUTUATING CRAP scenarion. Egads!

-Chris

So, the SACD-version was mastered less-agressive because of a "less-competitive" market....... does that imply that once the SACD or DVDA market becomes "more competitive" that they will the just start another loudness-race?

If only replaygain would be a standard inplementation in almost any hardware-player, then this whole mess would solve itself..... because it would be the hot-mastered records which would sound "less competitive" then.

So in other words - what would really improve "dynamic-range" in the music-world, is NOT another audio-format...... but instead a technology like replaygain which makes the loudness-race irrelevant.

- Lyx

In my case, it means it's not unreasonable to consider the possibility of bias, when a person might make a profit from his claim (in your case), or when he's paid a lot of money for a mod (in the customer's case). It would be unreasonable to dismiss the possibility outright or to ignore it. I hope you don't hate me for that.

Which particular belief of Doug Self's are you referring to, btw? I have found his work to be pretty reasonable...




I know . ;>





Well, your DBT results might make you one. ;>

Seriously, we need more *good* research into audible difference between formats; ther are *far* too many claims out ther alraedy that seem to believe the matter is settled (usually on the side of 'SACD is clearly superior to CD'). I look forward to detailed presentation of your DBT methodologies and results. ANd *this* is definitely a good place to have both dissected! This forum has the means and experience to help you conduct a multiple subject test,if you want to go through with that. Another place to get good feedback from skeptics -- at least a few of whom have engineering/audio component design backgrounds -- is rec.audio.high-end.




The NHK labs results have already been mentioned, and a search of 'Oohashi' on many audio forums (including this one) will produce the usual objections. Your hypothesis about the ultrasonic area remain that...you still need to demonstrate that they are making the difference, and not something else, because as you surely know, not all measurable differences are audible.




He says *might be* -- which is an important qualifier.



Actually, how well-respected that journal is, is something I'd have to ask a neurophysiologist about. I know from my own field that 'respect' is hard for people outside a field to gauge. I know which journals are 'respected' in developmental biology, but I wouldnt' be surprised if a physicist had no idea whatever.



What matters much more, is who the authors are, not how many.



YEs, that's a standard one. You also cite Dr. Greisinger...I wonder what he'd make of your data.




An interesting idea, worthy of more research. If you can set up some sort of audio comparison through this forum, I'll be happy to participate.

[quote=lexor,Aug 19 2004, 02:40 PM]That is incorrect/misleading! Though Microsoft does provide such capabilities in their WMEncoder, such audio is NOT part of HD-DVD specification ratifiend by the DVD forum and will not be found on future HD-DVD disks.
Dolby, DTS, LPAC (lossless), and mp4 (including HE and that new stuff Nero is working on to improve it, I'm bad with acronims) are the codecs. Only WMV9 Video has been approved.[/quote]
You are right, this information will be corrected. WMV9 is approved, not WMA9. Thank you.
[quote=WmAx,Aug 19 2004, 02:48 PM]Do you realize the 'audibility' chart on your site pertains to pure sinuisodal waveforms? The tolerances for audbility are far higher with music program. Even, with a purist piano recording, the JNDs are in the 1-2% average range of total THD of detectiablity thresholds for listeners:[/quote]
This is a different story. The audibility of op amps was created using RMAA, and this is a measuring software, which uses sine waves for measurements. I compared the graphical results with Albini graphs. This just illustrates the suggestion. It is not related to DBT (musical samples).
[quote=krabapple,Aug 19 2004, 08:02 PM]In my case, it means it's not unreasonable to consider the possibility of bias, when a person might make a profit from his claim (in your case)[/quote]
I don't provide the mod services anymore (till over two months) to be able to avoid such questions.
[quote]Which particular belief of Doug Self's are you referring to, btw?  I have found his work to be pretty reasonable...[/quote]
http://www.dself.dsl.pipex.com/ampins/webbop/opa627.htm
[quote]The OPA627 (Burr-Brown) is a very expensive opamp with excellent DC precision. But is it any good for audio? The answer is yes.[/quote]
http://www.anthemav.com/OldSitev1/pdf/sfd2mk3.pdf - page 3
OPA627+OPA2134 is a recommended combination for PCM1704 by Burr Brown engineers as well (take a look in PCM1704 data-sheet)
[quote]Well, your DBT results might make you one. ;>[/quote]
No, I had made DBT only after some attacks on AVS forum, I was too lazy to make it before. I hear the difference, it was not necessary to make DBT from my point of view, but people wanted it ...
[quote]Seriously, we need more *good* research into audible difference between formats; ther are *far* too many claims out ther alraedy that seem to believe the matter is settled (usually on the side of 'SACD is clearly superior to CD').  I look forward to detailed presentation of your DBT methodologies and results. ANd *this* is definitely the place to have both dissected (not avsforum)!  This forum has the means and experience to help you conduct a multiple subject test,if you want to go through with that.   Another place to get good feedback from skeptics -- at least a few of whom have engineering/audio component design backgrounds  -- is rec.audio.high-end.[/quote]
Thank you, it is a good advice, seriously
[quote]The NHK labs results have already been mentioned, and a search of 'Oohashi' on many audio forums (including this  one) will produce the usual objections.  Your hypothesis about the ultrasonic area remain that...you still need to demonstrate that they are making the difference, and not something else, because as you surely know, not all measurable differences are audible. [/quote]
For me NHK tests are significantly less valuable then Oohashi's. First of all - it is not possible to fool DBT system when you want to hear the difference, but it is very easy to fool it when you don't want or just too lazy to hear the difference ... some people who just don't want to pay attention or don't know what they intend to hear can break any test. Then - the used equipment. Why use in the listening tests not professinal monitors, which can reveal anything, but High-End speakers, which are designed to provide the most pleasurable sound, hiding imperfections?
[quote]He says *might be* -- which is an important qualifier.[/quote] As well as yours "potential"? You guys are trying to have no responsibility and never be caught I don't need such word as I listen to DVD-A, SACD and CDs almost every day in my system, all types of them. I have my own opinion and can share it, right? If people who financially benefit from the High-Res formats, don't make massive DBT to scientifically prove that they sound better, why do I need to bother to do it by myself for free?
[quote]Actually, how well-respected that journal is, is something I'd have to ask a neurophysiologist about.  I know from my own field that 'respect' is hard for people outside a field to gauge.  I know which journals are 'respected' in developmental biology, but I wouldnt' be surprised if a physicist had no idea whatever.
What matters much more, is who the authors are, not how many.[/quote] The results of measurements can be flawed intentionally or just because of mistake and signed by one person, the probability with 10 is significantly less ... don't you think so?
[quote]YEs, that's a standard one.  You also cite Dr. Greisinger...I wonder what he'd make of your data.[/quote] I have no doubts that Dr. Griesinger is a great scientist and I made my measurements just because I hear the difference on my equipment ...
[quote]An interesting idea, worthy of more research.  If you can set up some sort of audio comparison throuhg this forum, I'll be happy to participate.[/quote]
I am not ready to spend so much time yet May be a bit later.
[quote=Lyx,Aug 19 2004, 04:02 PM]So, the SACD-version was mastered less-agressive because of a "less-competitive" market....... does that imply that once the SACD or DVDA market becomes "more competitive" that they will the just start another loudness-race?[/quote] SACD version had been mastered less aggressive because it has more reserve in DR ... IMO

These are not valid arguments. The Ooashi test, in the so-called audibility section, was poorly documented and produced.

(1) No individual data is provided for the audibility section.
(2) Vague details of the actual session.

Remember, Ooashi's main focus was the brain scan section. The NHK lab paper was carefully produced and focused on audibility. Their is no cmoparision, the NHK paper is extremely well documented on the audbility information and all data. The equipment used was cutom made precision systems. It is irrelevant if it was a name-brand monitor,e tc. Actually, the speakers they made would have to perform superiorly to any standard monitor, since they were so careful to divide, filter and actively power so many discrete multi-way speakers to insure no IM products contaminating the test. The listeners were many, and nearly all audio experts. This is a far better test of actual audibility then Oohashi's. Besides, the Plenge, Shone and Jakbuwskis paper that is the established standard landmark, has yet to be rebutted successfully. It still stands.



You can have any opinion you wish. :-) However, it's not supported by factual information in this case. With a properly dither, a RBCD format can approach 96dB dynamic range. This is far beyond what an be used in any pratical listening environment, or what speakers/equipment can even produce. Let's say you have an extraordinarily quiet special built room that has a noisefloor of 35dB(RARE): let's start the noisefloor of the RBCD at or under this, your playback system would ahve to be able to produce nearly 130dB cleanly. Not considering potential hearing damages, the only speaker that can manage this in a home environment are large line arrays or super-efficient large horn speakers.

Besides, analyse the waveforms on some of your CDs. You'll find around 20dB RMS used on much modern rock/pop, perhaps 40dB on some jazz, etc. and no more then 65-70dB RMS average used on very dynamic classical recordings. So, the dynamic compression that is imposed on RBCD has nothing to do with any physical limitation, but a purposefully imposed limitation by the recording engineer(s).

Besides, the [1]optimal dynamic range in most practical situations for speakers is approx. 75dB, and headphone about 80dB. RBCD is signficnatly more capable.


[1] Signal-to-Noise Ratio Requirement for Digital Transmission Systems
Spikofski, Gerhard
AES Preprint: 2196

-Chris

It's a sad day when Telarc are competing in the loudness race.





That's an irrational statement, isn't it?

1. The person who made the CD has said that it was dynamically compressed to compete with loud sounding CDs. That was the reason he gave.
2. Telarc make hundreds of non-squashed CDs - some of them with very quiet passages far below the lowest level that this recording would have without dynamic range compression. They don't squash these CDs for classical markets where there is no need to compete in the loudness race.
3. Telarc are very pro DSD and will tell you how much better SACDs sound than CDs whenever they have the chance - yet this time there's no suggestion that this is the reason for the squashed CD sound - it's just a commercial competative decision.
4. The background noise limit on a CD is 96dB below full scale or lower (I don't think I'm telling you anything new here ) so this recording isn't going to fall below it. Even the (highly) subjective improvement that can be gained by going from 16-bit to 24-bit or DSD is tiny/minute/inconsequential compared to the audible damage that is done by dynamic range compression. So the recording would have sounded better on CD without the compression.

Despite all this, you suggest that the SACD version has less dynamic range compression because SACD has a greater dynamic range?

The cynic in me thinks that, if there is any other reason, it's to make sure that the better sounding version is on the more expensive format. To be fair to Telarc, I don't think they're that cynical - I'm sure other labels are though.

I reckon other labels will try to degrade the CD compared to the SACD or DVD, but they will find that users just copy the SACD or DVD-A through the analogue outputs, convert it to 44kHz, encode it with MusePack (or even mp3), and send it across the net. Then anyone using a file sharing network will be able to obtain a better CD by downloading the files and burning them to a CD-R than by going to the shop and buying the over compressed commercial CD. The industry couldn’t be so stupid, could they?


Cheers,
David.

Come on, Alex, there's nothing on that page about audible effect of using this op amp versus another. He just shows that it's a technically fine audio op amp.



Again, the question was whether the mods make an audible difference. Not whether one could swap one op amp for a *measurably* different/better one!

Self is about as 'objectivist' as they come...maybe i'm missing something, but I don't see him advocating modding of devices like the top-line Pioneer receivers.

I am sorry for the late response, guys. Some corrections had been made in the article, for example concerning Microsoft.

Chris,
I am not qualified enough to discuss non-perfection of scientific articles. I would like to remind you, though, about Legato Pro audible difference, it is not a natural way to add audible ultrasonic, but using slow roll-off in digital filters instead of sharp roll off some ultrasonic images had been artificially restored. The system sounded differently, I don't want to argue here better or worse, but no doubts we could hear the difference, and it is only in ultrasonic frequencies.

Actually it was a provocative answer, I am sorry for it. I had a hope that someone would explain mentioned by jheoaustin two SACD modes:
http://www.avsforum.com/avs-vb/showthread....899#post4196899
"The plots are really interesting, and reminds me of the Sharp demo/presentation for DSD to the industry in 1997. They were showing 2 DSD modes, and one was high quality mode of 120dB SNR and 20 ~ 22kHz bandwidth. The other was the wide band mode of 96 ~ 100dB SNR and 100kHz bandwidth."
Since my experiment confirmed some kind of those two modes http://members.cox.net/alex_lat/Tests/SACD_modes.PNG , I had a hope that someone would explain to me -100dB 100kHz SACD mode
I know that with dithering and noise-shaping DR in CD can significantly exceed the limit for S/N 96dB and can be:
http://www.megabitmax.com/content/technicalinfo
"MEGABITMAX provides up to 14 dB increase in perceived dynamic range over the standard white noise TPDF dithering, and up to 6 dB increase over the most powerful of today’s Word Length Reduction systems"
Over 14dB increase vs already dithered signal!
Therefore we have some technologies, which can increase the DR for CDs above the level of DR for 100dB 100kHz SACD, right? And "higher dynamic range" can't be a reason for non-clipped signals in SACD
Some additional information about 16/44.1kHz vs 24/96kHz
http://www.nanophon.com/audio/antialia.pdf
http://www.tnt-audio.com/sorgenti/2496e.html


Agreed.

David,
It really is The explanations are above. Thank you for your input.


krabapple, you had ignored another link, posted by Sonic frontiers:
http://www.anthemav.com/OldSitev1/pdf/sfd2mk3.pdf

Take a look here as well, it is in German, but anyone can use babelfish now, right? http://www.hoer-wege.de/dacupPCM1704.htm

Actually I modded it for myself (swap was for 3 op amps per channel, and what is most important, in I/V stage - see also the information about importance of this stage in Hawksford's article), and only after that decided to share such information Do you think that guys as Chris Johnson (SFI/Anthem/Part Connexions http://www.partsconnexion.com/assemblage.php), Walter Liederman (Underwood Hi-Fi http://www.underwoodhifi.com/whoweare.html), Dan Wright (ModWright http://www.modwright.com/) don't deserve to be respected just because they "advocate" modifications in High-End devices?
Take a look here and http://web.telecom.cz/macura/opamp_mer_en.html

THD point of view is not 100% relative to the sounding, though if we know the spectrum in the audible range, quite informative.
May be this paper will be helpful fior you? http://www.essex.ac.uk/ese/research/audio_...20amplifier.pdf
Here is a discussion about op amps for I/V in DIYAUDIO forum: http://www.diyaudio.com/forums/showthread/t-34324-p-1.html
And finally I had made some measurements with different op amps and DBT as well, this is an idea for another article, though no time for it yet ...

P.S. After the discussions in Audio Asylum forum there is an interesting suggestion concerning ultrasonic images in measurements, as assumed by some people, produced in 49TXi. I have only one suggestion at the moment - if we don't have such images on multi-channel tracks of James Taylor "Hourglass" and Steely Dan "Two against Nature" DVD-A 24/96, as claimed by Dave Kingsland and John Kotches, (I still have some doubts about that due to the following information):
http://www.warnervision.com.au/bio.asp?id=1883
http://www.scotthullmastering.com/Articles.html
"But for this project, the entire signal path was digital – but utilizing the higher sample rates and bit widths. The Genex source went directly to my Sonic Solutions editing system recording at the full bandwidth of 96x24. After the mixes were edited and compiled, the 96x24 bit AES signal was routed to my Z-Sys 96k 24-bit 6 channel digital EQ. We only needed 2 channels for this version, but this is such a nice sounding digital EQ."
Then this statement of Pioneer is not truthful: http://members.cox.net/alex_lat/images/Audio_scaler.PNG
and so called "Audio Scaler" works for DVD-A and SACD as well as for CDs and DD/DTS sources, adding artificial ultrasonic images, and I can see only one ability for that at the moment - switching from sharp roll-off to slow roll-off in digital filters DF1706 (exactly as in Legato). I need to make some additional measurements to prove what is correct.

There is no need for me to responsd to each article. I did skim through some of them, but I did not notice anything proving positive results in a listening test that woudl hold up under scrutiny.

That's really my only point -- people like to speculate and assume things that have not been proven under the strict controls required to be accepted as a truth. It really does not matter what the title or what degree the person holds that makes this speculation. Speculation is speculation, regardless of who spreads it around. :-)

Don't you find it curious that a big compnay like Sony has not produced a single scientifically valid test to prove that it's new format has an audible advantage? Such tests were paramount in development of RBCD format. In truth, they developed the RBCD specs based on parameters that exceeded the overwhelming majority of human discernability in the relevant playback conditions. Such recent audibility tests as the NHK labs test, that took every precaution possible to prevent errors(special speakers, special amplifier/filter systems, special recordings, audio expert listening panel, etc.) have only failed to produce positive results of audbility. The only examples I have seen of positive results were those that had many open-ended variables not accounted. Of course, you can feel free to critize the filter that NHK implemented. But it was ideal, in that it allowed a simple switch in/switch out of the ultrasonic spectrum. But then, criticize the filter method used in the other paper (What Bandwidth Is Necessary for Optimal Sound Transmission?, Jakubowski, Schone, Plenge)I referred to, which simply increased the filter cuttoff point until it was not disceranbe in reference to the unfiltered signal. NOt even 20kHz was reached before it was not discernable.

-Chris

EDITED FOR FACTUAL ERROR/TYPO.

There is actually a word for that, axiom. I'll give you an example of why your logic doesn't hold. Please prove to me that "up" is where you think it is. You just know that it's there, but can you even begin to prove it to a person that has no notion of direction (brain demage/trauma) and would say there is no such thing as "up"?

The imposibility of proving something doesn't translate to it's incorectness. There is a few things in math that are known to be true, but it is believed that they cannot be proven (Set Theory is full of that, the basics of Set Theory cannot be proved, as in not possible at all, but all math of today has been re-derived from Set Theory).

That's an uneeded statement. I never claimed such a thing, if you that is an implication. But, the illogic of assuming ideas that have no strong evidence to support them as true is ultimately counterproductive. That's exactly why scientific methodology is the standard. Do you know of a better/more efficient way? I don't.

In this particular case of audiblity of ultrasonic bandwidth -- I'm not aware of credible data(that can hold under any degree of scrutiny) has proven, or strongly suggested audibility should be possible.

-Chris

That sentence I don't understand, could you refraise "if you that is an implication", you mean if I'm implying that? If so then you seem to be forgetting your main argument which you actaully repeated in this last post above. You argue against the author of the article because he cannot provide anything solid to support the claims.

All of moderm mathematics has been re-derived from Set Theory, the very basic assumptions and statements of Set Theory (such as what is empty set) cannot be proved and have no credible data(that can hold under any degree of scrutiny) supporting it, and I mean absoluteley nothing, we as humans just have an understanding of what empty means.

To give a more tangible example, Set Theory backs up simple arithmetic, arithmetic says 2-4=-2, that is based on the fact that Set Theory tells us so, but set theory has no prove or any evidence to even remotley support it's basic axioms, so according to you, we can't say 2-4=-2.

Ok, it seems clear now that you are(understandably) confusing the methdology of math with that of science, which in fact are two different things:

1. Science is a systematic approach developed for efficiency of progress/development/understanding of the physical universe. This site is founded on such. Hypothesis and Theories are developed, and tested in science. They are not assumed as true until they have been proven. Even then, they may only be provisionally accepted. This is a system developed for a purpose. Mathematics has a purpose, too, and it's systematic operations for development seems to be rather different from that of what has been found optimal for physical analysis. IT seems that each has developed a system based on it's own requirements for efficiency. Unless you have a yet more efficient system, it seems useless to debate the validity of science based on the optimal system for a different application. Mathematics are not a physical entity. Consider the alternative of science in the physical world [1]

2. You refer to this as "according to you(me)", but I'm hardly the one that developed/established scientific methedology. But, since math has different needs to be efficient, I don't make any claims that it should be approached in exactly the same way as the physical universe, nor be subjected to precisely the same burdens of 'proof'. I follow it (science) because it's highly efficient in comparison to the alternative: [1] Abandonment of science would open up every matter to an infinite number of possible variables, never being able to quantify anything. To abandon means you must accept all theories/ideas/speculations equally. No requirement for evidence, remember. If you want to abandon the requirement for proof, go ahead please. But will you enjoy living in a world of where ANY idea deserves merit without substantiation? I personally think your just arguing for the sake of argument; it's hard for me to believe you assume everything/every idea true by default. Perhaps you enjoy philosiphy? See next item....

3. One thing I am definately not interested in is a debate of science vs. rational/justification of mathematical assumption(s), which is really a philisophical debate. A google newsgroup search will probably yeild hundreds if not thousands of results, if this is of interest to you.

-Chris

Chris,
Let's prove the results of this scientific research, OK?
What Bandwidth Is Necessary for Optimal Sound Transmission?, Jakubowski, Schone, Plenge
Very old research using 500Hz tone with harmonics till 25kHz. Actually they were defining the audibility of harmonics of one tone, not ultrasonic images. Their result - 15kHz is enough even with the phase distortions of analog cut-off filters. Enough for what? For mp3 or for 500Hz tone? May be ... It is THAT easy to break those scientific research results for the music signal of high resolution ... here are two files for download, both made from the same source - the spectrum of one is limited at 30kHz (as you suggest, anyway we don't hear ultrasonic images, then no reason to limit 30kHz spectrum to 25kHz to match the results of those measurements), the spectrum of another one is limited at 15kHz. Use DBT (I recommend ABX plug-in for foobar2000). Please note, that I used digital filtering WITHOUT phase shift. Score your result.
I got 13/20 in the beginning, and 15/20 (2.1% of guessing) after some time ... the limit is 5%, right? It means I can hear the difference. I am an average listener, because I made a sonogram at the doctor for my listening abilities ... the limit is 18kHz for me.
Then - this is REALLY IMPORTANT - I didn't use my perfect M&K 150 THX monitors here by purpose!!! I used Sony DP-IF5100 IR headphones, from analog output of Lynx L22 soundcard to analog input of DP-IF5100 ... I would like to remind that those headphones can handle only 48kHz sample rate ... and they convert the analog signal to digital ... we have the regular situation here with an average listener and average electronics ...
here are the results:
In the beginning: http://members.cox.net/alex_lat/Tests/beginning.PNG
After some training (I am sure I can do it better): http://members.cox.net/alex_lat/Tests/tests.PNG
And here are the files for you to do DBT:
I'd put two files at my server with 30kHz and 15kHz FR limits. My server's limit is 300mb, so please download the files fast enough: http://members.cox.net/alex_lat/Tests/Test.zip
Used software - Adobe Audition 1.5, Marquee tool.
Here is the comparison of spectrums for two signals:
http://members.cox.net/alex_lat/Tests/test1.PNG

Well, we can say 2-4=-2, but we can't say "This statement is false" is false.

More to the point of the topic: you are grossly confusing inductive reasoning (science and statistics - from the specific case to the general case) with deductive (mathematical - from the general to the specific) reasoning. Physical phenomena can't be directly reasoned with deductively, because we can't test each statement across all possible values and cases. When we make any statements about reality, we must implicitly weigh them against the weight of the observations backing up those statements.

I don't know what you mean by prove. That is a peer-reviewed journal paper. Has not been successfully discredited in the over 2 decades it's been around. :-)



And, what is the problem with the test signal? It's consistent test signal with spectral content that is broad and high in amplitude. As for ultrasonic image... that is an unproven concept, at least if you mean an 'audible' image.



Easy? 1st, it seems you have not followed the other threads such as the 96khz listening test thread, etc.. You can't break this reseach so easily with so many open variables that can potentially cuase false positives. Their were some positive results obtained in that thread too(you should read it), but they are inconclusive due to lack of necessary controls/analysis of hardware.

Then their is the NHK labs experiment that used music program material, and more extensive precautions then the original JVC(plenge, schone, jubowski) research. Where were the ultrasonic images? Why could not one person identify these elusive ultrasonic images? A test performed using equipment and recordings that were insured to have rich ultrasonic content. Maybe the test was 'fixed'? :-)

A positive result from conditions/equipment/hardware/signals with unknown behaviour in ultrasonic frequencies does not conclude anything.

-Chris

"Prove" means - let's try to use those results for our music signals, as you tried to use them ... OK?

The problem is that 500Hz tone signal with harmonics is not equal to the misic signal, as well as a sine wave ...

I don't even need to break it You can do it by yourself, if you will use the provided samples Hardware? I had provided the used hardware, but it doesn't really matter
Let's do it step by step First of all let's finish with the first scientific research, then we will talk about another one

Why "unknown behaviour" if we know the schematics and used elements with their limitations?

If you believe that, then consider it already broken. Their have been numerous 'tests' that established audibility WITHOUT closely analysing the source signal, ampllfiers, transducer distortions.....



Well, you seemed to not like the idea of non music signals. So with that NHK lab test using music signals, you can't use that as an objection.




No, we don't know. That's the point. We don't know what differences(errors) the filter may have caused in the software(careful analysis of test waveforms needs to be performed first, in the software before/after this filtering), the possibility of hardware distortion(transducers will very probably do this if the ultrasonic amplitude is anywhere near the ampllitude of the audible band content, see earlier referenced Griesenger presentation), possibility of error by person prepping the file(s), etc.

I have no intention of ABXing the files you presented(becuase of the above reasons), but I did download and the first thing I noticed is that you already made one error. ONe file is mono, while the other is stereo. Errors are easy to make. An error free test is very difficult to pull off. That's why I don't give much respect to such loosely controlled tests such as your trying to perform. No where near the level of controls are being used to near scientific validity.

-Chris

What is the relevance of ABXing a 15kHz low pass? Is it just to demonstrate that a previous experiment (which showed a 15kHZ low pass was inaudible) was wrong? I don't think we need to prove that again - there are enough threads already! Anyway, the experiment wasn't wrong - it showed that those listeners with those equipment listening to that particular sound couldn't detect a 15kHZ low pass filter. It's almost irrelevant to high quality music listening - drawing the conclusion that 15kHZ was enough for music was a false conclusion, but it's the applied logic that's faulty, not the scientific results.

As for the CD/SACD/DVD argument, none of them use a lowpass of 15kHz, so again, what are you trying to prove here?

Cheers,
David.

ok I did stear the discussion off topic so I'll just correct this last statement and switch gears.

The reason I said 2-4=-2 can't be said is because WmAx claimed we need to back-up with solid evidence, I made no statement of incorrectness of above math.

Second "This statement is false" is not a mathematical statement. A mathematical statement has to be either true or false and never both or neither. You are getting confuddled by logic statements, so it's better to stop with this here. Believe it or not this could get hairier than the audio discussion we have going on.

I make this a separate post, just so it doesn't get lost in above mess.


Is it possible (aka did someone try, successfully or not) to create a sound file which would be flat lined in the scope of current CD resolution (aka should be silent/have no information encoded) but have varying ultrasonic data (or anything else that CD format wouldn't store), varying for the purpose of creating interference pattern (i.e. don't just stick monotonous wave in there), and see if interference distortions propagate into audible segment (aka on playback we get spikes in what's considered audible segment)?

Someone experienced with Audition/SoundForge could attempt. Unless of course this has been tried and refuted before?

BTW:

last time I checked that thread the guy doing the listening could ABX 96 vs 44 flawlessly, what are you alluding to there?

Dude, it was a joke about Godel's Incompleteness Theorem! Laugh!

Alluding? If you paid attention to the enire poist, you would realize the results were inconclusive due to unverified hardware and distortion analysis.

-Chris

I'm not impressed by the grammar of the claim , and even less by the lack of scientific listening data in that white paper. No description of listening test methods, controls, quantitation and statistical analysis of results...just a dubious claim that 'warmth, speed (!) and smoothness were the key words described by members of the listening panel'. I already told you what my criteria for convincing comparison reports are..this doesn't even come close.

So, so far, neither of your links supports your claim, AFAIK..



Even Babelfish failed to find any proper listening test results on that site.

That's three sites, no joy.




And so far, no evidence other than anecdotal (from the manufacturer, in two cases).



Depends on what the modification is, and what the evidence is for audible difference. They deserve respect as far as their claims are supported by good reasoning and empirical evidence, no more, no less. It's quite possible -- I've seen it many times -- that smart people can believe wrong things as well as right things ...hey, it's happened to me once or twice too. Chris Johnson, I am familiar with, and he *generally* takes care to support his claims....but I've also seen him get into some huge wars with objectivists over on RAHE overs some of his claims. Have any of these people reported results of listening tests for particular mods?




No listening test results there.



Page doesn't open.. does it report properly done listening tests?



Again, SHOW ME THE PROPERLY DONE LISTENING TESTS.
Measurably differences do not necessarily translate into audible ones. That's a foundational concept in this forum... which is why here, unlike on 99% of forums about audio, claims of audible difference have to be backed up by proper empirical tests.