An interesting argument against sampling frequencies above 96kHz

Discussion in 'DACs' started by doctor fuse, Oct 30, 2017.

  1. E-Stat

    E-Stat Super Member

    Messages:
    4,146
    You're missing the point. The SACD layer of some were sourced from upsampled Redbook masters. I, Robot was recorded analog in 1977. Patricia Barber's Nightclub was recorded in 2000 prior to the introduction of SACD! Surely you understand that DSOTM was recorded nearly forty-five years ago. Speaking of which, the engineer who produced the SACD remix is an avid promoter of high resolution recordings, including 192/24.

    Ever seen a spectrum analysis of recorded content? You can clearly see when they are bandwidth limited.

    Read the description again:

    "I do not currently have a detailed equipment list for this venue, but the speakers were very large and capable high-end monitors, approximately 7 feet tall, and the power amps were sufficient to drive the speakers to very high levels without audible distortion."

    Perhaps that impresses you...

    That makes no sense whatsoever. Crappy gear won't reveal what better gear can.
     
    Last edited: Nov 14, 2017 at 5:56 PM
  2. E-Stat

    E-Stat Super Member

    Messages:
    4,146
    I find that you cannot recreate what was never captured despite tricks to "interpolate" the missing information.

    The Music Hall DAC 25.3 in my garage system allows me to choose native resolution or to upsample to either 96 or 192. Native sounds better overall to me.
     
  3. gvl

    gvl Well-Known Member

    Messages:
    500
    Definitely cannot recreate, however for example there is an obvious difference when listening to 44.1 vs 96kHz on a NOS DAC, and on an OS DAC it is much closer.
     
  4. E-Stat

    E-Stat Super Member

    Messages:
    4,146
    You're welcome to your opinion - which as previously noted is different than mine. I can simply switch between modes.

    If you're comparing different DACs, there is much more to the results than merely the front end.
     
    Last edited: Nov 14, 2017 at 7:48 PM
  5. bachdude

    bachdude New Member

    Messages:
    28
    Of course, upsampled Redbook would not qualify as high res. But you name one. How many were there? Patricia Barber's Nightclub would have been recorded in High Res PCM and converted to DSD and so counts as a high res in my book. And I think it was very smart to include SACD releases of earlier analogue recordings since many at the time, and some still today, claimed that DSD was closer to analogue than PCM and gave a more faithful representation. That is why so many early analogue recordings were re-released in SACD. So yeah, I agree that should have been included.

    Yes, I probably see a spectrum analyzer several times a week. But my question was based on a misunderstanding regarding what you meant by, "They didn't even bother to verify that the chosen disks were actually of higher than Redbook resolution". You said there were many. Can you tell me how many? Because the numbers based on over 500 listening tests don't look good for your position.

    I saw the description. I just don't presuppose that the Boston Audio Society are a bunch of incompetents as you do. I was involved in recording projects when I lived in Boston and took a couple of my own to mastering studios and I know that the level of the mastering studios in Boston is very high - Soundmirror is one example. There are a ton of recording studios out there run by very competent recording engineers that would be laughed at if they called themselves a "mastering studio". I highly doubt that the Boston Audio Society took their study to some hole in the wall studio and then claimed they took it to a mastering studio in the Boston area. For Boston Audio to call a place a "mastering studio" of classical music means it has, as a basic requirement, an unbelievably stellar audio system suitable for classical mastering.

    In case you want more proof, here is the studio at sound mirror that has speakers that meet the description,

    http://www.soundmirror.com/studio/

    When I read the description of the speakers I immediately thought of Soundmirror - one the the top mastering studios in the country - maybe the world. I've heard their systems I can verify that...they will blow you away.

    Disagree. It is harder to design a 44/16 DAC and many cheap ones, and older units, have crappy filters.

    But keep in mind, this thread is really about sampling rates even higher than 96k. Both Lavry and Benchmark claim that converters do not function as well in those extremely high samples rates. Why bother with them?
     
  6. E-Stat

    E-Stat Super Member

    Messages:
    4,146
    I'm delighted you have now come to that conclusion.

    I named three, but the Steely Dan recordings are also not natively high resolution digital recordings. Any answer over zero invalidates the bogus test.

    Would have been? What does that mean to you when it was recorded in 2000 using the technology available at the time? Do tell of the high res format used in 2000.

    I thought the comparison was between multiple digital formats.

    Do review my previous post where I've listed a few and with this post have listed more. At the expense of confusing the issue with facts, the answer should have been : NONE.

    Comparing Redbook resolution to Redbook resolution. Brilliant!

    Feel free to speculate as you choose in the absence of knowledge.

    Extremely high sample rates? Perhaps you are unaware of DXD. And as I've previously documented, the "measurable differences" are virtually non-existent. Did you follow my link?

    Indeed I would not since they employ cheap op amp analog stages. :)
     
  7. gvl

    gvl Well-Known Member

    Messages:
    500
    Do you realize that you cannot fully defeat oversampling in your DAC? Your notion of "native" just means less oversampling.
     
  8. botrytis

    botrytis Trying not to be a Small Speaker Hoarder Subscriber

    I will put this out there - work by Oohashi J. Neurophysiology. 2000. Vol 83: 3548 . The work has to do with the inaudible high frequencies and their affect audible sounds. So having these inaudible sounds affect how other sounds to us. So high res may add to the sound, not take away.

    Also, Rupert Neve has done some work in that area.
     
  9. E-Stat

    E-Stat Super Member

    Messages:
    4,146
    Clearly, you don't have the foggiest clue about the Music Hall DAC 25.3.

    "Lock: locks at input audio bit rate and sampling rate (upsampling off - bypasses SRC4192)"

    Let's repeat that in case you didn't get it the first time: upsampling off
     
  10. gvl

    gvl Well-Known Member

    Messages:
    500
    You cannot turn off the oversampling filter that's built into the PCM1796 DAC chip, that's the cruel reality of Delta-Sigma converters.
     
  11. E-Stat

    E-Stat Super Member

    Messages:
    4,146
    You are welcome to believe whatever you please.

    "Optionally the incoming signal can be upsampled to 192kHz, using a TI SRC4382 high end sample rate converter."
     
  12. gvl

    gvl Well-Known Member

    Messages:
    500
    I don't believe, I read the specs. You can buy into marketing BS as much as you want. If you want "native" the only way to accomplish it for real is a R2R DAC w/o digital filtering.
     
  13. E-Stat

    E-Stat Super Member

    Messages:
    4,146
    As do I. You can choose to employ the SRC4192 to upsample the input or bypass it.
     
  14. bachdude

    bachdude New Member

    Messages:
    28
    Wow, seriously? DVD audio came out in 2000 and SACD came out in 1999. Recording studios had high res recording capabilities years before the general population had access to these formats! The first 96/24 recordings were made in 1996.

    Anyway, I think it was totally valid to include releases of analogue recordings for the reasons I gave.

    I know the recording scene in Boston and Soundmirror would have been the obvious choice for their study and the speakers fit the description. It's a guess but an educated one.

    So believe as you wish. I have a performance to practice for...
     
  15. E-Stat

    E-Stat Super Member

    Messages:
    4,146
    Patricia Barber's Nightclub recording was:

    "Recorded and mixed at "Chicago Recording Company" Studio 4, May 15-19, 2000.
    Mastered at 24-Bit resolution from analog sources at "Foothill Digital", New York City."


    Not natively SACD.
     
  16. bachdude

    bachdude New Member

    Messages:
    28
    So? Like I said, still high res in my book. They transfer quite well. And when SACD first came out there was no native editing or processing. It was converted to PCM and then converted back to DSD. It wasn't until later in Pyramix that you could edit natively.
     
  17. botrytis

    botrytis Trying not to be a Small Speaker Hoarder Subscriber

    Gents, this is getting into a religious-type of argument. It doesn't do anyone any good if we are not civil about this. We are having a discussion and as such please let's keep it civil.
     
    mfrench likes this.
  18. gvl

    gvl Well-Known Member

    Messages:
    500
    I'll try one last time. The PCM1796 chip your DAC is based on has a built in oversampling filter, at least 8x. You cannot turn it off. What you can turn off is the SRC4192 ASRC, this reduces the overall oversampling rate that is the product of the OS done in ASRC and in the DAC chip, but you're left with at least 8x oversampling.
     
  19. botrytis

    botrytis Trying not to be a Small Speaker Hoarder Subscriber

    Not true DCS helped Sony put together DSD recording equipment - http://www.prosoundnetwork.com/archives/dcs-904-a/d-and-954-d/a-converters/25565 this was from 2000 so there was DSD native at that time. Most of DCS DACs are discreet, not chip based and they use FPGA's to programming the DAC - they are pretty upgradable but very expensive, They invented the ring DAC, which is their discreet DAC archetecture.
     
    Last edited: Nov 14, 2017 at 9:17 PM
  20. RGA

    RGA Super Member

    Messages:
    1,082
    Location:
    Hong Kong
    All DAC designers use measurements and by the way all dac designers can make bomb proof ultra excellent measuring DACs because all they would have to do is buy one and reverse engineer it. Jim Austin of Stereophile insinuated that Benchmark designs to get the best possible meausurement without listening to the products that they sell. Other companies can make such a thing and then listen and say - well that sounds bad. Let's try again and again and again and again and they discover that "unfortunately" the best sound comes from doing A, B, and C which makes the measurements look worse but the device makes music sound much better.

    And while the measurement result may lose them some business with the magazine shopper and graph gurus - they will gain more when people sit down and actually listen and compare in the same room at the same time. The manufacturer's job is to deal with the measurements. They are responsible for making the first product and the 500th product off the assembly line to sound/measure the same.

    I never said the manufacturer doesn't need measurements - the end user though should be able to "listen" and evaluate without the need whatsoever to see any measurement or graph on any part of the system.

    If you need to see the graph to determine if something sounds elite or not then you need a different hobby.

    When I began seriously evaluating stereos - I sat shut my eyes and listened - detemined if the system had the overall Gestalt/Goosebump/something-something factor of the music being played. If yes - the system is a winner if not it's not. Then over the years looking at notes taken certain generalities became apparent as to what actually sounded the best. Generalities - NOT absolutes as there would be some exceptions.

    With digital in the mid 2000s it became apparent that NOS at $3k was trouncing $10k-$20k players of the oversampling upsampling variety. And some other generalities were kicking in with no-feedback amps SS/tube versus high feedback amps and the size and shape and efficiency of loudspeakers.

    Some of these generalities were confirmed again recently covering an audio show recently where I listened and selected a room as one of the best three rooms. I chose the room first and THEN discovered that the DAC was a NOS tube DAC from Border Patrol. Coincidence? Perhaps but 2 other rooms in the top 3 also used NOS DACs. All three rooms used completely different loudspeakers and amplifiers as well. So the source kind of matters.

    These rooms that also had hi-resolution computer audio from a computer also all sounded worse than when the rooms played a CD in a CD transport - granted all three used very good transports but still. These experiences indicate support for my previously held generalizations though.

    Perhaps I am more focusing (and not in the spirit of the initial post) on the quality of the DACs and should be focused on the hi-res recording or something. But if a DAC that downcoverts to 16/44 can sound better than an equally expensive DAC playing 24/192 then it could be possible for a converter to sound better converting 24/192 to 24/96. Maybe not in the same machine but it would seem possible. Which brings us back to the quality of the analog output stage. The computer chip computer buffs may be good at the computer chip stuff - but may not be audiophiles or know how to actually get the "analog" part of the machine to sound decent.
     

Share This Page