Home of the Squeezebox™ & Transporter® network music players.
Page 4 of 15 FirstFirst ... 2345614 ... LastLast
Results 31 to 40 of 149
  1. #31
    Senior Member chill's Avatar
    Join Date
    Mar 2007
    Posts
    501
    What a good article. I loved his statement in the outro:
    The more that pseudoscience goes unchecked in the world at large, the harder it is for truth to overcome truthiness...
    Almost by way of an example (although he doesn't make the link himself), in footnote 18 he includes a quote from Wired magazine:
    "Some purists will tell you to skip FLACs altogether and just buy WAVs. [...] By buying WAVs, you can avoid the potential data loss incurred when the file is compressed into a FLAC. This data loss is rare, but it happens."
    I wonder if Wired picked that up from TAS.

  2. #32
    Senior Member
    Join Date
    Jan 2007
    Posts
    751
    Quote Originally Posted by bhaagensen View Post
    In conclusion i agree with you mnyb and others. Its just that wrt the end result the rigidness of mathematical sampling theory does (alone) not carry through all the way.
    Human hearing does "not carry through all the way" either. We are also not infinite.

    Audiophiles spend a lot of time rearranging deck hairs on the Titanic. But it does make them feel better. ;-)

  3. #33

    lots of holes in his article

    Unfortunately, the picture isn't as clear or as simple as he tries to portray.

    The Boston Audio DBT had flaws which are pointed out in many critiques all over the net.

    Among other things, he seems to thing the point of high res recordings is so we can hear high frequenies (above 20k); that of course isn't the point.

    He talks about macro dynamics of music, but not microdynamics.

    Finally, he may be correct. But it may also be irrelevant that he is. Some have made the case that you can hear everything on a well produced Redbook file that you can hear on a hi-res file, only some of the detail is easier to hear on a high res file. This alone can account for the subjective impression that the hi res sounds different or better.


    In any case, even the authors of the DBT admit that there best sounding files were those from SACD. This they attribute to better/different mastering of the hi-res files, and not to any inherent superiority of the format. That may be. But then that alone is a good reason to buy them and listen to them.
    GIK Acoustics Room Treatments. Tranquil PC fanless server running Windows 7 via FW> Mytek 192 DSD DAC;Odyssey Kismet (Khartago case) Stereo Amp; Devore Gibbon 9 Speakers; Dual 506 + Ortofon M20 (occasional use); SB Touch slaved to Empirical Audio Pace Car;SB Boom and SB Touch in additional rooms. Arcam CD82 which I don't use anymore, even though it's a very good player.

  4. #34
    Senior Member Soulkeeper's Avatar
    Join Date
    Dec 2009
    Location
    Norway
    Posts
    1,220
    Quote Originally Posted by firedog View Post
    He talks about macro dynamics of music, but not microdynamics.
    But what is microdynamics? Does it have a sensible definition, or is it just one of those woo-words that can mean anything that the user wants it to? (something like Spiritual Holistic Quantum Dynamics?)

  5. #35
    Senior Member
    Join Date
    Jan 2007
    Posts
    751
    Quote Originally Posted by firedog View Post
    Unfortunately, the picture isn't as clear or as simple as he tries to portray.

    The Boston Audio DBT had flaws which are pointed out in many critiques all over the net....

    Among other things, he seems to thing the point of high res recordings is so we can hear high frequenies (above 20k); that of course isn't the point.

    He talks about macro dynamics of music, but not microdynamics....
    The inability of humans to achieve perfection in blind testing seems to be the touchstone used by those audiophiles to dismiss the results of any test that challenges their beliefs.

    As many others have pointed out, conducting good blind tests is difficult and even the best tests will never cover all theoretical possibilities. That's true whether the subject is audio or something else.

    OK, even if "perfection" in such tests is impossible, what does one do with the accumulated evidence that says that many of the breathtaking differences audiophiles hear under sighted conditions are either wildly exaggerated or even imaginary?

    In the post quoted above, the case for supersonic frequency response is damaged, so the issue is switched to "microdynamics" - another audiophile term that has no specific meaning beyond the vague impression assigned by each individual.

    Can one give an example of a sound that gets lost on a CD that would be heard on a higher-rez recording? In addition to the Empire Brass (brass is always a challenge to record well), I also listened to some Boston Camerata last night, an early music choral group. I could clearly pick out individual voices from the 30 or more singers, whether massed, lead or background singers. What did the CD lose that a high-rez would have revealed? Clearly and accurately presenting the voices of 30 people singing together would seem a good test for the ability of a recording to maintain clarity and not lose articulation.

    Yes, I've heard a number of high-rez recordings and they do tend to be excellent. But I also have many CDs with superb audio quality. I rather strongly suspect the quality of the high-rez recordings is far more due to extra care given in the recording, mixing and mastering process than any inherent technical advantage. These recordings are marketed to a discriminating group of listeners and typically cost more, so they better be giving them something for their money.

  6. #36
    Senior Member bluegaspode's Avatar
    Join Date
    Jul 2009
    Location
    Berlin, Germany
    Posts
    3,156
    Quote Originally Posted by bhaagensen View Post
    Its not that I'm questioning Nyquist - but its a *mathamatical* theory involving the concept of infiniteness - in particular the reconstruction theorem involves an infinite sum.
    I also think this is the weakest spot in the argumentation of the paper.
    While in theory with Nyquist I can reconstruct the waveform with the lower bound of samples (i.e. 2x frequency) perfectly, I need infinite sums for it - so speaking in algorithms: I cannot compute it in reasonable time.

    So what is done in DACs is to approximate the original waveform with as much sums as possible in a given timeframe.

    Let's say for proper playback I can do 100calculations (take whatever number you like) in 1ms.
    Based on 100 calculations: can I get a better approximation of the original waveform if I have 100 samples or 200 samples ?

    Reading Wikipedia articles about Nyquist doesn't give me an answer for that.
    But that's the central question that needs to be answered, not if Nyquist in theory can reproduce a waveform with a given number of samples.
    Did you know: SqueezePlayer will stream all your music to your Android device. Take your music everywhere!
    Remote Control + Streaming to your iPad? Squeezebox + iPad = SqueezePad
    Want to see a Weather Forecast on your Radio/Touch/Controller ? => why not try my Weather Forecast Applet
    Want to use the Headphones with your Controller ? => why not try my Headphone Switcher Applet

  7. #37
    Senior Member
    Join Date
    Jan 2007
    Location
    Midlands, England
    Posts
    627
    Quote Originally Posted by bluegaspode View Post

    So what is done in DACs is to approximate the original waveform with as much sums as possible in a given timeframe.
    NO

    The DAC generates a stepped ladder function, the reconstruction filter turns this into the recovered waveform.

    In the case of many modern, oversampling DACs, the reconstruction filter is a simple 2nd order analogue filter. No limited number of calculations. Yes the transform theory for the filter is a set of infinite sums, but the filter is a filter.

  8. #38
    Senior Member
    Join Date
    Jan 2007
    Location
    Midlands, England
    Posts
    627
    Quote Originally Posted by bluegaspode View Post

    Let's say for proper playback I can do 100calculations (take whatever number you like) in 1ms.
    Based on 100 calculations: can I get a better approximation of the original waveform if I have 100 samples or 200 samples ?
    NO

    Assuming the 100 sampling frequency met Nyquist criteria and the ADC didn't use old style brickwall anti-alias filters, and that the anti alias filters were set to the same frequency, the post DAC signal would be identical in both cases! You could get Phil to check using his difference test, there would be nothing.

    Above Nyquist sampling means all the data is available to recover the total signal. It doesn't need infinite processing power, it is an analogue technique.

    The modern approach using oversampling take the anti aliasing ADC filters and the reconstruction DAC filters way out of audible effects range.
    However there are still the matters of ADC and DAC linearity, are the steps really equal;
    what is the clock v data jitter presented to the DAC, that effectively changes step sizes with time (note the higher the sample rate the tighter the jitter spec needs to be);
    what is the underlying noise level. 24 bit DACs don't usually achieve much more than 21 to 22 bits of signal to noise performance.

  9. #39
    Senior Member pippin's Avatar
    Join Date
    Oct 2007
    Location
    Berlin
    Posts
    11,748
    This is just wrong.
    You don't need an "infinite sum" or anything, in theory you don't even need a DAC, all you need is a low-pass filter. A DAC is nothing else than a number of carefully tuned band-pass filters followed up by a low-pass filter to clean things up.

    No sums involved. No infinity involved.
    ---
    learn more about iPeng, the iPhone and iPad remote for the Squeezebox and
    Logitech UE Smart Radio as well as iPeng Party, the free Party-App,
    at penguinlovesmusic.com
    New: iPeng 8, the Universal App for iOS 7 and iOS 8

  10. #40
    Senior Member Wombat's Avatar
    Join Date
    Feb 2006
    Posts
    816
    Quote Originally Posted by firedog View Post
    He talks about macro dynamics of music, but not microdynamics.
    Can you explain what this term "microdynamic" correpondents with from a technical standpoint of view?
    I lately linked to that thread already if you talk about time resolution:
    http://www.hydrogenaudio.org/forums/...opic=91126&hl=
    You can always ask J.J. Atkinson for some math that underlines your statement. I am sure he will help you.


    The issue playing back signals above 20kHz can damage playback in the audible range is really interesting but not much sources i can find.
    The intermodulation of amps can be measured but even this is not easy audible. The speakers themself do so much more harm it doesn┤t really matter.
    What i recognize is that with some newer B&W speakers like the CM8 i see a hefty +10dB resonance on-axis at ~30kHz in frequency response graphs i found on the net. So playing back any content at that frequency may indeed trigger the tweeters break-up and distort sounds below.
    Transporter (modded) -> RG142 -> Avantgarde Acoustic based 500VA monoblocks -> Sommer SPK240 -> self-made speakers

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •