Digital audio and misconceptions about hi-res audio, sample rates, and bitdepth

Discussion in 'Digital: DACs, USB converters, decrapifiers' started by lm4der, Jul 16, 2016.

  1. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    You could relax the strict 20 kHz band limitation to something higher in order to allow for more benign filtering. Like Jason said, 96 kHz is probably nice to work with.
     
  2. lm4der

    lm4der A very good sport - Friend

    Pyrate
    Joined:
    Sep 26, 2015
    Likes Received:
    1,461
    Trophy Points:
    113
    Location:
    Seattle, WA
    Right, it would be nice if that had been the standard for those reasons, but... I suppose it's really only a matter of time. Except for this MQA thing, which smells like fish.
     
  3. Hrodulf

    Hrodulf Prohibited from acting as an MOT until year 2050

    Pyrate
    Joined:
    Sep 28, 2015
    Likes Received:
    4,276
    Trophy Points:
    113
    You sure about the tap count? Best I can do from a SHARC processor is 4096.
     
  4. lm4der

    lm4der A very good sport - Friend

    Pyrate
    Joined:
    Sep 26, 2015
    Likes Received:
    1,461
    Trophy Points:
    113
    Location:
    Seattle, WA
  5. Madaboutaudio

    Madaboutaudio Friend

    Pyrate
    Joined:
    Sep 28, 2015
    Likes Received:
    545
    Trophy Points:
    93
    Location:
    Singapore
    On the topic of digital filters, Rob Watts, Dac designer for Chord thinks that non-closed form filter is more natural sounding(which is kinda paradoxical?)

    http://www.head-fi.org/t/784602/cho...-please-read-the-3rd-post/16170#post_12513178
     
  6. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    All I got from that was, "I once was stuck in this trap, but now I'm stuck in this trap and have gone through the appropriate mental hoops to convince myself that I'm not actually stuck in a trap because of what I am going to pretend is a different thought process that led me to this not-trap. Also, buy my DACs!"
     
  7. Psalmanazar

    Psalmanazar Most improved member; A+

    Pyrate Slaytanic Cliff Clavin
    Joined:
    Sep 27, 2015
    Likes Received:
    5,345
    Trophy Points:
    113
    @Hands basically "I have no respect for the original samples and don't care what the recording actually sounds like but Chord DACs sound good!"

    Reminds me of Lars Urlich saying Death Magnetic sounded "smoking" in his car when people complained about how compressed and clipped it was.
     
  8. GoodEnoughGear

    GoodEnoughGear Evil Dr. Shultz‎

    Pyrate
    Joined:
    Oct 25, 2015
    Likes Received:
    3,070
    Trophy Points:
    113
    Location:
    Cape Town, South Africa
    I do actually see a valid aspect of Rob's point. Given say an instrument like a cello playing in free air, with the goal to reproduce that sound through recording and playback we need to acknowledge that the goal is reproduce the sound of the actual instrument playing in free air.

    We may feel the best way to do that is to preserve original samples, but if one can get closer by manipulating samples and diverging from the originals, then so be it. There's no holy ground here for me.

    It's just a f'ing dubiously big "if".
     
  9. Madaboutaudio

    Madaboutaudio Friend

    Pyrate
    Joined:
    Sep 28, 2015
    Likes Received:
    545
    Trophy Points:
    93
    Location:
    Singapore
  10. Merrick

    Merrick A lidless ear

    Pyrate
    Joined:
    Jan 6, 2016
    Likes Received:
    12,569
    Trophy Points:
    113
    Location:
    Portland, OR
    That's assuming the goal of a DAC is to reproduce the real sound of a cello playing in free air, as we would hear it at a concert hall. And I think a lot of audiophiles assume that's what a DAC does. However, I always understood it that the purpose of a DAC is to reproduce the recording as accurately as possible, since it's impossible for a DAC to determine how the original performance actually sounded.

    Even the best recording is going to deviate from the original sound of the performance. To the best of my knowledge we have no microphones and recording equipment capable of perfectly capturing a performance as it sounds to the human ear in the real space. So how can a DAC try to reproduce how the performance sounds and not the recording?
     
  11. swamp

    swamp Acquaintance

    Joined:
    Jun 27, 2016
    Likes Received:
    52
    Trophy Points:
    18
    If reproducing the recording is the DAC's job, and reproducing the performance is the transducer's job, then where does that leave the amp? ;)
     
  12. Thad E Ginathom

    Thad E Ginathom Friend

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    14,134
    Trophy Points:
    113
    Location:
    India
    Refer: Old-fashioned meaning of hifi: High Fidelity. The playback chain, in a perfect world, plays back, faithfully, exactly what is on the recording.

    That's the bottom line.

    It leaves it needing to be transparent. colourless. Ahem... with tone controls for imperfect rooms, imperfect ears and, hey, people who want colour.

    But an amp may have many inputs, and the output of a DAC is just one --- and it is the input to the DAC, the digital source material, and what the DAC can and can't do with it, that is the focus here. so perhaps we should not get ahead of ourselves.
     
    Last edited: Jul 18, 2016
  13. ThePianoMan

    ThePianoMan Facebook Friend

    Joined:
    Apr 5, 2016
    Likes Received:
    189
    Trophy Points:
    33
    A lot of audiophiles have a totally inaccurate conception of what happens in the studio, and recording industry in general. Reproduced audio is exactly that: a reproduction of recorded sound filtered through mics, mic preamps, mixing desks, DAWs, ADCs, DACs, amps, speakers, etc. It's amazing that we can get sound as good as we do... a testament to the technology and how far it's come (among other things)
    That said, I prefer recording at 24/96, and only occasionally record on a 32-bit float if I have nine million mics setup. That's just about headroom. If you take a 24 bit recording down to 16-bits, the conventional wisdom I've always gotten in all my studio experience, classes, EE teachers, etc. is to use Dither. It's to hide the crap that comes from quantization. It's actually quite easy to hear on a relatively transparent system, I can best describe it not as low level noise, but as little digital garbage. Quiet, but present. Dither (from the recording side) helps a lot with this. As for the idea of a less-linear DAC sounding nicer... well, we all make decisions about what we color our signal chain with. I would personally choose tubes in an amp over non-linearities in a DAC. That's all I'll say about that.
     
  14. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    Late to the party, but...

    Yup. That's what Jason and Mike said. I don't know if it's an efficient implementation. I know half-band filters for example can skip quite a few coefficients. Dunno.

    "So eventually I tried eliminating the reconstruction aliasing, and boy did this make a big improvement".... reconstruction aliasing... I mean, isn't that what the interpolation filter in most DACs (mega-burrito or Whatchamacallit) and the analog front end of also almost all DACs is supposed to do anyway?

    Like Hands said... Buy my stuff.
     
  15. Madaboutaudio

    Madaboutaudio Friend

    Pyrate
    Joined:
    Sep 28, 2015
    Likes Received:
    545
    Trophy Points:
    93
    Location:
    Singapore
  16. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    I'm tired. Actually aliasing is eliminated by the ADC analog filter before things get sampled and converted to digital. It is also used when converting from a fast rate to a low rate. Basically, filter input signal using a filter with a corner frequency located at less than half the sampling frequency, to avoid fold over (aliasing) of the high frequencies into the low frequencies.

    The reconstruction filters in the DAC remove replica images associated when going from a low rate to a fast rate. They are interpolators.

    So I actually don't know what he means by reconstruction aliasing. Like I said, maybe it's because I'm tired and my brain is fried.
     
  17. Rex Aeterna

    Rex Aeterna Friend

    Pyrate
    Joined:
    Sep 28, 2015
    Likes Received:
    212
    Trophy Points:
    63
    Location:
    Cinnaminson, nj
    finally...someone agrees how ghey led zeppelin is.
     
  18. Ash1412

    Ash1412 Friend

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    667
    Trophy Points:
    93
    Happened to open up this thread and had a realization. This discussion is directly relevant to the Schiit MP. This is what Mike said about it:
     
    Last edited: Apr 21, 2017
  19. GoodEnoughGear

    GoodEnoughGear Evil Dr. Shultz‎

    Pyrate
    Joined:
    Oct 25, 2015
    Likes Received:
    3,070
    Trophy Points:
    113
    Location:
    Cape Town, South Africa
    I'm necro-ing this thread because I'm breaking my head on bit depth.

    I get that each additional bit we add doubles the available number of values for sampling. 1 bit = 2 values, 2 bits = 4 values, 3 bits = 8 values and so on. And the more bits we have the more we reduce the degree quantization error.

    Where I'm getting lost a bit is in how it translates to the physical world. Given the spreadsheet below, I can see how 26-bit equates to 96dB of range and 24-bit to 144dB.

    Untitled.jpg

    Given a fictitious orchestral piece of 84dB SPL (0dB to 84 dB)range, if encoded in 24-bit with the SPL peak near/at 0dbFS do I effectively reach silence by bit 15 as in the sheet above? So I am basically just not using bits 15 to 24? Or am I using all those bits, and I have finer degradations (higher actual resolution), ie: I am chopping that 85dB range into 1677216 pieces vs 16384?
     
  20. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,777
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    db is relative scale, so you are using finer gradations. However, it's finer gradations of noise limited by microphones and microphone preamps.

    As for silence, since it's a relative scale, there is really is no total zero silence (even when bits run out), but there is effective silence for humans, especially considering environment and ambient noise.
     

Share This Page