Digital that sounds like analog - diyAudio
Go Back   Home > Forums > Blogs > abraxalito
Home Forums Rules Articles diyAudio Store Gallery Wiki Blogs Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Rate this Entry

Digital that sounds like analog

Posted 10th December 2012 at 04:18 AM by abraxalito
Updated 15th February 2013 at 06:11 AM by abraxalito

For those who missed Frank (fas42)'s link on a thread I started then here's where I'm continuing my minimal oversampling DAC developments for the time being : Digital that sounds like analog
Views 3902 Comments 108
Total Comments 108


  1. Old Comment
    abraxalito's Avatar
    I agree that virtually all DACs are broken - which is what's led me to design my own
    Posted 19th December 2012 at 01:27 AM by abraxalito abraxalito is offline
  2. Old Comment
    Another interesting round of chatter on the 'Far Side', very pleasing to see the level of the discussion ...

    With regards to the bass being more intense, this is a positive sign, meaning underlying distortion is reducing. This will go hand in hand with the treble being more distinctive, more clearly part of the picture. The last thing one wants is for treble to be smooth, real sound is not smooth, it's intense, and if that is not conveyed correctly then the sound will never quite get there. When you talk of colour I would translate that to the sound having a more natural 'ring' to it, which is exactly what one should be after ...

    I have noted that 'Bitstream' issue many times, yes, those recordings are definitely quite a bit more troublesome -- I have not the knowledge, haven't bothered to try to analyse the underlying elements in the encoding to understand it better -- but fear not, they can be rendered in full glory, they will emerge, complete, into the sunshine! Another example of 'difficult' recordings being your friend, in this journey ... :)

    I like your idea of the polarity issue being as a result of the asymmetry of the signals being fed through the rest of the circuitry, I hadn't seen that before, and I see it as making a lot of sense ...

    The silver vs. copper thing, I see a lot in Don's comment about the interaction of the wire and components, his quote from the ad copy about the "hardness" being relevant I do agree with, but not for the silly explanation then tacked on.

    Posted 21st December 2012 at 02:11 AM by fas42 fas42 is offline
  3. Old Comment
    abraxalito's Avatar
    Yes, tonal colours - they are always there but with noise modulation they get 'greyed out' - additional noise confuses our ear/brain and we don't separate out the wanted (instrument, voice) from the background noise which is correlated with it, the result being we perceive a noisy voice/instrument. This gives a washed out perception.

    As for Bitstream recordings - noise modulation can't be removed, those leading edges are lost forever due to choice of wrong ADC. Analog tape has a similar effect on high level piano transients I've noticed but subjectively less objectionable - it adds softening rather later on the attack, subjectively masked slightly.
    Posted 21st December 2012 at 03:17 AM by abraxalito abraxalito is offline
  4. Old Comment
    I agree that the problem is that what you call noise modulation, what I would just call distortion, is what makes digital reproduction difficult, and when the job of separating is too hard for the brain then 'greying', 'washing out' occurs. However, I have found that the musical content in the recording is [I]still[/I] sufficient for the ear/brain to do the job of filtering comfortably, if enough effort is made on refining the sound. Even for Bitstream. I don't know what your experiences are in this area, but I find Yehudi Menuhin, especially the later stuff, a very severe test for extracting decent 'tonal colour'. So, he and Bitstream are a dynamite combo; if one gets it right there then you've earnt your oats ...

    Meaning, I dispute you saying leading edges are lost, for Bitstream. It may be the ear/brain filling in the gaps, I don't know and won't claim to understand precisely how it works ... but I do know if I make sufficient effort then I do experience the full dynamics and impact with those recordings ...

    I might mention now that I'm doing some experiments using relatively crude recording techniques, a cheap digital camera with HD video, records 16 bit PCM, noisily, no level control, but with a bit of luck well enough to demonstrate some of the points I've been making in my posts. Preliminary fiddling is looking promising -- once I have some material that hangs together well enough I'll post it in some area of the net, hopefully not losing too much more quality on the way ...

    Posted 21st December 2012 at 04:29 AM by fas42 fas42 is offline
    Updated 21st December 2012 at 04:32 AM by fas42
  5. Old Comment
    abraxalito's Avatar
    You listen on bitstream DACs right? So how do you know you "experience the full dynamics and impact" ?

    I am calling your bluff here Go look at the paper I just linked to at the other place and tell me your DACs have all the issues in regard to noise modulation licked.
    Posted 21st December 2012 at 07:18 AM by abraxalito abraxalito is offline
  6. Old Comment
    I'll raise the bid, and request the link to the paper that "proves" that noise modulation actually causes audible artifacts; here I'm not talking about getting a, say, -80dB encoded signal and showing that there is a measurable difference caused by this "noise", but rather that a typical music recording is perceived differently when noise modulation at the levels that are typically measured are in the picture.

    Which means, not good enough to say that S-D DAC sounds different, therefore it must be noise modulation causing it. Someone needs to "prove" that other interference and distortion mechanisms are not the root cause for variations in sound ...

    I've looked at the paper, and nowhere is there any linking of the "problems" of S-D with audible impacts; the best is that [I]perhaps[/I] it [I]can[/I] alias back into the audio band.

    As regards the listening experience, I like to experience the real thing on a regular basis, that's what I base my judgement call on. Judging by what the "analogue" sounded like at the recent hifi show I attended, that's a pretty low bar to be aiming at with what you're doing ... :D

    Posted 21st December 2012 at 09:52 AM by fas42 fas42 is offline
  7. Old Comment
    abraxalito's Avatar
    Its enough that I hear it, no proof required. Correlation I agree is not the same as causation, but then I'm not condemning S-D, just avoiding it. S-D DAC sounds inferior, the best hypothesis for that is noise modulation. No proof. I did not base my 'digital that sounds like analog' on the analog sounds that you heard, it was the closest descriptive metaphor that I could come up with at the time. Its simply much more succinct than 'digital minus the fatigue and harshness'

    I am of the view that digital can indeed do so much better than analog, but in my designs I won't be attempting that feat with S-D DACs, there's already enough challenge in this without tying one hand behind my back and putting ground up glass in my shoes
    Posted 21st December 2012 at 01:20 PM by abraxalito abraxalito is offline
  8. Old Comment
    That's fair enough ...

    Yes, you're creating a path into the headspace of people and so using a tag that resonates with many is a good move. I [I]have[/I] heard excellent analogue over the years, so in that sense was surprised by the somewhat tedious quality of nearly all examples on display at that show.

    My "hypothesis" is that the S-D chips have too much high speed circuitry as part of their intrinsic nature of operation, and that effectively pollutes their own nest; they don't need external interference to cause issues, they quite happily provide enough in their own right.

    That said, simple measures help a great deal: these converters need to run continually so that the circuitry is fully stabilised at all times; and there is an optimum data stream format for best sound. Just yesterday was doing a lot of playing around on the PC, with YouTube, and resampling tracks, and the dramatic improvement when feeding the converter a hi-res version of a 'mediocre' sample is so obvious.

    Posted 21st December 2012 at 10:06 PM by fas42 fas42 is offline
  9. Old Comment
    abraxalito's Avatar
    I agree with your hypothesis about S-D converters - in my own quest for sound nirvana, slower has always been better. My last change was from 2X OS with a digital anti-imaging filter, back to NOS with a passive anti-imaging filter. More tonal colours resulted, more dynamics apparent. So making things run faster is, IME always a recipe for greyness.
    Posted 22nd December 2012 at 12:39 AM by abraxalito abraxalito is offline
  10. Old Comment
    Richard, a sometime poster over there, Robert, has been an enthusiastic tweaker, understands where I'm coming from; and fairly recently posted a number of YouTube clips of his system running. One of those clips in particular has a certain element in it, that indicates something to me: [URL=""]Avalon String Sextet - YouTube[/URL]

    Any thoughts or comments ...?

    Posted 22nd December 2012 at 09:03 PM by fas42 fas42 is offline
  11. Old Comment
    Just a further thought, glanced at that Devialet thread and saw:

    [QUOTE]But I can confirm your point about the extreme sensitivity of the Devialet to power cables, conditioners and even digital cables - it would be a nice system to convert people from the anti-cable brigade![/QUOTE]Welcome to the world of high quality digital ... this is just the way it is, at the moment! I made this "discovery" over 25 years ago, and this was with classic R2R technology, :p :p ...

    My experience over the years has drilled into me the need to be obsessive, totally obsessive about such matters; otherwise, you're wrestling with a python ... :)

    Edit: the fact that some people find the Devialet brilliant, and others, that its sound is quite mediocre, is also a perfect giveaway of the "true" situation. Digital sound, for me, has alway been " ... , when she was good she was very, very good, and when she was bad she was hor... ".

    Posted 22nd December 2012 at 09:55 PM by fas42 fas42 is offline
    Updated 22nd December 2012 at 10:04 PM by fas42
  12. Old Comment
    Funny little conversation I'm having here ... ;)

    If you want strings, this is a recording that should sound staggeringly good, in every possible way: [url=] Vivaldi: The Four Seasons: Music[/url]

    Posted 22nd December 2012 at 10:15 PM by fas42 fas42 is offline
  13. Old Comment
    abraxalito's Avatar
    I predict I would not like the Devialet - noise modulation. I was surprised though to see the pic on that thread of my old colleague Laurence Dickie - the designer of the Vivid brand speakers, playing with it feeding his Giyas. I'd had imagined he'd know better
    Posted 23rd December 2012 at 09:15 AM by abraxalito abraxalito is offline
  14. Old Comment
    abraxalito's Avatar
    Originally Posted by fas42
    Any thoughts or comments ...?
    I'm listening now but my laptop's speakers are probably the worst speakers in my place - the sound is fairly dynamic to my ears, probably this isn't DSD but its really a guess given the transducer limitations...
    Posted 24th December 2012 at 05:26 AM by abraxalito abraxalito is offline
  15. Old Comment
    Yes, this is classic Redbook, coming from a Wadia tweaked by a highly regarded exponent of such doings. However, what I'm referring to is whether you detect a change in the quality of the reproduction as the clip progresses, from the beginning through to about 2/3'rds in. I have very basic Harmon Kardon powered speakers to monitor this, on a commercial quality HP Compaq desktop, but the characteristic I'm concerned about is very clear to me ...

    Posted 24th December 2012 at 06:37 AM by fas42 fas42 is offline
  16. Old Comment
    [QUOTE]It does mean that for [B]some[/B] playback there exists a relaxation, a coherence that I do not find with digital recordings.


    So why is it that digital - good digital at that - has so much trouble with this "inner detail"? [/QUOTE]Digital [I]is[/I] capable of creating very intense, yet relaxed replay; which goes hand in hand with revealing "inner detail" - it's all part of the same package. But it's not terminally linked to the S-D "problem" - otherwise recordings done with dreaded Bitstream ADCs could not be lifted to this performance level.

    I'll aim to get my primitive recording setup to a point where it clearly shows retrieval of fine, musical detail -- S-D replay of a S-D recording -- to point this out ...

    Posted 26th December 2012 at 11:06 AM by fas42 fas42 is offline
  17. Old Comment
    [QUOTE=fas42;bt2045] However, what I'm referring to is whether you detect a change in the quality of the reproduction as the clip progresses, from the beginning through to about 2/3'rds in. I have very basic Harmon Kardon powered speakers to monitor this, on a commercial quality HP Compaq desktop, but the characteristic I'm concerned about is very clear to me ...

    Okaaay, time's up!! :D

    The answer I was looking for is that there is a marked deterioration in SQ as the clip progresses: at the very beginning the sound's fine, string tone is crisp and clear, but from then on it starts going downhill, the sparkle steadily erodes and a murkiness, or lack of clarity and acoustic separation between the sound elements becomes worse and worse.

    Can anyone pick this? This has always been the Achilles Heel of digital sound for me, and eliminating this behaviour is one of the key struggles in achieving high quality audio playback.

    Posted 3rd January 2013 at 11:26 AM by fas42 fas42 is offline
  18. Old Comment
    abraxalito's Avatar
    Perhaps admitting to the S-D problem (as contrasted with "problem") would be a solution to the loss of sparkle you encounter?
    Posted 7th January 2013 at 06:43 AM by abraxalito abraxalito is offline
  19. Old Comment
    I know you enjoy pointing the finger at that black hatted, moustache twirling villain, :), but, 'fraid it's not S-D in origin, this is generic digital behaviour -- though I have to admit the latest units that I came across at the recent hifi show were doing better. The system in that clip uses a Wadia processing unit, which runs 8 B-B PCM1704 chips per channel, "in a balanced, time-shifted array" -- sound vaguely familar ... ;)? My first headaches involved a Yamaha unit, built with ancestors of those IC's ...

    No, this is in part due to earthing issues, and static buildup, and myriads of other nasty gremlins -- it's a hard life being an audio perfectionist ... :D

    Posted 7th January 2013 at 07:57 AM by fas42 fas42 is offline
  20. Old Comment
    abraxalito's Avatar
    Ah, OK then let's rule out S-D from this. What other things can change over those timescales of minutes? Could it be transformers and common-mode noise issues? CM noise is related to earthing. Does the effect occur regardless of the level being played out or only happen when playing above a certain volume?
    Posted 8th January 2013 at 03:07 AM by abraxalito abraxalito is offline
Hide this!Advertise here!

New To Site? Need Help?

All times are GMT. The time now is 01:38 PM.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright 1999-2017 diyAudio

Content Relevant URLs by vBSEO 3.3.2