Interesting white paper on digital phase filters!

earwaxxer · 2280

0 Members and 1 Guest are viewing this topic.

Offline earwaxxer

  • Hero Member
  • *****
    • Posts: 1336
on: August 25, 2012, 06:53:15 PM
 I have been playing with this stuff for awhile now in software (foobar). Freeware at that! See! I'm not THAT crazy! Everything I have been listening to for the past couple of years has been upsampled to 24/96, min phase, dither applied and allow for aliasing. When 24/96 is played through the Transporter I am told that the DAC leaves it alone (NOS). I may be wrong on that. I have tried different sample rates (max. is 24/96), and phase settings. Current settings are so far the best. I didnt mention replay gain, but that has also tamed transients in a major way.  All I know is the sound is mucho bueno! I have tried several times to switch back to native redbook playback - no way. By comparison, the sound of native is kind of disjointed and unnatural. Not pleasant. I wouldnt know its off unless I heard the comparison. I also revisited my MSB Link DAC lll. It also sounds fantastic with the upsampled data. I think it may sound better than my Transporter. It always did sound good (natural), just a bit 'dated' and 'unsophisticated', in comparison to the modern quiet, resolving, and detailed Transporter.

 Anyone else playing with this software?

Eric
Emotiva XPA-2, Magnepan MMG (mod), Quickie (mod), JRiver, Wyrd4sound uLink, Schiit Gungnir, JPS Digital power cord, MIT power cord, JPS Labs ultraconductor wire throughout, HSU sub. powered by Crown.


Offline earwaxxer

  • Hero Member
  • *****
    • Posts: 1336
Reply #1 on: August 30, 2012, 05:32:14 AM
Hey Tim - I totally agree with what you are saying! Its a bit counter intuitive and somewhat of a phenomena of sorts, but I think an explanation of why software may be 'better' may have to do with the algorithms used. Hardware upsample has to happen in real time, software does not. I would have to assume that the code (algorithm) for each uses different techniques. Not being a coder I would have to defer on that. The second would be how the DAC chip itself handles 'native' redbook and high res. material. I know that the AK4396 in the Transporter handles them differently (normal, dual, and quad speed mode for 44.1,96,192khz). IMO its all about the filters. With software you can control them. With some very recent DAC's you can select for filters (slow, fast etc.). With the Transporter I can control the AK4396 hardware filter via squeezbox server (either slow or fast). There is a bit of a change in sound from a bit sharper with 'fast' to a bit more analog or smoother with 'slow'. With this white paper, and other signs of adoption of these concepts in DAC makers, I think we are seeing an evolution of thinking on digital sound. We can thank the use of computers as music server for that.

Eric
Emotiva XPA-2, Magnepan MMG (mod), Quickie (mod), JRiver, Wyrd4sound uLink, Schiit Gungnir, JPS Digital power cord, MIT power cord, JPS Labs ultraconductor wire throughout, HSU sub. powered by Crown.


Offline earwaxxer

  • Hero Member
  • *****
    • Posts: 1336
Reply #2 on: August 30, 2012, 02:51:56 PM
Unfortunately, I think we are still getting 'crappy resampling'! I dont think its intended to be crappy, its just a way of doing things thats hard to change. Same thing with going from CD players to PC based music server. Its a change of thinking. It doesnt make sense, in some ways of thinking, that a PC could sound better than a CD player. The experience says that they can, and do sound better. It requires a different way of thinking.

Eric
Emotiva XPA-2, Magnepan MMG (mod), Quickie (mod), JRiver, Wyrd4sound uLink, Schiit Gungnir, JPS Digital power cord, MIT power cord, JPS Labs ultraconductor wire throughout, HSU sub. powered by Crown.