will
|
I had an interesting set of experiences.
When I first got my 2012 4 core Mini i7, I had been running with a Jitter Bug followed by a 2nd gen (or 3rd...not the iso) Regen, followed by a short Curious Cable. I liked DSD resampling in Audirvana to 11.2, or DSD 256 at the time. My modified Gustard x20pro does not up sample, so can't compare software to hardware, but thought Audirvana did a very nice job with a carefully adjusted filter balance.
Then, doing amp mods I knew would take a long time, I decided to recheck. To keep consistency through a long mod cycle, what was my best sound to stay with? I was always a little on the fence about whether my DAC was better with DSD or PCM, and finally determined I liked 352.8 PCM best, upsampled from Redbook in Audrivana, and the same basic Audirvana settings I had liked for DSD resampling. The DSD was more apparently clean, but the PCM was plenty clean, and more textural and complex, more musical to me.
With a Mac using DoP to transfer DSD, the DSD has to be packed in a PCM stream if I get it correctly. My top DSD level was 256, and since DoP uses a lot of PCM data to carry the DSD, my computer was running its fan a lot. This stuff messes with my dyslexia, but I think it goes something like this Pal. Native DSD 512 would be 22.6 MHz, 1 bit x 44100 Hz x 512 = 22.6 GHz.... which is 2 x 11.2 MHz/DSD 256. For DoP, with a word length of 16 bits used to pack a DSD bitstream, 352.8K x 16 = 5.6 MHz, or DSD 128. So for DoP, the setup needs to be good with 352.8K PCM to create and utilize DSD 128. In turn, 705.6K PCM allows for DSD 256 (11.2 MHz) via DoP and so on. Does that sound right?
I figured all this processing had to have some effect on the sound, though it sounded surprisingly good. But perhaps this helps explain why I was getting more complex sound from the less computer intensive 352.8K PCM? Don't know, but I kept liking it better for the most part.
Then, as Torii IV and CSP3 modifications progressively presented a more complete and complex signal, more fine detail and spacial information changed the picture, refining harder detail into musically complex detail, the Regen setup started sounding worse to me...sort of metallic and mechanical. Different now, but familiar, this was the reason I had added the Jitter Bug originally.
I started liking a prototype USB cable Eric Hider sent me to test best (not his final, but a really nice cable), or my own DIY USB cable, preferring each without the Regen and/or Jitter Bug. Both were made with power and ground separated from data, and I had used both with the Regen, sometimes liking one better, and sometimes the other. Mind you, at this point I had changed to Eric's dBAudioLabs Revolution OS, which was a very nice refinement for digital disappearing into music here, in large part because it was more revealing of subtle information that was there in the files. The OS more revealing, and amps more revealing, the USB treatment that once seemed beneficial sounded off.
This showed up many months ago as I was re-tuning/refining the power supply bypasses in the amps. I had decided to pull all very small bypasses in order to get the best foundation I could with 0.1 uF being the smallest for the most part, and things were getting really good. Finally balanced and beautiful, along with attenuators, resistors, wires, connectors, all was quite good enough to stop. But I knew from before I could get more from bypassing the bypasses. Many tests later, in one position I might have two complimentary 0.1 uF caps that together sounded better than 0.22s I tried alone. Then I started to reveal more very fine detail and spacial info with a single, or combinations of 0.0047, 0.01, and 0.022 bypassing the primary bypasses for a given power supply position.
During this process, as the amps continued to be more revealing and musical, I went from Eric's very nice revealing, transparent, and balanced USB cable, back to my also revealing, but smoother and a little sweeter DIY USB cable. It is made with pure silver in cotton for data, and really good copper for power and ground, those wires separated from the data.
Not sure if any of this will help anyone else, but an interesting ride. As the OS and amps got better and better, I could not use the Regen anymore. I guess the moral of the story is that I consistently find it is best to recheck things now and then to see if previous decisions are still valid.
|