Simple DSD modulator for DSC2

Jump to Latest
#121
chientechnical said:
Tks u. I'm refreshing the 4x DSC parallel different balance (or 8x).
I used LT3045 regulator for all DSC, Amanero, and isolator too. Crytek clock clean are on DAC side with support of Arduino count frequently for switch select 44.1/48Khz family and mute function, and isolated too.
How the Arduino can detect the PCM or DSD being transmitted. Sorry I'm my-learn from the Internet and have a bit of a hobby for self-made audio.
Click to expand...

what do you mean by "4x DSC parallel" ?
who has made that board ?
 
#124
pjotr and olek said it works... so why doubt ?

my problem is only implementation-wise for my current setup... as many people pointed out, should be only a problem for my personal non-senses... hence why i needed to design a "gating" board for the outputs.

You will be using the module "as-intended", so i think you are good go 🙂

Be aware of some "juggling" needed to properly flash the memory for the fpga. but everything is online. If i was able to sort it out, you surely can 10x better than me.

Enjoy
 
Last edited:
#125
A question might be whether or not either of the FPGA codes posted so far can sound as good or better at PCM to DSD conversion verses an expertly implemented AK4137 converter?

Also of possible interest, one may recall that Chord claims to use a 17th order modulator (they call it a 'noise shaper') in their best dacs to get audio band noise as measured in the digital domain down to -300dB. They claim if any higher than that then its effects are still audible. They don't claim to know why, they only claim it as an empirical observation from listening tests.
 
#127
Understood.

John Westlake claimed AK4137 DSD conversion sounded better to him than anything from HQ Player.

However, I would note that AK4137 is an asynchronous converter, so it takes some care to keep the PLL very stable for best results.
 
#129
HO Player sounds impressive, but unnatural - at least to me. Especially if I am not doing critical listening, i.e. if I am doing something else with the music playing in the background. This combination does not let me focus fully on what I'm doing.

I'd rate AK4137 implementation inside May DAC quite highly and I prefer its sound much better than HQ Player or any other upsampling done "externally"

To clarify: I like original files, processed by May NOS implementation, most of the time. However, with some 44.1kHz material, I feel that the high-frequency spectrum is lacking somewhat (due to that high-frequency roll-off). Files recorded with higher sample rates sound great in NOS mode - nothing else is required here and NOS mode sounds just amazing.

The best job at overcoming that roll-off (with 44.1kHz files) is done in fact by AK4137 implementation inside May DAC. It corrects the high-frequency roll-off while not changing the sound in any other way. I can still perceive the depths of the sound stage and the natural harmonics like I do with true NOS. HQ Player obliviates both, but it does sound "impressive", especially with the 1.536MHz PCM / ASDM7EC DSD modulator. I suppose it's good to have choices... but ultimately, the closest to an analog record I can get is with the original unprocessed files (especially if recorded at 88.2 / 96kHz or higher sample rates), using NOS mode.
 
#133
Agreed it is not like ESS. But ASRC with PPLL still only attenuates jitter, it doesn't eliminate it. The question then that interests me is how low does incoming jitter have to be, and how low does power supply noise have to be before no further audible improvement can be detected by a panel of trained listeners. IME, both incoming jitter and PS noise have to be VERY low before a good ASRC such as AK4137 audibly sounds its best.


EDIT: I might mention that AK4137 has multiple operating modes. It must have a PLL when operated without a crystal MCLK and when it set to act as I2S slave to the dac that follows it. IME that mode doesn't sound very good. Don't know if anyone else (besides John Westlake) has spent much time finding out things that are not in the data sheet.
 
Last edited:
#134
Nowadays ASRCs usually have some kind of sample rate ratio estimating loop which is almost entirely digital, although you could argue that the points where a clock gets sampled with some other clock are not. In any case, it is nice when everything is synchronized to the DAC clock, so you can do without such loops because you know exactly what the ratio is.
 
#137
Markw4 said:
Agreed it is not like ESS. But ASRC with PPLL still only attenuates jitter, it doesn't eliminate it. The question then that interests me is how low does incoming jitter have to be, and how low does power supply noise have to be before no further audible improvement can be detected by a panel of trained listeners. IME, both incoming jitter and PS noise have to be VERY low before a good ASRC such as AK4137 audibly sounds its best.


EDIT: I might mention that AK4137 has multiple operating modes. It must have a PLL when operated without a crystal MCLK and when it set to act as I2S slave to the dac that follows it. IME that mode doesn't sound very good. Don't know if anyone else (besides John Westlake) has spent much time finding out things that are not in the data sheet.
Click to expand...

But I don't like music from ESS DAC (9038). It's bad compared to my HQ Player + Amanero + DSC2 output by OPT.
So, I want PCM2DSD for use the SBC as Rpi on Sever/Player side. It's cheep price and low power more.
I want to listen PCM2DSD this, then compare to HQ-Player and AK4137 Kit. This important more than all.
 
Last edited:
#138
Many people around here seem to complain about the ESS sound. Its better in fully synchronous DSD mode. IMHO it can be tweaked a bit the make the sound a little better in that mode by small changes in MCLK timing relative to BCLK, although ESS says it shouldn't matter. Moved it in 150pS steps and thought one step sounded best, but it was only only a slight effect. In any case, most of us seem to have moved on from ESS. Guess Rohm will be the next thing to try 🙂
 
#139
Even so, all No Chip DAC (R-2R or R Ladder) projects up to now are rated higher than Chip DAC. And the cost for the 74xxx595 logic chip and the resistor is significantly cheaper, regardless of the manufacturer, easily customized to suit the music ...

I just want to upgrade the sound quality to DSC2 for good more. I think End-Point-DSD, same as PCM2DSD is appreciated and can easily change, update digital filter algorithm. Hopefully, coders can use this platform to provide the community with many alternatives.
 
You must log in or register to reply here.
Top Bottom