Thanks @bohrok2610 I understand the sentences.
How should look like a reclock after PCM2DSD? My Dac is NoDac ( passive high pass filter).
I have seen some modules fited with a good OSCO clock rated at 10Mhz and two output 25M or 54M
How should look like a reclock after PCM2DSD? My Dac is NoDac ( passive high pass filter).
I have seen some modules fited with a good OSCO clock rated at 10Mhz and two output 25M or 54M
Attachments
I would just use the same MCK that you use on PCM2DSD and reclock DCLK (possibly also DSDL/DSDR) with a flip-flop (e.g. 74LVC1G79). But as I said reclocking after PCM2DSD is not necessary so you can add it later if needed.
I almost forgot having a Singxer F1 module in the drawer, which includes the two necessary 22/24 clocks, plus the pins and mounting holes compatible with the i2s of the Opt-Toslink receiver. I would like to connect them both together, but I don't know if it will work or if the clocks will be effective that way.
Attachments
Back home after a car trip all long today, I have recovered my original board bought two years ago already populated, it was freezed by a middleman.
So no more need replacement, thanks @Cestrian for kind free board offering, and please apologise me the query here. I just wanted this board badly.
So no more need replacement, thanks @Cestrian for kind free board offering, and please apologise me the query here. I just wanted this board badly.
No worries, I'm glad you got sorted.
Always best to have a built unit. I've built 5 of these so far (one i destroyed by mistake) so i can appreciate.
I hope you like it. I am very impressed with it and the work Pjotr and Olo put in.
Always best to have a built unit. I've built 5 of these so far (one i destroyed by mistake) so i can appreciate.
I hope you like it. I am very impressed with it and the work Pjotr and Olo put in.
IIRC, the gain of the last build was adjust so that DSD rather than being at -6dB compared to PCM, the DSD converter would make it at -4dB. IIRC it was a compromise to try to get the best sound with hi-res recordings produced with some headroom, versus some popular music released on CD with a little bit of intersample overs. The idea was that PCM recordings mastered to be more loud should be attenuated a little before going into the DSD converter.
Also interesting that the HQ Player manual includes some advice/warning to leave some headroom when converting PCM to DSD. Apparently it can be difficult to predict exactly how much headroom will be needed.
Also interesting that the HQ Player manual includes some advice/warning to leave some headroom when converting PCM to DSD. Apparently it can be difficult to predict exactly how much headroom will be needed.
Do you mean that full-scale PCM becomes -4 dB DSD or that it becomes +2 dB DSD? By definition, the percentage of ones is modulated between 25 % and 75 % at 0 dB DSD.
IIUC its the former, whereas the standard is -6dB to allow for headroom (plus, IIRC, maybe an optional extra -3dB if more headroom is desired). Seems to me something about that is mentioned in the standard.
Last edited:
I think the headroom is in the definition of the 0 dB DSD level: 0 dB PCM is normally converted to 0 dB DSD, but as 0 dB DSD only modulates the percentage of ones from 25 % to 75 %, the signal can actually go a bit beyond 0 dB DSD. Not all the way to 0 % to 100 % (+6.02059... dB DSD) though, as that would make the sigma-delta modulator unstable (among other issues).
-1 dB as in somewhere between -0.5 dB and -1.5 dB or -1 dB as in fairly accurately -1 dB?
I would say between -0.95dB and -1.05dB. I base this just on looking at output level through PCM2DSD compared to output without PCM2DSD (i.e. DSF playback bypassing PCM2DSD). DSF file was created by sox-dsd from wav.file. I'm not sure if that conversion has exactly 0dB gain.
It's nice to present new firmware.
First of all, I'd like to thanks bohrok2610 for your tests and MarcelvdG for accurate tips.
The firmware is based on the previous version, so the specifications are similar.
There was a problem at low levels with additional peaks, as noted by bohrok2610:
https://www.diyaudio.com/community/threads/return-to-zero-shift-register-firdac.379406/post-7635705
Here is our solution:
We attach 4 versions with minor changes. You can choose whether you want to use ODDR2 primitives and what the channels layout is.
The channels layout has been adapted to the Pavel DSC or MarcelvdG RTZ DAC.
DIY use only!
First of all, I'd like to thanks bohrok2610 for your tests and MarcelvdG for accurate tips.
The firmware is based on the previous version, so the specifications are similar.
There was a problem at low levels with additional peaks, as noted by bohrok2610:
https://www.diyaudio.com/community/threads/return-to-zero-shift-register-firdac.379406/post-7635705
Here is our solution:
We attach 4 versions with minor changes. You can choose whether you want to use ODDR2 primitives and what the channels layout is.
The channels layout has been adapted to the Pavel DSC or MarcelvdG RTZ DAC.
DIY use only!
You choose with ODDR2 or without ODDR2, or try both.
The DSC or RTZ versions only have the channels swapped.
The DSC or RTZ versions only have the channels swapped.