Problems with DSD_U32_BE

Need help with MPD?
max
Forum team
Posts: 1029
Joined: January 15th, 2013, 3:43 pm

Re: Problems with DSD_U32_BE

Post by max »


J.L.C.
Posts: 27
Joined: January 9th, 2017, 10:12 pm

Re: Problems with DSD_U32_BE

Post by J.L.C. »

Could it be in the updates he did to the plugins (DsdiffDecoderPlugin.cxx, DsdDecoderPlugin.cxx, and DffDecoderPlugin.cxx)?

max
Forum team
Posts: 1029
Joined: January 15th, 2013, 3:43 pm

Re: Problems with DSD_U32_BE

Post by max »

I'm not yet sure. The changes to existing plugins are serious bugs, but they are nullified by serious bugs in other parts of MPD that guy added. Did I already express how much I dislike this obscure coding style? ;-)

J.L.C.
Posts: 27
Joined: January 9th, 2017, 10:12 pm

Re: Problems with DSD_U32_BE

Post by J.L.C. »

max wrote:I'm not yet sure. The changes to existing plugins are serious bugs, but they are nullified by serious bugs in other parts of MPD that guy added. Did I already express how much I dislike this obscure coding style? ;-)
I ran into a few problems trying to compile it; it took some hacking and commenting out some code blocks.

I'd much rather use the main/vanilla trunk!

Thanks again for all of your fantastic work on this project.

max
Forum team
Posts: 1029
Joined: January 15th, 2013, 3:43 pm

Re: Problems with DSD_U32_BE

Post by max »

So, now I went to great lengths, and compared MPD.git with this xxxbugxxxx branch. I don't have any DSD_U32 hardware, but I added some debug code to both which fakes a DSD_U32 capable DAC, but instead of playing it, it writes the data to a file, and observes the ALSA hw/sw parameters.

What I got was the worst possible outcome: everything is completely identical! Bit by bit!

Once identical data with identical parameters gets passed from whatever process to the ALSA kernel driver, in my understanding there cannot possibly be any audible difference. But you say that vanilla MPD sounds worse... and I have no idea how this can be possible.

J.L.C.
Posts: 27
Joined: January 9th, 2017, 10:12 pm

Re: Problems with DSD_U32_BE

Post by J.L.C. »

I'll recompile and try again using a different machine.

Thanks again!

J.L.C.
Posts: 27
Joined: January 9th, 2017, 10:12 pm

Re: Problems with DSD_U32_BE

Post by J.L.C. »

Downloaded again from the current master branch on both my desktop and laptop then compiled on both....

No hiss!

Working perfectly on both machines!!

Solved!

max
Forum team
Posts: 1029
Joined: January 15th, 2013, 3:43 pm

Re: Problems with DSD_U32_BE

Post by max »

Hooray! 0.20.2 coming really soon.

J.L.C.
Posts: 27
Joined: January 9th, 2017, 10:12 pm

Re: Problems with DSD_U32_BE

Post by J.L.C. »

In case anyone else finds their way here, my Encore mDSD is also working with 0.20.2, as long as I set: dop "yes" in .mpdconf. The mDSD shows the data as DSD (blue light) and sounds awesome.

Perfect playback, everything sounds great!

Many thanks for the rapid development!

J.L.C.
Posts: 27
Joined: January 9th, 2017, 10:12 pm

Re: Problems with DSD_U32_BE

Post by J.L.C. »

This is likely outside the scope of MPD development, but is there anything I can do to get full functionality from the Encore mDSD?

Is it possible to do something with modprobe or alsa.config to indicate the device is capable of native DSD?

lsusb shows:

ID 16d0:09dd MCS

and aplay -l shows:

card 3: mDSD [Encore mDSD], device 0: USB Audio [USB Audio]
Subdevices: 1/1
Subdevice #0: subdevice #0

But, it appears to only be working with DoP.

Even though it's using an XMOS chip, the vendor and device ID's don't match up with what is currently in the kernel. Is it just a matter of patiently waiting for it to be added, or is there anything I can do?

It supports USB Audio Class 2.0, but will it require modules to be compiled for this particular vendor and device ID?

Can I somehow "spoof" it as another XMOS device that is currently supported?

Post Reply