Dice firewire support

Melvin Vermeeren mail at mel.vin
Fri May 22 13:36:13 CEST 2020


Hi all,

On Friday, 22 May 2020 11:07:06 CEST Pascal Bomel wrote:
> Thank you so much Takashi.
> Actually i have a debian 4.9.12 "stretch" release with 4.9.211 kernel
> version (uname -v)
> This kernel has snd-dice module inside isn't it ?
> I am not very familiar with linux, that's why i appreciate your help.

You can either upgrade to Debian buster, which has 4.19, or install a kernel 
from stretch-backports, which is also 4.19. Assuming you have amd64 hardware 
that would be https://packages.debian.org/stretch-backports/linux-image-amd64.
See also https://backports.debian.org/Instructions/ for more instructions.

Some background information below, from how I remember things. Do note that 
there were many details and I may be mistaken about some of them.

At the time of 4.18 there was some issue with the timing that caused non-bit-
perfect playback, resulting in audio artefacts that were hard to detect as it 
was near perfect. DSD over PCM (DoP) mode however made it very obvious this 
issue existed, as it would not work reliably at all. It was also noticeable 
when listening to a fixed Hz tone, it wasn't perfectly smooth.

I believe there was some significant refactoring going on with snd_dice at the 
cause, which caused some previously working things to stop working as snd_dice 
became more compliant with device's listed features and protocols.

Older versions "just sent a new stream" when frequency changed (iirc), which 
actually caused the DAC to sometimes not catch the signal and made it silent. 
You then had to restart jack (or ALSA stream) a few times. The new version 
used proper lock mechanism (AMDTP?) meaning it worked 100% of the time, except 
for the non-bit-perfect issue.

Since those days of experimenting I have been running snd_dice and 
snd_firewire_lib from sound-4.4-rc7-692-g2eb65d67afbf out of tree, with minor 
patches to make them compatible with modern kernels and overrides for the DAC.

Thanks for the mention Takashi, I appreciate it. Since I am running 5.5.17 
(buster-backports) kernel right now I will do some testing later today to see 
if there are still any issues in this version, the Mytek is still used daily.

Cheers,

-- 
Melvin Vermeeren
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: This is a digitally signed message part.
URL: <http://mailman.alsa-project.org/pipermail/alsa-devel/attachments/20200522/de265ed4/attachment-0001.sig>


More information about the Alsa-devel mailing list