Re: [Alsa-user] Fwd: snd-usb-audio - USB DAC can't set sample frequency

[Daniel M. <zo...@gm...>]

Hi,

On 12/18/2013 11:54 PM, Michal Šmucr wrote:
>> I was about to recommend you look at the mixer values in case there are
>> any, but the display detail actually let's me assume that there's
>> something wrong in the setup sequence.
>>
> 
> There were no controls in mixer. At first i also missed some typical
> audiocontrol device descriptors like typical CLOCK_SOURCE and
> CLOCK_SELECTOR subtypes. But after looking at USB dump from Zodiac
> Gold, which was working flawlessly, there are also no controls and
> almost same values at rest of descriptors. Only visible difference is
> declared bLockDelay, but this could be just due to hardware
> functionality difference and maybe it is not necessary for Platinum.

We don't have support for bLockDelay in Linux currently, but that
shouldn't cause any trouble in your case, as it's set to 0 anyway.

> I'll try to settle for DAC loan for weekend.. I'll be more comfortable
> with its further testing at my place. Unfortunately I don't have any
> lower level hardware analyser like Beagle. Do you think USB
> communication dumps in pcap format from usbmon module on Linux and
> USBPcap on Windows will be enough for checking and comparison at least
> for initialization stage?

Yes, that should do. You should at least see any non-standard compliant
messages that are transferred. Another thing you could have a look at is
the message ordering and timing. The device might be picky about that,
who knows. Also see the delay quirks I added for another vendor, maybe
you need something like this as well (0959f22ee6).

> Aside from that, i found your test tool for playback of DSDIFF files
> to ALSA at github. I wasn't really aware, that recent ALSA library
> have internally defined DSD format and could pack DSD frames to PCM at
> audio driver by itself, so it is not necessary to do it in
> application, that is cool.

I added this in 3.10, yes.

> Maybe i could later do also test with that
> tool, only thing will be replace hardcoded device name for hw:MPD3 and
> adjusting defines either for 64fs or 128fs DSD input file.

Yes, we need more testers around here, and support for more devices. If
you can have a look at this, that'd be much appreciated :)


Thanks,
Daniel