Takashi Iwai wrote:
> At Tue, 13 Feb 2007 18:36:35 +0100,
> Karsten Wiese wrote:
>> Hi Rui,
>>
>> Am Dienstag, 13. Februar 2007 schrieb Rui Nuno Capela:
>>> Hi Karsten,
>>>
>>> In my quest in completing us428control implementation, I'm now stumbling
>>> on a fundamental question regarding the number of tracks/channel-faders
>>> (per bank) each particular usx2y device supports. The question boils down
>>> to what is the actual numeric value of the y in the actual usx2y model
>>> name :)
>>>
>>> In practice, I need to know which usx2y model is actually connected.
>>>
>>> So, what is the best/deterministic way to detect (programmatically)
>>> whether us428control is connecting to a US224 (y=4, as mine) _or_ a US428
>>> (y=8, as yours) ?
>>>
>>> Is snd_hwdep_info_get_name() expected to be the most correct method?
>> Yes, I fear. this gives the /proc/bus/usb/002/004 filename.
>> I just did cat /proc/bus/usb/004/005 > x
>> and check x with hexedit: the 1604 and 8001 is in there.
>> I don't know how to portably interpret a read from
>> /proc/bus/usb/004/005 though.
> 
> Note that /proc/bus/usb is deprecated on many distros...
> 

How about using snd_pcm_hw_params_get_channels_max() ?
At least on capture the US224 should give 2 and the US428 should give 4
as the maximum number of channels.

It would be handful for me if the attached test program could be ran
against a US428, just for the records ;)

On my US224 it gives:

$ snd_hwdep_info hw:3
---hwdep---
snd_hwdep_info_get_id   (hw:3) = 'USX2Y Loader'
snd_hwdep_info_get_name (hw:3) = '/proc/bus/usb/001/004'
---pcm (capture)---
snd_pcm_hw_params_get_channels_min(hw:3) = 2
snd_pcm_hw_params_get_channels_max(hw:3) = 2
---pcm (playback)---
snd_pcm_hw_params_get_channels_min(hw:3) = 2
snd_pcm_hw_params_get_channels_max(hw:3) = 2

Karsten?
-- 
rncbc aka Rui Nuno Capela
rncbc@rncbc.org