Duncan Webb wrote:
> Lars Hanisch wrote:
>> Hi,
>>
>>   I'm playing a bit around with detecting installed video output 
>> devices. I'm having a PVR350, which has various devices:
>>
>> /dev/video0
>> /dev/video16
>> /dev/video24
>> /dev/video32
>> /dev/video48
>>
>>   If I'm querying the capabilities of these devices with 
>> VIDIOC_QUERYCAP, I get the same result for every device: 10702f3.
>>
>>   How can a program choose the right device, if it needs to write to the 
>> YUV-decoder without knowing that ivtv puts that at /dev/video48?
>>
>>   VIDIOC_ENUMOUTPUT and VIDIOC_ENUM_FMT also return the same for every 
>> device:
>>
>> output 0, type 2: S-Video + Composite
>> output 1, type 2: Composite
>> output 2, type 2: S-Video
>> output 3, type 2: RGB
>> output 4, type 2: YUV C
>> output 5, type 2: YUV V
>> format 0:32314d48: HM12 (YUV 4:2:0)
>> format 1:4745504d: MPEG
>>
>>
>>   What am I missing?
> 
> 
> AFAIK you're not missing anything, the only way I've found is to look in 
>    /sys/class/video4linux/*/name to find out what a specific device is. 
> As this is a string, it's not useful for applications.
> 
> My guess is that this is an oversight in the V4L2 API, that there is 
> doesn't seem to be a way to determine what a specific device is. In the 
> early days V4L only supported YUV formats.
> 
> Actually, I would like to be  able to determine what a specific device 
> can really do. It would be nice to have an ALSA driver for audio instead 
> of a video device.

  I'm working on a better method for detecting the PVR350-output in the 
ivtv-vidix part of mplayer, so this limitation is not too bad. It's ivtv 
specific, so the code can presume some things.

  I'll need to study the v4l2-spec a bit more before I can provide a 
good proposal for the api. Are there any discussions which regard this 
kind of limitation?

Lars.

_______________________________________________
ivtv-devel mailing list
[email protected]
http://ivtvdriver.org/mailman/listinfo/ivtv-devel

Reply via email to