On Wed October 17 2012 02:25:00 Laurent Pinchart wrote:
Hi Hans,
On Monday 15 October 2012 13:35:45 Hans Verkuil wrote:
During the Plumbers Conference a few weeks ago we had a session to resolve V4L2 ambiguities. It was very successful, but we didn't manage to tackle two of the harder topics, and a third one (timestamps) cause a lot of discussion on the mailinglist after the conference.
So here is the list I have today. Any other ambiguities or new features that should be added to the list?
Small topic that we've briefly discussed on IRC: if a device doesn't tell the driver what color space it uses, should the driver guess or tell the application that the color space is unknown ? I've ran into that issue for the uvcvideo driver, while I agree with you that in that case the color space is very likely sRGB, and that the driver is probably in a better position to make that guess than the userspace application (as the driver knows it handles a webcam), what should be the rule ?
I'll add it to the list.
- Make a decision how to tell userspace that the monotonic timestamp is
used.
Several proposals were made, but no decision was taken AFAIK. Can someone (Sakari?) make a summary/current status of this?
- Pixel Aspect Ratio
Pixel aspect: currently this is only available through VIDIOC_CROPCAP. It never really belonged to VIDIOC_CROPCAP IMHO. It's just not a property of cropping or composing. It really belongs to the input/output timings (STD or DV_TIMINGS). That's where the pixel aspect ratio is determined.
While it is possible to add it to the dv_timings struct, I see no way of cleanly adding it to struct v4l2_standard (mostly because VIDIOC_ENUMSTD is now handled inside the V4L2 core and doesn't call the drivers anymore).
Isn't that an implementation issue instead of an API issue ?
True, but it will require changing all drivers. But adding it to v4l2_standard would be the best place for this, so perhaps we should think about doing all this work.
Regards,
Hans