On Sun, Jan 12, 2014 at 7:20 PM, Ed Beroset <[email protected]> wrote:
> I could, but perhaps it's a moot point.  When I use "vlc
> v4l2c:///dev/video:input=2" it works perfectly.  When I use "vlc
> v4l2:///dev/video:input=2" (which is what I'd been attempting up until your
> message) I get the same errors as listed in ticket 10258.
>
> So, I'm happy that a solution seems to be at hand, but vlc only showed me
> v4l2:// devices under the heading "Capture Devices" and I did not know of
> the existence of v4l2c:// devices until just now (and don't know what they
> are.)  Thank you for bridging the gap for me -- what can we do to help the
> next person encountering this problem?

The root of the issue is that the V4L2 spec had some ambiguity, so
until very recently devices did a really bad job of advertising
whether they were providing compressed video or raw uncompressed video
(the issue was particularly bad in cases where a given device could
provide both).  The use of "v4l2c://" explicitly tells VLC that the
PVR-150 is providing compressed video, which is why it works for you.
If you instead tell it "v4l2://" then you're telling VLC that the
device provides uncompressed video, which the PVR-150 doesn't support
- that's why you're getting the error.

Presumably drivers do a better job advertising what they support
nowadays (primarily as a result of work Hans Verkuil did over the last
year or so), so VLC should probably be changed such that specifying
"v4l2://" causes it to ask the driver whether the video is compressed,
and handle it accordingly.  Then in the longer term the "v4l2c://"
method would be deprecated entirely.

Even there you probably want to add some sort of option to allow the
caller to specify which to use when a device supports both, since
there are advantages to each approach depending on what you're
ultimately doing with the video.  For example, if you're streaming it
over a network then compressed is the better choice - unless you want
to stream it in H.264, in which case you're better doing uncompressed
so that you avoid the cost of decompressing the stream before
re-encoding it to H.264.

Long story short, none of this is really a problem with the IVTV
driver, and if you want to see VLC's logic improved then it would
probably makes sense to start a new thread over on the vlc-devel
mailing list.

Devin


-- 
Devin J. Heitmueller - Kernel Labs
http://www.kernellabs.com

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

Reply via email to