On Saturday 07 October 2006 05:10, Stephen Mucher wrote:
> Hans,
>
> Thanks for your insight!
>
> See below ...
>
> - Show quoted text -
>
> On 10/6/06, Hans Verkuil < [EMAIL PROTECTED]> wrote:
> > On Friday 06 October 2006 20:12, Stephen Mucher wrote:
> > > Hi!
> > >
> > > I've been eagerly awaiting CC support for my new PVR-500, and
> > > with the dawn of the 2.6.18 kernel and ivtv 0.8 I had hoped the
> > > wait was finally over ....
> > >
> > > Alas, that was not the case.
> > >
> > > MythTV 0.20 does not seem to detect any closed captions at all,
> > > and I have reason to believe the culprit is somehow rooted in
> > > ivtv.
> > >
> > > I have configured each tuner for cc sliced vbi mode & stream
> > > embedding via the v412-ctl utility:
> > >
> > > v4l2-ctl --set-fmt-sliced-vbi=cc -c stream_vbi_format=1 -d0
> > > v4l2-ctl --set-fmt-sliced-vbi=cc -c stream_vbi_format=1 -d1
> > >
> > > Doing this does not seem to help -- in fact, the video stream
> > > becomes intermittently corrupt as a result. This video
> > > corruption is present in MythTV, as well as when I capture the
> > > video stream directly (cat /dev/video0 > test.mpg). Side note:
> > > Is there a way to extract any embedded vbi data from the mpeg
> > > stream to rule out MythTV altogether?
> > >
> > > I tried --get-fmt-sliced-vbi, and the results seem off:
> > >
> > > Format:
> > > Type : Sliced VBI Capture
> > > Service Set : cc
> > > Service Line 0: /
> > > Service Line 1: /
> > > Service Line 2: /
> > > Service Line 3: /
> > > Service Line 4: /
> > > Service Line 5: /
> > > Service Line 6: /
> > > Service Line 7: cc / cc
> > > Service Line 8: cc / cc
> > > Service Line 9: cc / cc
> > > Service Line 10: cc / cc
> > > Service Line 11: cc / cc
> > > Service Line 12: cc / cc
> > > Service Line 13: cc / cc
> > > Service Line 14: cc / cc
> > > Service Line 15: cc / cc
> > > Service Line 16: cc / cc
> > > Service Line 17: cc / cc
> > > Service Line 18: cc / cc
> > > Service Line 19: /
> > > Service Line 20: /
> > > Service Line 21: /
> > > Service Line 22: /
> > > Service Line 23: /
> > > I/O Size : 0
> > >
> > > Shouldn't I expect to see cc on service line 21?? I guess I dont
> > > really understand what this indicated.
> >
> > I think this is a bug in the implementation of the VIDIOC_G_FMT
> > ioctl. Lines 7-18 should be shifted three lines down (10-21). It
> > looks like the table is filled incorrectly for NTSC. The internal
> > driver settings are not affected, it's a reporting bug only. I'll
> > investigate to see where it goes wrong.
I can now confirm that this is the case. I think I can get the fix into
2.6.19. I've posted the patch to the v4l-dvb-maintainer list.
> >
> > The MPEG corruption you see is something I'm hunting down (although
> > not much progress the last two weeks due to lack of time). I know
> > what is causing it and I know how to work around it, but the devil
> > is in the details and that's taking me more time.
>
> Is the work around something I could patch on my end?
No, it's a complete redesign of the way the driver handles the
interrupts and DMA. Not a simple patch at all.
>
> But nevertheless, I would have expected to see MythTV pick up the CC
>
> > data. If the vbi test tool works, then CC is read correctly, so it
> > should also be available in MythTV.
> >
> > Possible exception: if you run a 64-bit kernel, then you also need
> > to apply patch misc/MythTV-0.20.diff over MythTV-0.20: that fixes a
> > 64-bit bug in MythTV.
>
> Nope -- running the 32-bit version here.
>
> Check with v4l2-ctl --log-status whether VBI embedding is really
> turned
>
> > on when a MythTV capture is in progress.
>
> Hmmm ... no mention of vbi embedding in the dmesg excerp:
Oops, I'd forgotten that that was a post-2.6.18 change.
Try 'v4l2-ctl -C stream_vbi_format' instead. It should return 1.
Yes -- it returns 1. The 'ivtv-enc-vbi' process is running, as well.
~Stephen
Regards,
Hans
_______________________________________________
ivtv-users mailing list
[email protected]
http://ivtvdriver.org/mailman/listinfo/ivtv-users
_______________________________________________ ivtv-users mailing list [email protected] http://ivtvdriver.org/mailman/listinfo/ivtv-users
