Mailing List Archive

On Sat, 2009-02-07 at 14:49 -0500, Jeff Campbell wrote:
> Hi Andy,
>
> Just going through the detailed specs on the cx18 chip and noticed the
> Conexant marketing sheet says the max bitrate for the chip is 15 Mbps
> in either CBR or VBR mode.
>
> "Variable and constant bit rate up to 15 Mb/sec" under Features and
> Benefits / Video Encoding
>
> At the moment the cx18 still reports 27 Mbps as the max in v4l2-ctl
> -L, which I think is just a carry over from the earlier 15/16 chip.
> Just wanted to flag it before it slipped my mind.

Yes, I saw that the other week myself. I hadn't bothered to fix the
control yet. It's wrapped up in some callouts to cx23415/6/7/8 common
code and some v4l2-common stuff:

linux/drivers/media/video/v4l2-common.c
linux/drivers/media/video/cx2341x.c

I put it low on the list, because mucking with those means more
regression testing than normal to make sure I didn't screw up the ivtv,
pvrusb2, cx88, or cx23885 drivers, as they all call out to those too
(IIRC).


I suppose I could intercept that control and fix it here:

linux/drivers/media/video/cx18/cx18-controls.c

Thanks.

Regards,
Andy

> -Jeff



_______________________________________________
ivtv-devel mailing list
ivtv-devel@ivtvdriver.org
http://ivtvdriver.org/mailman/listinfo/ivtv-devel

>
> > "Variable and constant bit rate up to 15 Mb/sec" under Features and
> > Benefits / Video Encoding
> >
> > At the moment the cx18 still reports 27 Mbps as the max in v4l2-ctl
> > -L, which I think is just a carry over from the earlier 15/16 chip.
> > Just wanted to flag it before it slipped my mind.
>
>
> I put it low on the list, because mucking with those means more
> regression testing than normal to make sure I didn't screw up the ivtv,
> pvrusb2, cx88, or cx23885 drivers, as they all call out to those too
> (IIRC).
>
>
I suspected that might be the case, just wanted to mention it before I
forgot.

In a similar vein, we were doing some testing with VBR versus CBR last night
and noticed that the current structure, which I believe is also carried over
from ivtv, does not allow changing the bitrate on the fly. However, the
chip does allow changing the bitrate on the fly, but changing that will also
create regression issues potentially. However, that being said, I used to
be able to change the BR on the fly with ivtv and some time in the last year
or so it changed so you could not write that while the device was in use.
Not sure why that was done, perhaps for a different device in the same
family?

-Jeff

On Sunday 08 February 2009 14:01:34 Jeff Campbell wrote:
> > > "Variable and constant bit rate up to 15 Mb/sec" under Features and
> > > Benefits / Video Encoding
> > >
> > > At the moment the cx18 still reports 27 Mbps as the max in v4l2-ctl
> > > -L, which I think is just a carry over from the earlier 15/16 chip.
> > > Just wanted to flag it before it slipped my mind.
> >
> > I put it low on the list, because mucking with those means more
> > regression testing than normal to make sure I didn't screw up the ivtv,
> > pvrusb2, cx88, or cx23885 drivers, as they all call out to those too
> > (IIRC).
>
> I suspected that might be the case, just wanted to mention it before I
> forgot.
>
> In a similar vein, we were doing some testing with VBR versus CBR last
> night and noticed that the current structure, which I believe is also
> carried over from ivtv, does not allow changing the bitrate on the fly.
> However, the chip does allow changing the bitrate on the fly, but
> changing that will also create regression issues potentially. However,
> that being said, I used to be able to change the BR on the fly with ivtv
> and some time in the last year or so it changed so you could not write
> that while the device was in use. Not sure why that was done, perhaps for
> a different device in the same family?

If I remember correctly I disabled it for ivtv after experimentation showed
that changing the bitrate while capturing would sometimes (always? Can't
remember...) result in a corrupt stream. At least the cx23415/6 seems to be
very sensitive to changes in the compression parameters or in the video
format coming from the video digitizer. Perhaps the cx23418 can handle that
better, but I have my doubts given that it has the some core hardware.

Regards,

Hans

--
Hans Verkuil - video4linux developer - sponsored by TANDBERG

_______________________________________________
ivtv-devel mailing list
ivtv-devel@ivtvdriver.org
http://ivtvdriver.org/mailman/listinfo/ivtv-devel

We will do some more testing and report back on how the cx18 handles it.

-Jeff

On Sun, Feb 8, 2009 at 8:25 AM, Hans Verkuil <hverkuil@xs4all.nl> wrote:

> On Sunday 08 February 2009 14:01:34 Jeff Campbell wrote:
> > > > "Variable and constant bit rate up to 15 Mb/sec" under Features and
> > > > Benefits / Video Encoding
> > > >
> > > > At the moment the cx18 still reports 27 Mbps as the max in v4l2-ctl
> > > > -L, which I think is just a carry over from the earlier 15/16 chip.
> > > > Just wanted to flag it before it slipped my mind.
> > >
> > > I put it low on the list, because mucking with those means more
> > > regression testing than normal to make sure I didn't screw up the ivtv,
> > > pvrusb2, cx88, or cx23885 drivers, as they all call out to those too
> > > (IIRC).
> >
> > I suspected that might be the case, just wanted to mention it before I
> > forgot.
> >
> > In a similar vein, we were doing some testing with VBR versus CBR last
> > night and noticed that the current structure, which I believe is also
> > carried over from ivtv, does not allow changing the bitrate on the fly.
> > However, the chip does allow changing the bitrate on the fly, but
> > changing that will also create regression issues potentially. However,
> > that being said, I used to be able to change the BR on the fly with ivtv
> > and some time in the last year or so it changed so you could not write
> > that while the device was in use. Not sure why that was done, perhaps for
> > a different device in the same family?
>
> If I remember correctly I disabled it for ivtv after experimentation showed
> that changing the bitrate while capturing would sometimes (always? Can't
> remember...) result in a corrupt stream. At least the cx23415/6 seems to be
> very sensitive to changes in the compression parameters or in the video
> format coming from the video digitizer. Perhaps the cx23418 can handle that
> better, but I have my doubts given that it has the some core hardware.
>
> Regards,
>
> Hans
>
> --
> Hans Verkuil - video4linux developer - sponsored by TANDBERG
>

Hans Verkuil wrote:
> If I remember correctly I disabled it for ivtv after experimentation
> showed
> that changing the bitrate while capturing would sometimes (always? Can't
> remember...) result in a corrupt stream. At least the cx23415/6 seems to be
> very sensitive to changes in the compression parameters or in the video
> format coming from the video digitizer. Perhaps the cx23418 can handle that
> better, but I have my doubts given that it has the some core hardware.
>
> Regards,
>
> Hans
>
>
We'll experiment a bit with that - I saw it was a limitation in the
common module for these drivers, so I'll try it without the restriction
and see how it goes on the cx18. The reason we were interested in this
was that the documentation for the cx18 makes explicit mention of the
ability to change it while the stream is in progress.

Mike

_______________________________________________
ivtv-devel mailing list
ivtv-devel@ivtvdriver.org
http://ivtvdriver.org/mailman/listinfo/ivtv-devel