On Fri, Oct 15, 2010 at 02:24:27PM +0200, Jan Stary wrote:
> On Oct 14 01:59:39, Jacob Meuser wrote:
> > On Thu, Oct 14, 2010 at 12:05:09AM +0000, Jacob Meuser wrote:
> > > On Thu, Oct 14, 2010 at 12:59:37AM +0200, Jan Stary wrote:
> > > > On Oct 13 22:07:27, Jacob Meuser wrote:
> > > > > On Wed, Oct 13, 2010 at 11:51:47PM +0200, Jan Stary wrote:
> > > > > > This is a fresh upgrade of current/amd64 on a Macbook3,1.
> > > > > > Basically, everything works except the wifi and uvideo.
> > > > >
> > > > > > I installed the uvideo firmware as documented in uvideo(4).
> > > > > > I can capture an image with 'fswebcam image.jpg';
> > > > > > I have problems capturing a video stream with
> > > > > > ffmpeg -y -f video4linux2 -s vga -r 25 -i /dev/video0 /tmp/out.mpg
> > > > > > (the fps eventually drops to 0 and results in a tiny file
> > > > > > containing just a few frames) but I haven't really tried
> > > > > > tweaking the format options.
> > > > >
> > > > > ffmpeg doesn't work with video(4). frankly, I don't understand why
> > > > > video(4) support is enabled in ffmpeg, giving the the impression it's
> > > > > supposed to work, when it clearly doesn't.
> > > >
> > > > On another machine (HP EliteBook 8530w running a recent -current)
> > > > I can capture video(4) just fine - using certain resolutions and frame
> > > > rates, that is.
> > >
> > > with ffmpeg? can you share your command, I've never gotten it to work.
>
> Yes, with ffmpeg.
> I will get my hands on the HP during the weekend and will come back.
>
> > fwiw, this is what I use, because using ffmpeg to capture directly from
> > video(4) has never worked for me:
> >
> > $ video -f /dev/video1 -e yuy2 -s 640x480 -r 20 -o - | ffmpeg -y \
> > -f rawvideo -pix_fmt yuyv422 -s 640x480 -r 20 -i - \
> > -itsoffset 0.5 -f sndio -ar 16000 -ac 1 -i sun:1 out.avi
>
> Thank you for mentioning video(1), I didn't even know it existed.
> However, it only segfaults on this macbook
>
> $ video
> video: /dev/video: returned size not as requested
> Segmentation fault (core dumped)
probably the segfault is because of symbol conflicts with libxcb, which
has already been fixed, but I'd need to see a backtrace to know
for sure.
> $ video -O video.raw -r 25 -s 640x480 -v
> video device /dev/video:
> encodings: uyvy
> sizes: 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
> 0x0, 0x0, 0x0
> controls: brightness, saturation, gamma, sharpness
> Xv adaptor 0, Intel(R) Textured Video:
> encodings: yuy2, uyvy
> max size: 1280x800
> using yuy2 encoding
> using frame size 0x0 (0 bytes)
> video: /dev/video: returned size not as requested
> Segmentation fault (core dumped)
>
> I have tried various options (-r 15, -r 25, -s 640, -s 1280, -s 320)
> with always the same result.
>
> Why is it that the only sizes supported are 0x0, 0x0, etc?
> Is this the problem ("returned size not as requested")?
video(1) isn't getting the sizes correctly for some reason.
can you compile a kernel with UVIDEO_DEBUG defined (or just
add a #define it in sys/dev/usb/uvideo.c) and send me the
resulting dmesg?
> The video(1) manpage contains a few typos, see diff below.
thanks.
> The camera is a bultin Micron iSight; the manpage mentions
> "Apple builtin iSight" as known to work.
>
> Also, the driver seems to first attach to bluetooth, then detach,
> and later attach to the camera:
>
> uvideo0 at uhub1 port 4 configuration 1 interface 0 "Apple Computer
> Bluetooth" rev 2.00/0.0c addr 2
> uvideo0 detached
> uvideo0 at uhub1 port 4 configuration 1 interface 0 "Micron Built-in iSight"
> rev 2.00/1.84 addr 2
> video0 at uvideo0
>
> Is this normal?
looks like it is from what I see in uvideo.c.
>
> Thank you for your time
>
> Jan
>
>
>
> --- video.1.orig Fri Oct 15 09:43:18 2010
> +++ video.1 Fri Oct 15 09:43:32 2010
> @@ -108,7 +108,7 @@
> is
> .Ql - ,
> frames will be read from standard input.
> -.It Fl O Ar ouput
> +.It Fl O Ar output
> File to which frames will be written.
> If
> .Ar output
> @@ -121,7 +121,7 @@
> .Nm
> will also display the frames via
> .Xr Xv 3 .
> -.It Fl o Ar ouput
> +.It Fl o Ar output
> File to which frames will be written.
> If
> .Ar output
--
jake...@sdf.lonestar.org
SDF Public Access UNIX System - http://sdf.lonestar.org
