gatos-devel Mailing List for GATOS

On Mon, 31 Dec 2001, Nikolai wrote:

> Hi,
> It seems you are not alone in getting frames rotated somewhat left ;)
> I noticed this phenomena when updated from km 0.1.1 to 0.1.7.
> I gave up for now. However, ffmpeg is able to capture for me and
> nothing hangs at least. Though, not a usable quality video yet.
> I have old Cel 366 and old AIW Pro, Red Hat 7.1.
> BTW, ATI's player for Windows manages to capure easily on this same
> box at 360x240, but it locks the box tightly from time to time. There
> are probably lots of bugs in ATI's player and .DLLs, just nobody will
> ever fix them. These mach64-based cards, like mine, are not manufactured
> any more, so ATI does not need to care much about. Still, it seems that
> in theory reasonable quality capturing could be achieved with km+ffmpeg
> or something like, even not on a very fast CPU. At least not worse than
> with ATI's player, and probably better. I think the work is now going on,
> and it will happen some happy day :) thanks Volodya, Peter and other people.

I'll try km 0.1.1 :-)

Brain

--------------------------------
Petr `Brain' Kulhavy
<br...@ar...>
http://artax.karlin.mff.cuni.cz/~brain
Faculty of Mathematics and Physics, Charles University Prague, Czech Republic

---
A man without a woman is like a statue without pigeons.

On Mon, 2001-12-31 at 13:56, Vladimir Dergachev wrote:
>=20
>    New tag checkpoint_3 is available in modules ati.2, drm-kernel and km.
>=20
>    What you get: stable source with 2d, 3d (Rage128 and Radeon only),
>            TV-in, capture. Initial support for 8500DV (no 3d). No TV-out.
>=20
>    This is meant to become the "official" driver set for 4.2.0. This is
>    also the code I'll be pushing to XFree after 4.2.0 is released.
>=20
>    So please pound on it :))
>=20
>    Note: the stable ati.2 should no longer work with standard drm drivers
>    for Radeon. You need to get drm-kernel.=20
>=20

Wee!

I just installed, first impressions are great. Im about to do a capture,
i'll let you know if anything interesting happens. Lockups with xawtv
are gone :> I would test avview but I don't have alsa installed
currently.=20

Great work.

--=20
Morgan Collins http://sirmorcant.morcant.org
Crypto Doesn't Kill - People Do.

   New tag checkpoint_3 is available in modules ati.2, drm-kernel and km.

   What you get: stable source with 2d, 3d (Rage128 and Radeon only),
           TV-in, capture. Initial support for 8500DV (no 3d). No TV-out.

   This is meant to become the "official" driver set for 4.2.0. This is
   also the code I'll be pushing to XFree after 4.2.0 is released.

   So please pound on it :))

   Note: the stable ati.2 should no longer work with standard drm drivers
   for Radeon. You need to get drm-kernel. 

                       best

                                Vladimir Dergachev

I'll try the newer version.  My monitor is a NEC XV15+.  It does not flicker 
on the normal X tree (ie: actually runs at 85 Hz)

 Rob

> > Secondly, even though xvidtune says my refresh rate is 85 Hz, my monitor
> > was flickering quite a lot - like it was at 60 Hz.  Is there any way to
> > fix this?
>
> Which monitor ? Which branch are you using ?
>
>                           Vladimir Dergachev
>
> > And lastly, glxgears gives me a frame rate of about 1/3 of what I got
> > with the normal X driver, and direct rendering is working.  Is there
> > anyway to improve this?
> >
> > Thanks.
> >
> > _______________________________________________
> > Gatos-devel mailing list
> > Gat...@li...
> > https://lists.sourceforge.net/lists/listinfo/gatos-devel
>
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel

On Mon, 31 Dec 2001, Rob Weryk wrote:

> Hi, I tried the ati.2 driver (from CVS a couple of days ago) along with the a 
> recent CVS pull of X.  I just have a few questions.

Please try again :)) There was some major upgrade last night.

> 
> Is the video in support only for cards with a TV tuner?  I have a Radeon VIVO 
> with no tuner and I could not get any video in - just a black picture.

In principle yes, but this card is doing something strange with video
input assignment.. Try looking in radeon_video.c, RADEONReadMM_TABLE
function and force "xxxConnector" to different values.

> 
> Secondly, even though xvidtune says my refresh rate is 85 Hz, my monitor was 
> flickering quite a lot - like it was at 60 Hz.  Is there any way to fix this?

Which monitor ? Which branch are you using ?

                          Vladimir Dergachev

> 
> And lastly, glxgears gives me a frame rate of about 1/3 of what I got with 
> the normal X driver, and direct rendering is working.  Is there anyway to 
> improve this?
> 
> Thanks.
> 
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
> 

Hi, I tried the ati.2 driver (from CVS a couple of days ago) along with the a 
recent CVS pull of X.  I just have a few questions.

Is the video in support only for cards with a TV tuner?  I have a Radeon VIVO 
with no tuner and I could not get any video in - just a black picture.

Secondly, even though xvidtune says my refresh rate is 85 Hz, my monitor was 
flickering quite a lot - like it was at 60 Hz.  Is there any way to fix this?

And lastly, glxgears gives me a frame rate of about 1/3 of what I got with 
the normal X driver, and direct rendering is working.  Is there anyway to 
improve this?

Thanks.

Hello!

I've tested new stuff: SetTransform.
It works! Great work! Thanks you!
I've applied it into my radeon_vid too.

HAPPY NEW YEAR!

Best regards! Nick

> All code in current GATOS CVS has been updated to 4.1.99.4+. (tree of
>December 30th). It compiles fine on my system. If you keep XFree86 CVS
>closeby please try and report if anything breaks (it shouldn't !). 
>
> Modules updated:
>
>     ati.2
>     ati.2/devel
>     ati.2/orig
>     drm-kernel
>     drm-kernel/orig
>
>(/XXX indicates branch XXX).
>
>                           Vladimir Dergachev

Just tried the main ati.2 CVS without problems. A big improvement, the mouse 
cursor no longer sticks when close to the left edge of the screen.

very nice.

Jean-Yves Coupal

 All code in current GATOS CVS has been updated to 4.1.99.4+. (tree of
December 30th). It compiles fine on my system. If you keep XFree86 CVS
closeby please try and report if anything breaks (it shouldn't !). 

 Modules updated:

     ati.2
     ati.2/devel
     ati.2/orig
     drm-kernel
     drm-kernel/orig

(/XXX indicates branch XXX).

                           Vladimir Dergachev

On Mon, 31 Dec 2001, Peter Surda wrote:

> On Mon, Dec 31, 2001 at 12:09:10AM -0200, Marcio Calixto Cabral wrote:
> >  Hi there,
> hi
> 
> >  I have been struggling the past few days to make my Rage Fury Pro
> > capture work under linux. I have followed what Mark have said, but  I am
> > getting stuck with this message:
> > 
> >  make test
> > sync
> > insmod ./km_api.o
> > insmod ./km_drv.o km_debug=0
> > ./km_drv.o: init_module: No such device
> > Hint: insmod errors can be caused by incorrect module parameters,
> > including invalid IO or IRQ parameters
> > make: *** [test] Error 1
> Grr volodya you forgot to fix the docs <g>

Actually this is on purpose - I don't want people installing km in
/lib/modules so they don't load old versions accidentally.

(I.e. insmod km_drv will try to load the module from /lib/modules/... path
first, to load module from the current directory use insmod ./km_drv.o)

                            Vladimir Dergachev

> 
> correct should be:
> make && make install
> add into /etc/modules.conf:
> ----
> options km_drv km_debug=0
> alias char-major-81 km_drv
> ----
> depmod -a
> 
> and then it should work automagically, supposing you have all modules
> compiled.
> 
> >  I am not getting Direct Rendering under X now after unpacking ATI.tgz and
> > DRM.tgz. I would really like to get this thing working this year :). Any
> > help would be really appreciated.
> use the source :-) And happy new year.
> 
> Bye,
> 
> Peter Surda (Shurdeek) <shu...@pa...>, ICQ 10236103, +436505122023
> 
> --
>                    Press every key to continue.
> 

 Well, I figured out the busy IRQ conflict problem but now I am facing
another one:

 I have km_drm and km_api loaded. When I start xawtv I get a fuzzy image
from the input camera. Does anybody know what is wrong? I tried changing
the input display, norm, etc, but no success.

 Any help would be greatly appreciated.

 Marcio Cabral

On Mon, Dec 31, 2001 at 12:09:10AM -0200, Marcio Calixto Cabral wrote:
>  Hi there,
hi

>  I have been struggling the past few days to make my Rage Fury Pro
> capture work under linux. I have followed what Mark have said, but  I am
> getting stuck with this message:
>=20
>  make test
> sync
> insmod ./km_api.o
> insmod ./km_drv.o km_debug=3D0
> ./km_drv.o: init_module: No such device
> Hint: insmod errors can be caused by incorrect module parameters,
> including invalid IO or IRQ parameters
> make: *** [test] Error 1
Grr volodya you forgot to fix the docs <g>

correct should be:
make && make install
add into /etc/modules.conf:
----
options km_drv km_debug=3D0
alias char-major-81 km_drv
----
depmod -a

and then it should work automagically, supposing you have all modules
compiled.

>  I am not getting Direct Rendering under X now after unpacking ATI.tgz and
> DRM.tgz. I would really like to get this thing working this year :). Any
> help would be really appreciated.
use the source :-) And happy new year.

Bye,

Peter Surda (Shurdeek) <shu...@pa...>, ICQ 10236103, +436505=
122023

--
                   Press every key to continue.

On Mon, 31 Dec 2001, Peter Surda wrote:

> On Mon, Dec 31, 2001 at 12:42:15AM +0100, br...@ar... wrote:
> > > Yes but is also says that the input size is 384x144 while in fact it is
> > > 640x240 (for NTSC). So, here is your problem !
> > No. I set the size using xatitv (it was in some of my previous mails).
> Please actually read the mails. This has no effect.
> 
> BTW U sure you have NTSC? Last time I checked Czech Republic had PAL?

NTSC is my line, I am merely trying to point out that my numbers are for
NTSC (as I don't remember them offhand for PAL and SECAM.. 720x288
something.. )

                    Vladimir Dergachev

> 
> > Brain
> Bye,
> 
> Peter Surda (Shurdeek) <shu...@pa...>, ICQ 10236103, +436505122023
> 
> --
> Actually, Microsoft is sort of a mixture between the Borg and the Ferengi.
> 

 Hi there,

 I have been struggling the past few days to make my Rage Fury Pro
capture work under linux. I have followed what Mark have said, but  I am
getting stuck with this message:

 make test
sync
insmod ./km_api.o
insmod ./km_drv.o km_debug=0
./km_drv.o: init_module: No such device
Hint: insmod errors can be caused by incorrect module parameters,
including invalid IO or IRQ parameters
make: *** [test] Error 1

 I am running a Debian with kernel 2.4.5. Here are some of my systems
configs:

lsmod
km_api                  2176   0
emu10k1                46096   0  (unused)
ppp_deflate            39136   0  (autoclean)
bsd_comp                4096   0  (autoclean)
ppp_async               6480   1  (autoclean)
sm56                 1345984   1  (autoclean)
ppp_generic            16784   3  (autoclean) [ppp_deflate bsd_comp
ppp_async]
vfat                    8848   1  (autoclean)
fat                    31648   0  (autoclean) [vfat]
phonedev                2816   0  (unused)
tuner-3036              1840   0  (unused)
parport_pc             23408   0  (autoclean)
i2c-parport             1344   0  (unused)
parport                25536   0  [parport_pc i2c-parport]
i2c-old                 3904   1  [i2c-parport]
videodev                5216   0
raw1394                 7328   0  (unused)
ieee1394               22288   0  [raw1394]
i2o_config              7472   0  (unused)
i2o_block              34144   0  (unused)
i2o_core               41104   0  [i2o_config i2o_block]
i2o_pci                 2576   1  [i2o_core]
i2c-algo-bit            7232   0  (unused)
i2c-core               13136   0  [tuner-3036 i2c-algo-bit]
mousedev                4160   1
hid                    11760   0  (unused)
usbmouse                2000   0  (unused)
input                   3648   0  [mousedev hid usbmouse]

 cat /proc/pci
PCI devices found:
  Bus  0, device   0, function  0:
    Host bridge: PCI device 8086:1130 (Intel Corporation) (rev 2).
      Prefetchable 32 bit memory at 0xf8000000 [0xfbffffff].
  Bus  0, device   1, function  0:
    PCI bridge: PCI device 8086:1131 (Intel Corporation) (rev 2).
      Master Capable.  No bursts.  Min Gnt=8.
  Bus  0, device  30, function  0:
    PCI bridge: Intel Corporation 82820 820 (Camino 2) Chipset PCI (rev
1).
      Master Capable.  No bursts.  Min Gnt=6.
  Bus  0, device  31, function  0:
    ISA bridge: Intel Corporation 82820 820 (Camino 2) Chipset ISA Bridge
(ICH2) (
rev 1).
  Bus  0, device  31, function  1:
    IDE interface: Intel Corporation 82820 820 (Camino 2) Chipset IDE U100
(rev 1)
.
      I/O at 0xa800 [0xa80f].
  Bus  0, device  31, function  2:
    USB Controller: Intel Corporation 82820 820 (Camino 2) Chipset USB
(Hub A) (re
v 1).
      IRQ 9.
      I/O at 0xa400 [0xa41f].
  Bus  0, device  31, function  3:
    SMBus: Intel Corporation 82820 820 (Camino 2) Chipset SMBus (rev 1).
      IRQ 3.
      I/O at 0xe800 [0xe80f].
  Bus  0, device  31, function  4:
    USB Controller: Intel Corporation 82820 820 (Camino 2) Chipset USB
(Hub B) (re
v 1).
      IRQ 9.
      I/O at 0xa000 [0xa01f].
  Bus  1, device   0, function  0:
    VGA compatible controller: ATI Technologies Inc Rage 128 PF (rev 0).
      IRQ 11.
      Master Capable.  Latency=64.  Min Gnt=8.
      Prefetchable 32 bit memory at 0xf4000000 [0xf7ffffff].
      I/O at 0xd800 [0xd8ff].
      Non-prefetchable 32 bit memory at 0xf2800000 [0xf2803fff].
  Bus  2, device  10, function  0:
    Multimedia audio controller: Creative Labs SB Live! EMU10000 (rev 7).
      IRQ 9.
      Master Capable.  Latency=64.  Min Gnt=2.Max Lat=20.
      I/O at 0xb800 [0xb81f].
  Bus  2, device  10, function  1:
    Input device controller: Creative Labs SB Live! (rev 7).
      I/O at 0xb400 [0xb407].
  Bus  2, device  14, function  0:
    Communication controller: Motorola SM56 PCI Modem (rev 0).
      IRQ 5.
      Master Capable.  No bursts.  Min Gnt=1.Max Lat=255.
      Prefetchable 32 bit memory at 0xf3000000 [0xf30000ff].

 I am not getting Direct Rendering under X now after unpacking ATI.tgz and
DRM.tgz. I would really like to get this thing working this year :). Any
help would be really appreciated.

 Thank in advance,
 Marcio Cabral

On Mon, Dec 31, 2001 at 12:42:15AM +0100, br...@ar... wrote:
> > Yes but is also says that the input size is 384x144 while in fact it is
> > 640x240 (for NTSC). So, here is your problem !
> No. I set the size using xatitv (it was in some of my previous mails).
Please actually read the mails. This has no effect.

BTW U sure you have NTSC? Last time I checked Czech Republic had PAL?

> Brain
Bye,

Peter Surda (Shurdeek) <shu...@pa...>, ICQ 10236103, +436505122023

--
Actually, Microsoft is sort of a mixture between the Borg and the Ferengi.

On Sun, 30 Dec 2001, Mark Roth wrote:

> Vladimir,
> 
> If I do that with the default RedHat Linux 7.2 kernel, I get the 
> following errors (I've also attached the build log so you can see the 
> warnings I get).  Any ideas?

Try disabling MODVERSIONS (module versioning).

                      Vladimir Dergachev

> 
> Thanks,
> Mark
> 
> >> > a) get the latest km
> >>
> >>     * Downloaded km-0.1.9.tgz from gatos.sourceforge.net.
> >>     * Compiled and installed to /lib/modules/.../misc/.
> >>
> > 
> > You don't need to install it. Instead become root and run from km
> > directory:
> > 
> >   /sbin/modprobe videodev
> >   make test
> > 
> >                            Vladimir Dergachev
> 
> 

Vladimir,

If I do that with the default RedHat Linux 7.2 kernel, I get the 
following errors (I've also attached the build log so you can see the 
warnings I get).  Any ideas?

Thanks,
Mark

>> > a) get the latest km
>>
>>     * Downloaded km-0.1.9.tgz from gatos.sourceforge.net.
>>     * Compiled and installed to /lib/modules/.../misc/.
>>
> 
> You don't need to install it. Instead become root and run from km
> directory:
> 
>   /sbin/modprobe videodev
>   make test
> 
>                            Vladimir Dergachev

On Sun, 30 Dec 2001, Vladimir Dergachev wrote:

>
>
> On Mon, 31 Dec 2001 br...@ar... wrote:
>
> > On Sun, 30 Dec 2001, Vladimir Dergachev wrote:
> >
> > > Are you sure ? Because, by default, ffmpeg wants yuv420. Also, km
> > > generates fields 640x240 in size (for NTSC) - this could also be a
> > > problem.
> >
> > Here's ffmpeg output:
> >
> > Input #0, video_grab_device, from '/dev/video':
> >   Stream #0.0: Video: rawvideo, yuv422, 384x144, 50.00 fps
> > Input #1, audio_device, from '/dev/dsp':
> >   Stream #1.0: Audio: pcm_s16le, 44100 Hz, mono, 705 kb/s
> > Output #0, avi, to 'test.avi':
> >   Stream #0.0: Video: msmpeg4, 320x240, 20.00 fps, 200 kb/s
> >   Stream #0.1: Audio: ac3, 44100 Hz, mono, 128 kb/s
> > Stream mapping:
> >   Stream #0.0 -> #0.0
> >   Stream #1.0 -> #0.1
> > Press [q] to stop encoding
> >
> >
> > ffmpeg says input is yuv422 so I believe it :)
>
> Yes but is also says that the input size is 384x144 while in fact it is
> 640x240 (for NTSC). So, here is your problem !

No. I set the size using xatitv (it was in some of my previous mails).

Brain

--------------------------------
Petr `Brain' Kulhavy
<br...@ar...>
http://artax.karlin.mff.cuni.cz/~brain
Faculty of Mathematics and Physics, Charles University Prague, Czech Republic

---
One nuclear bomb can ruin your whole day.

On Mon, 31 Dec 2001 br...@ar... wrote:

> On Sun, 30 Dec 2001, Vladimir Dergachev wrote:
> 
> > Are you sure ? Because, by default, ffmpeg wants yuv420. Also, km
> > generates fields 640x240 in size (for NTSC) - this could also be a
> > problem.
> 
> Here's ffmpeg output:
> 
> Input #0, video_grab_device, from '/dev/video':
>   Stream #0.0: Video: rawvideo, yuv422, 384x144, 50.00 fps
> Input #1, audio_device, from '/dev/dsp':
>   Stream #1.0: Audio: pcm_s16le, 44100 Hz, mono, 705 kb/s
> Output #0, avi, to 'test.avi':
>   Stream #0.0: Video: msmpeg4, 320x240, 20.00 fps, 200 kb/s
>   Stream #0.1: Audio: ac3, 44100 Hz, mono, 128 kb/s
> Stream mapping:
>   Stream #0.0 -> #0.0
>   Stream #1.0 -> #0.1
> Press [q] to stop encoding
> 
> 
> ffmpeg says input is yuv422 so I believe it :)

Yes but is also says that the input size is 384x144 while in fact it is
640x240 (for NTSC). So, here is your problem !

                       Vladimir Dergachev

> 
> Brain
> 
> --------------------------------
> Petr `Brain' Kulhavy
> <br...@ar...>
> http://artax.karlin.mff.cuni.cz/~brain
> Faculty of Mathematics and Physics, Charles University Prague, Czech Republic
> 
> ---
> A GOOD WAY TO THREATEN somebody is to light a stick of dynamite.  Then you
> call the guy and hold the burning fuse to the phone.  "Hear that?" you say.
> "That's dynamite, baby."
>                 -- Jack Handley, The New Mexican, 1988.
> 

On Sun, 30 Dec 2001, Vladimir Dergachev wrote:

> Are you sure ? Because, by default, ffmpeg wants yuv420. Also, km
> generates fields 640x240 in size (for NTSC) - this could also be a
> problem.

Here's ffmpeg output:

Input #0, video_grab_device, from '/dev/video':
  Stream #0.0: Video: rawvideo, yuv422, 384x144, 50.00 fps
Input #1, audio_device, from '/dev/dsp':
  Stream #1.0: Audio: pcm_s16le, 44100 Hz, mono, 705 kb/s
Output #0, avi, to 'test.avi':
  Stream #0.0: Video: msmpeg4, 320x240, 20.00 fps, 200 kb/s
  Stream #0.1: Audio: ac3, 44100 Hz, mono, 128 kb/s
Stream mapping:
  Stream #0.0 -> #0.0
  Stream #1.0 -> #0.1
Press [q] to stop encoding


ffmpeg says input is yuv422 so I believe it :)

Brain

--------------------------------
Petr `Brain' Kulhavy
<br...@ar...>
http://artax.karlin.mff.cuni.cz/~brain
Faculty of Mathematics and Physics, Charles University Prague, Czech Republic

---
A GOOD WAY TO THREATEN somebody is to light a stick of dynamite.  Then you
call the guy and hold the burning fuse to the phone.  "Hear that?" you say.
"That's dynamite, baby."
                -- Jack Handley, The New Mexican, 1988.

On Sun, 30 Dec 2001, Vladimir Dergachev wrote:

> Are you sure ? Because, by default, ffmpeg wants yuv420. Also, km
> generates fields 640x240 in size (for NTSC) - this could also be a
> problem.

How do I tell ffmpeg to use yuv422?

Brain

--------------------------------
Petr `Brain' Kulhavy
<br...@ar...>
http://artax.karlin.mff.cuni.cz/~brain
Faculty of Mathematics and Physics, Charles University Prague, Czech Republic

---
Women can keep a secret just as well as men,
but it takes more of them to do it.

On Sun, 30 Dec 2001 br...@ar... wrote:

> On Sun, 30 Dec 2001, Vladimir Dergachev wrote:
> 
> >
> >
> > On Sun, 30 Dec 2001 br...@ar... wrote:
> >
> > > Finally it's working!
> > >
> > > The problem was in wrong input size. Now I'am able to record 320x240/20Hz
> > > ac3-mpeg4 avi video.
> > >
> > > But the picture is still rotated left by 1/3 of picture width. I think somewhere in
> > > the code must be something like "address+width" instead of "address+3*width"
> > > (because we have 3 bytes per pixel: Y, U and V). It seems to be in km, because
> > > avview's screenshot has the same bug.
> >
> > This is very strange. Could it be that you specified a wrong input format ?
> > It is YUV422, not YUV420. And no, you do not have 3 bytes per pixel, but
> > only 2. Y is sampled for every pixel, but U and V are sampled
> > alternatively... (this is an oversimplification, see http://www.mir.com/DMG/ -
> > - quite a good read)
> 
> 
> No, ffmpeg says input is rawvideo yuv422, so it should be right.

Are you sure ? Because, by default, ffmpeg wants yuv420. Also, km
generates fields 640x240 in size (for NTSC) - this could also be a
problem.

                          Vladimir Dergachev

> 
> Brain
> 
> --------------------------------
> Petr `Brain' Kulhavy
> <br...@ar...>
> http://artax.karlin.mff.cuni.cz/~brain
> Faculty of Mathematics and Physics, Charles University Prague, Czech Republic
> 
> ---
> No problem is so formidable that you can't just walk away from it.
> 

On Sun, 30 Dec 2001, Vladimir Dergachev wrote:

>
>
> On Sun, 30 Dec 2001 br...@ar... wrote:
>
> > Finally it's working!
> >
> > The problem was in wrong input size. Now I'am able to record 320x240/20Hz
> > ac3-mpeg4 avi video.
> >
> > But the picture is still rotated left by 1/3 of picture width. I think somewhere in
> > the code must be something like "address+width" instead of "address+3*width"
> > (because we have 3 bytes per pixel: Y, U and V). It seems to be in km, because
> > avview's screenshot has the same bug.
>
> This is very strange. Could it be that you specified a wrong input format ?
> It is YUV422, not YUV420. And no, you do not have 3 bytes per pixel, but
> only 2. Y is sampled for every pixel, but U and V are sampled
> alternatively... (this is an oversimplification, see http://www.mir.com/DMG/ -
> - quite a good read)


No, ffmpeg says input is rawvideo yuv422, so it should be right.

Brain

--------------------------------
Petr `Brain' Kulhavy
<br...@ar...>
http://artax.karlin.mff.cuni.cz/~brain
Faculty of Mathematics and Physics, Charles University Prague, Czech Republic

---
No problem is so formidable that you can't just walk away from it.

On Sun, 30 Dec 2001 br...@ar... wrote:

> Finally it's working!
> 
> The problem was in wrong input size. Now I'am able to record 320x240/20Hz
> ac3-mpeg4 avi video.
> 
> But the picture is still rotated left by 1/3 of picture width. I think somewhere in
> the code must be something like "address+width" instead of "address+3*width"
> (because we have 3 bytes per pixel: Y, U and V). It seems to be in km, because
> avview's screenshot has the same bug.

This is very strange. Could it be that you specified a wrong input format ? 
It is YUV422, not YUV420. And no, you do not have 3 bytes per pixel, but
only 2. Y is sampled for every pixel, but U and V are sampled
alternatively... (this is an oversimplification, see http://www.mir.com/DMG/ - 
- quite a good read)

                                 Vladimir Dergachev

> 
> I set picture size in /dev/video using old xatitv (in capture menu you can set
> output size and xatitv tells the card to resize picture to given size).
> 
> Here's the hellish command :)
> 
> ffmpeg -f video_grab_device -s 384x144 -r 50 -i /dev/video -f audio_device -i
> /dev/dsp -acodec ac3 -ar 44100 -ab 128 -vcodec mpeg4 -s 320x240 -r 20 -qscale 2
> -g 3000 -me log -deinterlace -intra -y test.avi
> 
> I have several questions:
> 
> 1) Which codec is faster: mpeg4 or msmpeg4?
> 2) What does the -g option mean, what do I change with this option?
> 3) Is it possible to tell the ffmpeg to tell the card to give him pictures in
> size I want (so resizing won't be done by ffmpeg (using software) but by card
> (using HW))?
> 
> Brain
> 
> --------------------------------
> Petr `Brain' Kulhavy
> <br...@ar...>
> http://artax.karlin.mff.cuni.cz/~brain
> Faculty of Mathematics and Physics, Charles University Prague, Czech Republic
> 
> ---
> "Hello," he lied.
>                 -- Don Carpenter, quoting a Hollywood agent
> 
> 
> 
> 
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
> 

On Sun, Dec 30, 2001 at 05:48:13PM +0100, br...@ar... wr=
ote:
> Finally it's working!
>=20
> The problem was in wrong input size. Now I'am able to record 320x240/20Hz
> ac3-mpeg4 avi video.
hooray 4 brain :-)

> I set picture size in /dev/video using old xatitv (in capture menu you ca=
n set
> output size and xatitv tells the card to resize picture to given size).
I think this is the problem, it doesn't work, size is fixed (volodya said t=
hat
already).

> 1) Which codec is faster: mpeg4 or msmpeg4?
IMHO msmpeg4

> 2) What does the -g option mean, what do I change with this option?
translated to avi language it means keyframe frequency, the unit being numb=
er
of frames.

> 3) Is it possible to tell the ffmpeg to tell the card to give him pictures
> in size I want (so resizing won't be done by ffmpeg (using software) but =
by
> card (using HW))?
No, ffmpeg doesn't support the ioctl, km doesn't support it either and this
won't change in foreseeable future.

> Brain
Bye,

Peter Surda (Shurdeek) <shu...@pa...>, ICQ 10236103, +436505=
122023

--
               Dudes! May the Open Source be with you.