From: Jim Shaffer on
When I try to play a high-definition (1920x1080) .ts video file with
mplayer using the xv output driver, the video is only drawn on part of
the screen. A chunk at the right side is drawn as a pink rectangle.
This happens even if I'm not playing it full-screen (my monitor is
running at 1680x1050). Xine exhibits the same problem. I get normal
playback if I use the x11 output driver instead of xv, but then I
don't have hardware acceleration and it looks grainy even on my 2.8
GHz P4. I'm assuming this isn't a problem with the player software
unless xine and mplayer share the same code, but I don't know whether
it's a problem with the x.org software or a problem with my video
card. The card is an ATI All-In-Wonder (Radeon) 7500, and I'm using
x.org 6.8.2. mplayer displays a "aspect: Warning: no suitable new res
found!" message when using the xv output. Non-accelerated output
produces no such message. Also, the xv driver *will* scale videos of
lesser resolution up to full-screen 1680x1050 with no errors or
incorrect display.



From: Måns Rullgård on
Jim Shaffer <jmshaffer(a)alltel.net> writes:

> When I try to play a high-definition (1920x1080) .ts video file with
> mplayer using the xv output driver, the video is only drawn on part of
> the screen. A chunk at the right side is drawn as a pink rectangle.
> This happens even if I'm not playing it full-screen (my monitor is
> running at 1680x1050). Xine exhibits the same problem. I get normal
> playback if I use the x11 output driver instead of xv, but then I
> don't have hardware acceleration and it looks grainy even on my 2.8
> GHz P4. I'm assuming this isn't a problem with the player software
> unless xine and mplayer share the same code, but I don't know whether
> it's a problem with the x.org software or a problem with my video
> card. The card is an ATI All-In-Wonder (Radeon) 7500, and I'm using
> x.org 6.8.2. mplayer displays a "aspect: Warning: no suitable new res
> found!" message when using the xv output. Non-accelerated output
> produces no such message. Also, the xv driver *will* scale videos of
> lesser resolution up to full-screen 1680x1050 with no errors or
> incorrect display.

You've found a hardware limitation. Run xvinfo and check what maximum
supported overlay size is. Look for a line like this:

maximum XvImage size: 2046 x 2046

--
M?ns Rullg?rd
mru(a)inprovide.com
From: Jim Shaffer on
On Sun, 05 Feb 2006 11:15:02 +0000, M?ns Rullg?rd <mru(a)inprovide.com>
wrote:

>You've found a hardware limitation. Run xvinfo and check what maximum
>supported overlay size is. Look for a line like this:
>
> maximum XvImage size: 2046 x 2046

Mine says 2048 x 2048. That's smaller than both my video and my
monitor. So is this a bug/limitation in the x.org radeon driver, or
in both xine and mplayer?


From: Henrik Carlqvist on
Jim Shaffer <jmshaffer(a)alltel.net> wrote:
>> Run xvinfo and check what maximum
>> supported overlay size is. Look for a line like this:
>>
>> maximum XvImage size: 2046 x 2046
>
> Mine says 2048 x 2048. That's smaller than both my video and my
> monitor.

To me 2048x2048 sounds bigger than both 1680x1050 which you said you used
as monitor resolution and 1920x1080 which you said was the .ts file
resolution.

> So is this a bug/limitation in the x.org radeon driver, or
> in both xine and mplayer?

I am also using the radeon driver with my radeon 9200. Before I hadn't
tried to play any movie with bigger resolution than my screen resolution
as you are trying to do. However, by shrinking the size of a window I can
see that xv is able to scale down the video.

Now to test your problem I used transcode to scale up a short video to
1920x1080. When playing that test video using xv I get exactly the same
pink rectangle as you get. My guess is that this is a bug in the radeon xv
code. Even though 1920x1080 is smaller than 2048x2048 wich xv is supposed
to be able to handle it is bigger than my screen resolution of 1280x1024.

At start, mplayer creates a window that is bigger than my screen
resolution. The window is centered on my screen, it seems as if it is as
much to the left of my screen as the part of the window that is outside to
the right. When pressing "f" in mplayer to get fullscreen viewing the
window is resized to my screen resolution. Then I can see the left part of
the window that previously was outside my screen, but the right part that
was outside my screen is now pink. I think it is a bug in the xv radeon
code as both xine and mplayer have the same problem and the problem
disappears when using the x11 driver in mplayer.

regards Henrik
--
The address in the header is only to prevent spam. My real address is:
hc7(at)uthyres.com Examples of addresses which go to spammers:
root(a)variousus.net root(a)localhost

From: Jim Shaffer on
On Sun, 05 Feb 2006 20:08:50 +0100, Henrik Carlqvist
<Henrik.Carlqvist(a)deadspam.com> wrote:

>Jim Shaffer <jmshaffer(a)alltel.net> wrote:
>>> Run xvinfo and check what maximum
>>> supported overlay size is. Look for a line like this:
>>>
>>> maximum XvImage size: 2046 x 2046
>>
>> Mine says 2048 x 2048. That's smaller than both my video and my
>> monitor.
>
>To me 2048x2048 sounds bigger than both 1680x1050 which you said you used
>as monitor resolution and 1920x1080 which you said was the .ts file
>resolution.

Yes, I don't know why I typed "smaller".

>At start, mplayer creates a window that is bigger than my screen
>resolution. The window is centered on my screen, it seems as if it is as
>much to the left of my screen as the part of the window that is outside to
>the right. When pressing "f" in mplayer to get fullscreen viewing the
>window is resized to my screen resolution. Then I can see the left part of
>the window that previously was outside my screen, but the right part that
>was outside my screen is now pink. I think it is a bug in the xv radeon
>code as both xine and mplayer have the same problem and the problem
>disappears when using the x11 driver in mplayer.

I found some documentation of this bug at
https://bugs.freedesktop.org/show_bug.cgi?id=1462

It looks like it hasn't been fixed yet, they don't consider it
critical, and they think it's a hardware bug. But if it's a hardware
bug, it seems to exist on multiple chips. I wonder how difficult it
would be to install ATI's X drivers, if they exist for my hardware.