Latest And Possibily Last Of Atty's Mplayer

[tovarish]

It is available here.
http://atty303.googlepages.com/home
Meanie maybe you can repackage it for pdaxrom. Can you also do a favour. Your package is for cxxxx but i have a c860 so i cant use the bvdd module. Can you update your package such that if a c860 is detected then the bvdd module is not loaded and the driver w100 is used (-vo w100).

tovarish

[danboid]

Great news, and just in time for inclusion in pdaXii13 5.3!

Does anyone know whats changed 1.2.0?

[kkazakov13]

No idea, but it's bigger

I'm updating my mplayer package to have it instead of the previous one. Thanks tovarish for the tip.

[tovarish]

can you also make your package more friendly for us c7/8xx users who dont have bvdd and iwmmxt. Maybe you can have 2 packages, one basic which uses the video driver w100 and and additional packace which is for the cxxxx users.

The changes what i have been able to decipher is better wma/v support. More codecs from using more recent ffmpeg code and mplayer code.

tovarish

[Jon_J]

I haven't tried this with any of my "difficult" videos yet. (I have about 75 "difficult" videos)
I installed this on my Cacko C3100
The size difference is negligible.
First, I uninstalled mplayer-bvdd-iwmmxt_1.1.5-1_arm.ipk

Code: [Select]

df -h
                      used         available
no mplayer           54.6M        47.4M
mplayer 1.1.5-1      58.1M        43.9M
mplayer 1.2.0-1      58.5M        43.5MOnly a difference of .4M (MB)
EDIT:
In Cacko, zplayer has problems, but kino2 seems to play this new video just fine.    
There is just a little stuttering, but previously, very few of these played well enough to watch them.
5 minutes into this video, it gets about 2-3 seconds out of sync (audio)
Last night's episode of "Lost" on TV
lost.308.flashes.before.your.eyes.hdtv.xvid-xor.avi (349MB)
Here's the stats of this video reported by kino2
Video
Size: 608x336 | Codec :XVID
Bitrate: 1013456 | FPS: 23.976
Audio
Codec: mad | Rate: 48000
Bitrate: 128000 | Channels: 2

I forgot to mention, I have overclocking enabled in kino2, set to:
Turbo (624/208/104MHz)

[kkazakov13]

I'm having problems with this version, specially pressing "O" (OSD) - the movie freezes. can anyone else try and tell if it's true?

[Serge]

I'm having problems with this version, specially pressing "O" (OSD) - the movie freezes. can anyone else try and tell if it's true?
[div align=\"right\"][{POST_SNAPBACK}][/a][/div]

I also encountered this bug: [a href=\"https://garage.maemo.org/tracker/index.php?func=detail&aid=254&group_id=54&atid=269]https://garage.maemo.org/tracker/index.php?..._id=54&atid=269[/url]

Compilation with gcc 4.1 seems to resolve the issue. Actually it may be not a gcc bug, but something wrong in new mplayer code that causes undefined behaviour (which happens to be ok in latest gcc, but cause a freeze in 3.x). Maybe somebody should contact upstream mplayer developers

[speculatrix]

why is atty discontinuing bvdd-iwmmt?

[Meanie]

why is atty discontinuing bvdd-iwmmt?
[div align=\"right\"][a href=\"index.php?act=findpost&pid=154552\"][{POST_SNAPBACK}][/a][/div]

maybe because the zaurus is EOP?

[Serge]

why is atty discontinuing bvdd-iwmmt?
[div align=\"right\"][{POST_SNAPBACK}][/a][/div]

Thanks to atty for all that he has done, but it is possible to move on.

The first question, is atty's build any better than normal mplayer in terms of media decoding performance? If it really is better in some respects, where can be the sources for his build downloaded (or public repository with the possibility to do checkout)? Anything useful should be submitted upstream for all ARM powered linux devices to benefit from these optimizations.

As for bvdd and w100, I think that you should submit these video output modules for inclusion into upstream mplayer, but if you are happy with applying patches, that's ok. Feel free to join this thread if you are interested in improving mplayer: [a href=\"https://www.oesf.org/forums/index.php?showtopic=22280]https://www.oesf.org/forums/index.php?showtopic=22280[/url]

[tovarish]

the sources for 1.1.5 were/are available. I even used it to compile with really optimised settings and it still works faster than any of the mplayers for zaurus.
Sadly 1.20 sources are not available anywhere and i dont know japanese to ask him

[Serge]

the sources for 1.1.5 were/are available. I even used it to compile with really optimised settings and it still works faster than any of the mplayers for zaurus.

Yes, I had a look at 1.1.5 sources some time ago. Seems like most of the ARM decoder optimizations are already applied in upstream mplayer/ffmpeg. Did you benchmark atty's build and have any numbers? I wonder if it is IPP that makes the difference.

Sadly 1.20 sources are not available anywhere and i dont know japanese to ask him
[div align=\"right\"][a href=\"index.php?act=findpost&pid=154577\"][{POST_SNAPBACK}][/a][/div]

Surely it would be interesting to have a look at these sources to see what was changed. Not publishing the sources sounds like a GPL violation

[koen]

the sources for 1.1.5 were/are available. I even used it to compile with really optimised settings and it still works faster than any of the mplayers for zaurus.

Yes, I had a look at 1.1.5 sources some time ago. Seems like most of the ARM decoder optimizations are already applied in upstream mplayer/ffmpeg. Did you benchmark atty's build and have any numbers? I wonder if it is IPP that makes the difference.

[div align=\"right\"][a href=\"index.php?act=findpost&pid=154706\"][{POST_SNAPBACK}][/a][/div]

atty stopped being faster than upstream a few months ago. And since Ximageon now supports XV there's no need for  vo_w100 anymore as well.

[tovarish]

actually his w100 code comes from pdaxrom's aticore itself so theoretically it was never needed.