Hi, Nick: can I remove libavcodec from mplayer CVS? does it still have something interesting code, or all things are poretd to ffmpeg CVS? What about ffmpeg support? I think it could be added for ffmpeg cvs version, like: #ifdef USE_FASTMEMCPY #include "../libvo/fastmemcpy.h" #endif or so. USE_FASTMEMCPY is already defined in mplayer's config.h did you check divx4linux? it's much faster, but i can't imagine how. i think that DLL codecs are optimized enough, but it's 2 times faster in some cases! also, ffmpeg codec is usually faster than DLL, so maybe DLLs aren't so good as they said. btw divx4 is still buggy, while ffmpeg plays everything. and it's opensource, and portable, while divx4 is linux/windows only :( A'rpi / Astral & ESP-team -- mailto:arpi@thot.banki.hu http://esp-team.scene.hu _______________________________________________ Mplayer-dev-eng mailing list Mplayer-dev-eng@lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng
Hi, Arpi! On 2001-08-01 00:03:18 you wrote:
Hi,
Nick: can I remove libavcodec from mplayer CVS? does it still have something interesting code, or all things are poretd to ffmpeg CVS?
I don't know - Gerard did not accepted my code which allowed function inlining and have replaced it with double function's pointer. Instead he have performed some C-level optimization. He wants to run 386 optimized ffmpeg on non-MMX (and non-FPU) Intel cpus.
What about ffmpeg support? I think it could be added for ffmpeg cvs version, like: #ifdef USE_FASTMEMCPY #include "../libvo/fastmemcpy.h" #endif
or so.
USE_FASTMEMCPY is already defined in mplayer's config.h
I'm not sure that this code will live longer a several hours in ffmpeg-CVS :(
did you check divx4linux? it's much faster, but i can't imagine how. After purchasing a new video hardware - I forgot, what is non real-time playback. Thus codec's speed is interested for me only from theoretical point now.
i think that DLL codecs are optimized enough, but it's 2 times faster in some cases! also, ffmpeg codec is usually faster than DLL, so maybe DLLs aren't so good as they said.
AFAIK divx32c has enough aggressive MMX(1,2) optimized code so it's a little strange for me - maybe you should test a several movies?
btw divx4 is still buggy, while ffmpeg plays everything. and it's opensource, and portable, while divx4 is linux/windows only :(
Undoubtedly - mplayer must have ffmpeg stuff in any form since it's only solution for non Intel machines. Best regards! Nick (2001-08-01 10:14:27) _______________________________________________ Mplayer-dev-eng mailing list Mplayer-dev-eng@lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng
Hello! I've commited fastmemcpy stuff to ffmpeg. Please wait during several days - if Gerard will not suppress it then remove libavcodec stuff in mplayer. Best regards! Nick _______________________________________________ Mplayer-dev-eng mailing list Mplayer-dev-eng@lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng
Hi,
Hello!
I've commited fastmemcpy stuff to ffmpeg. Please wait during several days - if Gerard hehe. Gerard is away nowdays, se it will remain for a few days there :)
will not suppress it then remove libavcodec stuff in mplayer. ok
A'rpi / Astral & ESP-team -- mailto:arpi@thot.banki.hu http://esp-team.scene.hu _______________________________________________ Mplayer-dev-eng mailing list Mplayer-dev-eng@lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng
Hi,
I've commited fastmemcpy stuff to ffmpeg. I've fixed it.
You used #ifdef HAVE_CONFIG_H, but it is never defined. gerard uses HAVE_AV_CONFIG_H. But later I've removed it totally, because config.h is already included from common.h. this is why it worked for you, and you didn't get warning. your include was never executed. A'rpi / Astral & ESP-team -- mailto:arpi@thot.banki.hu http://esp-team.scene.hu _______________________________________________ Mplayer-dev-eng mailing list Mplayer-dev-eng@lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng
participants (2)
-
Arpi -
Nick Kurshev