Dmitry Borisov | 1 Nov 05:31 2004
Picon

ANN: PyMedia-1.2.3.0-pre1 on Windows.

Hi everybody.
Just wanted to introduce first release candidate for PyMedia-1.2.3.0
Right now I have Windows installer for prerelease version. You can download
it here:
http://pymedia.org/downloads/pymedia-1.2.3.0.win32-py2.3.exe

I appreciate if you try it out and let me know if you encounter any problems
with it
I still need to fix some things about PTS handling when creating movies, but
it will require somebody who's familiar with that to help me out.

What is it ?
----------------------------------------------------------------
PyMedia( http://pymedia.org ) is a media module for Python, based on the
following libraries:
  libavcodec( 0.4.8 ),  libavformat( 0.4.8 ),  libdvdcss, libdvdread
Supports following formats( extensions ):
    avi, mp3, divx, xvid, ac3, vob, mpeg, mp3, wma, mov, asf
Allows to manipulate the audio/video frame by frame, convert to/from any
supported format, playback DVD with HW support on some platforms and much
more...

It has simple interface and make it a snap multimedia programming in Python.
Check out this incomplete tutorial to see how simple is that:
http://pymedia.org/tut/

What's changed/fixed:
----------------------------------------------------------------

- VFrame has new method: convert(). Allow you to convert it to any other
(Continue reading)

Moritz Naumann | 1 Nov 14:00 2004

Re: Pseudo-stream playback of ffmpeg-generated mpeg2 streams fails on proprietary players

Tom Harper schrieb:
> At 10:18 AM 10/29/2004, Moritz Naumann wrote:
>> But when downloading and playing back at the same time (pseudo 
>> streaming), both players stop playback after a second.
[..]
> I have encountered the same issue- you may be able to fix this by
> changing the rc_max_rate number, which should be some reasonable
> value for the data that you are trying to encode.
> 
> With mpeg1 I have had to set video_rc_max_rate and video_bit_rate
> both to the same number to completely resolve this issue.
> 
> I am not sure which flags you can use for these but I know that they
> are settable options in ffmpeg.c.

Hey Tom and thanks for your hints. They helped me to workaround the bug. 
As I was too lazy to recompile I tried to figure out the ffmpeg binary 
call-time flags and succeeded:

-b $BITRATE_IN_KBITPS
-maxrate $MAX_BITRATE_TOLERANCE_IN_KBITPS
-bufsize $RATE_CONTROL_BUFFER_SIZE_IN_KB

Here's an example call:
ffmpeg -y -v 1 -i /firewire/3_in_pixelboxx_laden/abraftb2.mpg -b 100 
-maxrate 500 -bufsize 40 -s 354x288 -croptop 20 -cropbottom 20 -cropleft 
20 -cropright 20 -title "abraftb2.mpg" 
/usr/local/Pixelboxx-1.1-media235/pb-web/doc/streams/files/abraftb2.mpg

I don't know if it's possible to use -q instead of -b (didn't test).
(Continue reading)

compn | 1 Nov 16:10 2004
Picon

Fw: MPlayer ffv1.c fix for compile problem

"Eric Auer" <eric <at> CoLi.Uni-SB.DE> wrote:
> Hi, with gcc 3.3.1, the compiler gets an internal error with:
> 
>             for(p=0; p<3; p++)
>                 sample[p][i]= sample_buffer[p][(h+i-y)%ring_size]+3;
> 
> in ffv1.c around line 460. I replaced that by:
> 
>                 sample[0][i]= sample_buffer[0][(h+i-y) & 1]+3;
>                 sample[1][i]= sample_buffer[1][(h+i-y) & 1]+3;
>                 sample[2][i]= sample_buffer[2][(h+i-y) & 1]+3;
> 
> and now it works fine. Simple enough :-).
> 
> (that is of course a dirty patch because I hardcoded ring_size,
> but a more elegant patch would work, too...)
> 
> Eric

just posting for someone who does not wish to subscribe. do not cc me! :)

-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
plaum | 1 Nov 23:21 2004
Picon

libavcodec bug

Using todays CVS, configure called without options

While valgrinding ffplay_g, the following messages show up with most
 .mp4-Files (video fourcc mp4v), I tried, one example is at

http://www.archive.org/download/DuckandC1951/DuckandC1951.mp4
Video fourcc: mp4v

==2267== Thread 3:
==2267== Invalid read of size 4
==2267==    at 0x8215705: decode_vol_header (bswap.h:30)
==2267==    by 0x8211285: ff_mpeg4_decode_picture_header (h263.c:5969)
==2267==    by 0x82220B6: ff_h263_decode_frame (h263dec.c:480)
==2267==    by 0x809B7C3: avcodec_decode_video (utils.c:516)
==2267==    by 0x805A53C: try_decode_frame (utils.c:1626)
==2267==    by 0x80581A8: av_find_stream_info (utils.c:1743)
==2267==    by 0x805603C: decode_thread (ffplay.c:1324)
==2267==    by 0x402C36BA: SDL_RunThread (in /usr/lib/libSDL-1.2.so.0.0.5)
==2267==    by 0x402C38CE: (within /usr/lib/libSDL-1.2.so.0.0.5)
==2267==    by 0x402EE521: thread_wrapper (vg_libpthread.c:667)
==2267==    Address 0x41E1A5CB is 27 bytes inside a block of size 29 alloc'd
==2267==    at 0x40028C83: memalign (vg_replace_malloc.c:328)
==2267==    by 0x809C0BF: av_malloc (mem.c:56)
==2267==    by 0x809A538: av_mallocz (utils.c:40)
==2267==    by 0x806E56E: mov_read_esds (mov.c:572)
==2267==    by 0x806E144: mov_read_default (mov.c:381)
==2267==    by 0x806EB08: mov_read_stsd (mov.c:960)
==2267==    by 0x806E144: mov_read_default (mov.c:381)
==2267==    by 0x806E144: mov_read_default (mov.c:381)
==2267==    by 0x806E144: mov_read_default (mov.c:381)
(Continue reading)

Michael Niedermayer | 1 Nov 17:53 2004
Picon
Picon

Re: Fw: MPlayer ffv1.c fix for compile problem

Hi

On Monday 01 November 2004 16:10, compn wrote:
> "Eric Auer" <eric <at> CoLi.Uni-SB.DE> wrote:
> > Hi, with gcc 3.3.1, the compiler gets an internal error with:

this is ffmpeg-devel, not gcc-devel
i cant speak for the other ffmpeg developers but i will not apply a gcc bug 
workaround unless it either affects 2.95 or all gcc 3.* versions i have 
installed
especially not if its not a patch at all, but a copy and pasted piece of 
broken code

[...]
> just posting for someone who does not wish to subscribe. do not cc me! :)

ffmpeg-dev is not subscribers only, so no need to be subscribed

[...]
--

-- 
Michael

"I do not agree with what you have to say, but I'll defend to the death your
right to say it." -- Voltaire

-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
(Continue reading)

Michel Bardiaux | 2 Nov 10:22 2004
Picon

Re: How to ad a wav file to a mpeg audio stream?

soledad alborno wrote:
> Hello, I am working with ffmpeg v.0.4.9_pre1 in visual c++ compiled with 
> msys and mgw.
> I need to create a mpeg file with a couple of ppm files in video stream 
> and a wav file in audio stream.
> I could add ppm files using img_convert.
> But now I am having troubles in add the audio stream. My wav file has 
> this characteristics:
> - Audio PCM 8000 KHz, 16 bit, Mono channel. bitrate: 128 kb/s
>  
> Does anybody know how to add this file to the mpeg's audio stream, there 
> are any convert function to pass the CODEC_ID_PCM_S16LE format to a 
> CODEC_ID_MP2?

8KHz (I assume 8000KHz is a typo...) is not a legal sampling rate for an 
MPEG file, you will have to resample the audio to, say, 32KHz.

>  
> Thank you in advance.
> Soledad
> 

HaND,
--

-- 
Michel Bardiaux
Peaktime Belgium S.A.  Bd. du Souverain, 191  B-1160 Bruxelles
Tel : +32 2 790.29.41

-------------------------------------------------------
This SF.Net email is sponsored by:
(Continue reading)

Erik Brännlund (SA/TMS | 2 Nov 10:31 2004
Picon

Last (hopefully) muxing question...

Hi and thanx for the previous responses to my muxing questions. 

I have finally managed to remux my TS to a PS that dvdauthor turns into working DVD.
However this only works on a small recording. When I take a longer recording dvdauthor stops saying 
"WARN: Audio pts for channel _ moves backwards; please remultiplex input.".

Any hints on why this is happening. I dont seem to find any PTS:s moving 
backwards when I parse the file in my own code.

/Erik

-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
Michel Bardiaux | 2 Nov 10:49 2004
Picon

[PATCH] To avoid a rash of "enumeration value `CODEC_ID_...' not handled in switch"

Which happens with current CVS, gcc2.95, and
./configure  --enable-mp3lame --disable-strip --enable-amr_nb 
--enable-a52 --enable-gpl

HaND,
-- 
Michel Bardiaux
Peaktime Belgium S.A.  Bd. du Souverain, 191  B-1160 Bruxelles
Tel : +32 2 790.29.41
Index: libavcodec/dpcm.c
===================================================================
RCS file: /cvsroot/ffmpeg/ffmpeg/libavcodec/dpcm.c,v
retrieving revision 1.6
diff -d -u -r1.6 dpcm.c
--- libavcodec/dpcm.c	28 Sep 2004 03:09:48 -0000	1.6
+++ libavcodec/dpcm.c	2 Nov 2004 09:46:05 -0000
 <at>  <at>  -282,6 +282,7  <at>  <at> 
             }
         }
         break;
+    default:;
     }

     *data_size = out * sizeof(short);
Index: libavcodec/mpegvideo.c
===================================================================
RCS file: /cvsroot/ffmpeg/ffmpeg/libavcodec/mpegvideo.c,v
retrieving revision 1.443
(Continue reading)

Måns Rullgård | 2 Nov 11:34 2004

Re: Last (hopefully) muxing question...

Erik Brännlund (SA/TMS) <erik.brannlund <at> ericsson.com> writes:

> Hi and thanx for the previous responses to my muxing questions.
>
> I have finally managed to remux my TS to a PS that dvdauthor turns
> into working DVD.  However this only works on a small recording. When
> I take a longer recording dvdauthor stops saying "WARN: Audio pts for
> channel _ moves backwards; please remultiplex input.".
>
> Any hints on why this is happening. I dont seem to find any PTS:s
> moving backwards when I parse the file in my own code.

Without looking at the source code of dvdauthor, I can think of two
reasons for such a message.  You might have drift between your audio
and video PTS, or you have a problem with the SCR values such that a
PTS happens to be less than the last SCR, which is typically bad.

--

-- 
Måns Rullgård
mru <at> inprovide.com

-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_idU88&alloc_id065&op=click
Måns Rullgård | 2 Nov 11:35 2004

Re: [PATCH] To avoid a rash of "enumeration value `CODEC_ID_...' not handled in switch"

Michel Bardiaux <mbardiaux <at> peaktime.be> writes:

> Which happens with current CVS, gcc2.95, and
> ./configure  --enable-mp3lame --disable-strip --enable-amr_nb
> --enable-a52 --enable-gpl

I can sense a "gcc is broken. rejected" coming along...

--

-- 
Måns Rullgård
mru <at> inprovide.com

-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_idU88&alloc_id065&op=click

Gmane