headers
Michael Niedermayer CVS | 1 Dec 2004 03:28
Picon

CVS: ffmpeg/libavformat mpeg.c,1.75,1.76 avformat.h,1.111,1.112 utils.c,1.118,1.119

Update of /cvsroot/ffmpeg/ffmpeg/libavformat
In directory mail:/var2/tmp/cvs-serv7400/libavformat

Modified Files:
	mpeg.c avformat.h utils.c 
Log Message:
fifo_realloc()

Index: mpeg.c
===================================================================
RCS file: /cvsroot/ffmpeg/ffmpeg/libavformat/mpeg.c,v
retrieving revision 1.75
retrieving revision 1.76
diff -u -d -r1.75 -r1.76
--- mpeg.c	23 Nov 2004 22:25:12 -0000	1.75
+++ mpeg.c	1 Dec 2004 02:28:28 -0000	1.76
 <at>  <at>  -391,7 +391,7  <at>  <at> 
         default:
             return -1;
         }
-        fifo_init(&stream->fifo, 2*stream->max_buffer_size + 100*MAX_PAYLOAD_SIZE); //FIXME think about
the size maybe dynamically realloc
+        fifo_init(&stream->fifo, 16);
         stream->next_packet= &stream->premux_packet;
     }
     bitrate = 0;
 <at>  <at>  -1172,11 +1172,6  <at>  <at> 
         stream->predecode_packet= pkt_desc;
     stream->next_packet= &pkt_desc->next;
(Continue reading)

Permalink | Reply | 0 comments |
headers
Guilhem Tardy | 1 Dec 2004 04:16
Picon
Favicon
Subject: Re: CVS: ffmpeg/libavcodec imgconvert.c,1.55,1.56

Re: CVS: ffmpeg/libavcodec imgconvert.c,1.55,1.56

> > Could you confirm whether you intend to LGPL the swcaler (granted
> > that the related MPlayer G2 condition is met) for MPlayer G2 only,
> > or for all other applications too?
> 
> how could someone LGPL something only for a specific application?

There are GPL applications (e.g. GnomeMeeting) that have explicit exceptions
for a set of libraries. I suppose anything is possible, but I take your answer
as a "plain LGPL" (for all purposes), granted that your MPlayer G2 condition is
met.

		
__________________________________ 
Do you Yahoo!? 
The all-new My Yahoo! - Get yours free! 
http://my.yahoo.com 

-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
Permalink | Reply | 1 comment |
headers
D Richard Felker III | 1 Dec 2004 07:20
Subject: Re: Re: CVS: ffmpeg/libavcodec imgconvert.c,1.55,1.56

Re: Re: CVS: ffmpeg/libavcodec imgconvert.c,1.55,1.56

On Tue, Nov 30, 2004 at 07:16:46PM -0800, Guilhem Tardy wrote:
> > > Could you confirm whether you intend to LGPL the swcaler (granted
> > > that the related MPlayer G2 condition is met) for MPlayer G2 only,
> > > or for all other applications too?
> > 
> > how could someone LGPL something only for a specific application?
> 
> There are GPL applications (e.g. GnomeMeeting) that have explicit exceptions
> for a set of libraries.

This is very different though. The additional license granted in these
"exceptions" (not quite the right word) is not like the GPL or LGPL,
both of which require anyone receiving the code to be able to license
it to third parties under the same license. This is not rocket
science, people, it's logical thinking 101.

Rich

-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
Permalink | Reply | 0 comments |
headers
Luca Abeni | 1 Dec 2004 18:26
Picon
Favicon
Subject: Re: CVS: ffmpeg/libavcodec imgconvert.c,1.55,1.56

Re: CVS: ffmpeg/libavcodec imgconvert.c,1.55,1.56

Hi Michael,

On Mon, 2004-11-29 at 12:42, Michael Niedermayer wrote:
> 4:2:0 -> 4:2:2 upsamples the chroma, or in other words it guesses 2n samples 
> from n samples, its obviously not possible to guess the samples correctly 
> unless there are some additional constraints placed on the data, like 
> assuming its bandlimited, but that one leads to very low quality ...
> another option is to approximate the samples with some linear filter or use 
> some non linear method to select the most likely values based upon 
> downsampling some test images ...
Ok. I was just trying to understand how such filters are designed.

[...]
> the swscaler supports both accurate/slow and inaccurate/fast conversation in 
> most cases, u have just found the inaccurate/fast variant ...
Opss... I guess I just showed how small is my knowledge of the swscaler
code... ;-)
After this, I decided to try to directly use swscaler, without trying to
understand the algorithms.

Here is my plan:
step 1) include in libavformat support for linking and using an external
swscaler (compiled inside the mplayer tree). This will depend on
--enable-gpl. With this option enabled, libavcodec will use swscaler for
implementing img_convert() and img_resample().
Some glue code will be needed (example: IMGFMT_* <---> PIX_FMT_*
conversion)
Question: do you think that the img_convert() and img_resample()
interfaces are ok, or should they been changed? (for example,
img_convert() does not have any context...).
(Continue reading)

Permalink | Reply | 1 comment |
headers
Luca Abeni | 1 Dec 2004 18:30
Picon
Favicon
Subject: Re: CVS: ffmpeg/libavcodec imgconvert.c,1.55,1.56

Re: CVS: ffmpeg/libavcodec imgconvert.c,1.55,1.56

Hi Richard,

On Mon, 2004-11-29 at 16:28, D Richard Felker III wrote:
> On Mon, Nov 29, 2004 at 12:37:29PM +0100, Luca Abeni wrote:
[...]
> the algorithm in ilpack is made to handle interlaced content
> correctly. if you don't care about this, it's much simpler and looks
> slightly better for progressive content, but will horribly mess up
> interlaced content.
Ok, thanks for the info; I'll have a look at it.
Anyway, I finally decided to go for using swscaler without trying to
understand the algorithm (see my previous mail).

			Thanks,
				Luca
--

-- 
_____________________________________________________________________________
Copy this in your signature, if you think it is important:
                               N O    W A R ! ! !

-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
Permalink | Reply | 1 comment |
headers
D Richard Felker III | 2 Dec 2004 00:06
Subject: Re: CVS: ffmpeg/libavcodec imgconvert.c,1.55,1.56

Re: CVS: ffmpeg/libavcodec imgconvert.c,1.55,1.56

On Wed, Dec 01, 2004 at 06:30:45PM +0100, Luca Abeni wrote:
> Hi Richard,
> 
> On Mon, 2004-11-29 at 16:28, D Richard Felker III wrote:
> > On Mon, Nov 29, 2004 at 12:37:29PM +0100, Luca Abeni wrote:
> [...]
> > the algorithm in ilpack is made to handle interlaced content
> > correctly. if you don't care about this, it's much simpler and looks
> > slightly better for progressive content, but will horribly mess up
> > interlaced content.
> Ok, thanks for the info; I'll have a look at it.
> Anyway, I finally decided to go for using swscaler without trying to
> understand the algorithm (see my previous mail).

swscaler is very good, but imo it's overkill for the trivial case
where you're scaling by powers of two. you can write really simple
interpolation code.

rich

-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
Permalink | Reply | 0 comments |
headers
Michael Niedermayer | 3 Dec 2004 02:23
Picon
Picon
Subject: Re: CVS: ffmpeg/libavcodec imgconvert.c,1.55,1.56

Re: CVS: ffmpeg/libavcodec imgconvert.c,1.55,1.56

Hi

On Wednesday 01 December 2004 18:26, Luca Abeni wrote:
> Hi Michael,
>
> On Mon, 2004-11-29 at 12:42, Michael Niedermayer wrote:
> > 4:2:0 -> 4:2:2 upsamples the chroma, or in other words it guesses 2n
> > samples from n samples, its obviously not possible to guess the samples
> > correctly unless there are some additional constraints placed on the
> > data, like assuming its bandlimited, but that one leads to very low
> > quality ... another option is to approximate the samples with some linear
> > filter or use some non linear method to select the most likely values
> > based upon downsampling some test images ...
>
> Ok. I was just trying to understand how such filters are designed.
>
> [...]
>
> > the swscaler supports both accurate/slow and inaccurate/fast conversation
> > in most cases, u have just found the inaccurate/fast variant ...
>
> Opss... I guess I just showed how small is my knowledge of the swscaler
> code... ;-)
> After this, I decided to try to directly use swscaler, without trying to
> understand the algorithms.
>
> Here is my plan:
> step 1) include in libavformat support for linking and using an external
> swscaler (compiled inside the mplayer tree). This will depend on
> --enable-gpl. With this option enabled, libavcodec will use swscaler for
(Continue reading)

Permalink | Reply | 0 comments |
headers
Michael Niedermayer CVS | 5 Dec 2004 03:46
Picon

CVS: ffmpeg/libavformat mpeg.c,1.76,1.77

Update of /cvsroot/ffmpeg/ffmpeg/libavformat
In directory mail:/var2/tmp/cvs-serv2192

Modified Files:
	mpeg.c 
Log Message:
This patch takes into account that fifo_realloc may adjust fifo.wptr
patch by ("Chris" chris at garveycocker dot com)

Index: mpeg.c
===================================================================
RCS file: /cvsroot/ffmpeg/ffmpeg/libavformat/mpeg.c,v
retrieving revision 1.76
retrieving revision 1.77
diff -u -d -r1.76 -r1.77
--- mpeg.c	1 Dec 2004 02:28:28 -0000	1.76
+++ mpeg.c	5 Dec 2004 02:46:00 -0000	1.77
 <at>  <at>  -1172,6 +1172,8  <at>  <at> 
         stream->predecode_packet= pkt_desc;
     stream->next_packet= &pkt_desc->next;

+    fifo_realloc(&stream->fifo, fifo_size(&stream->fifo, NULL) + size + 1);
+
     if (s->is_dvd){
         if (is_iframe) {
             stream->fifo_iframe_ptr = stream->fifo.wptr;
 <at>  <at>  -1181,7 +1183,6  <at>  <at> 
         }
     }
(Continue reading)

Permalink | Reply | 0 comments |
headers
Michael Niedermayer CVS | 6 Dec 2004 00:21
Picon

CVS: ffmpeg/libavcodec rv10.c,1.59,1.60

Update of /cvsroot/ffmpeg/ffmpeg/libavcodec
In directory mail:/var2/tmp/cvs-serv6050/libavcodec

Modified Files:
	rv10.c 
Log Message:
next try at fixing rv20 decoding

Index: rv10.c
===================================================================
RCS file: /cvsroot/ffmpeg/ffmpeg/libavcodec/rv10.c,v
retrieving revision 1.59
retrieving revision 1.60
diff -u -d -r1.59 -r1.60
--- rv10.c	27 Nov 2004 18:10:05 -0000	1.59
+++ rv10.c	5 Dec 2004 23:21:39 -0000	1.60
 <at>  <at>  -350,6 +350,13  <at>  <at> 
     }
     av_log(s->avctx, AV_LOG_DEBUG, "\n");
 #endif
+#if 0
+    for(i=0; i<s->avctx->extradata_size; i++){
+        av_log(s->avctx, AV_LOG_DEBUG, "%2X ", ((uint8_t*)s->avctx->extradata)[i]);
+        if(i%4==3) av_log(s->avctx, AV_LOG_DEBUG, " ");
+    }
+    av_log(s->avctx, AV_LOG_DEBUG, "\n");
+#endif

     if(s->avctx->sub_id == 0x30202002 || s->avctx->sub_id == 0x30203002){
         if (get_bits(&s->gb, 3)){
(Continue reading)

Permalink | Reply | 0 comments |
headers
Michael Niedermayer CVS | 6 Dec 2004 00:21
Picon

CVS: ffmpeg/libavformat rm.c,1.29,1.30

Update of /cvsroot/ffmpeg/ffmpeg/libavformat
In directory mail:/var2/tmp/cvs-serv6050/libavformat

Modified Files:
	rm.c 
Log Message:
next try at fixing rv20 decoding

Index: rm.c
===================================================================
RCS file: /cvsroot/ffmpeg/ffmpeg/libavformat/rm.c,v
retrieving revision 1.29
retrieving revision 1.30
diff -u -d -r1.29 -r1.30
--- rm.c	19 Nov 2004 18:16:36 -0000	1.29
+++ rm.c	5 Dec 2004 23:21:39 -0000	1.30
 <at>  <at>  -634,6 +634,7  <at>  <at> 
                 /* ra type header */
                 rm_read_audio_stream_info(s, st, 0);
             } else {
+                int fps, fps2;
                 if (get_le32(pb) != MKTAG('V', 'I', 'D', 'O')) {
                 fail1:
                     av_log(&st->codec, AV_LOG_ERROR, "Unsupported video codec\n");
 <at>  <at>  -647,14 +648,24  <at>  <at> 
                 st->codec.width = get_be16(pb);
                 st->codec.height = get_be16(pb);
                 st->codec.frame_rate_base= 1;
-                st->codec.frame_rate = get_be16(pb) * st->codec.frame_rate_base;
+                fps= get_be16(pb);
(Continue reading)

Permalink | Reply | 0 comments |

Gmane