[FFmpeg-devel] [PATCH] png parser

[Peter Holik]

> On Wed, Jun 10, 2009 at 01:27:10PM +0200, Peter Holik wrote:
>> > On Wed, Jun 10, 2009 at 11:52:34AM +0200, Peter Holik wrote:
>> >> >>> On Thu, Jun 04, 2009 at 10:40:55AM +0200, Peter Holik wrote:
>> >> >>>> > On Mon, Jun 01, 2009 at 12:27:52PM +0200, Peter Holik wrote:
>> >> >>>> >> Hi!
>> >> >>>> >>
>> >> >>>> >> I've now reworked my patch to really parse a png for end of frame.
>> >> >>>> >>
>> >> >>>> >> Now you can filter out png's of any stream like
>> >> >>>> >>
>> >> >>>> >> cat garbage a.png garbage b.png garbage | ffmpeg -f image2pipe -vcodec png -i -
>> c%d.png
>> >> >>>> >>
>> >> >>>> >> cu Peter
>> >> >>>> >>  Makefile     |    1
>> >> >>>> >>  allcodecs.c  |    1
>> >> >>>> >>  png_parser.c |  124 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3
>> files
>> >> >> changed, 126 insertions(+)
>> >> >>>> >> 63f7cd1bb826726f1d2a9e7254410cd7ec98a81c  png-parser.patch
>> >> >>>> >> diff --git a/libavcodec/Makefile b/libavcodec/Makefile
>> >> >>>> >> index 14ef9f4..189231b 100644
>> >> >>>> >> --- a/libavcodec/Makefile
>> >> >>>> >> +++ b/libavcodec/Makefile
>> >> >>>> >> @@ -411,6 +411,7 @@ OBJS-$(CONFIG_MLP_PARSER)              += mlp_parser.o mlp.o
>> >> >>>> >>  OBJS-$(CONFIG_MPEG4VIDEO_PARSER)       += mpeg4video_parser.o h263.o mpeg12data.o
>> >> >>>> mpegvideo.o
>> >> >>>> >> error_resilience.o
>> >> >>>> >>  OBJS-$(CONFIG_MPEGAUDIO_PARSER)        += mpegaudio_parser.o mpegaudiodecheader.o
>> >> >>>> >> mpegaudiodata.o
>> >> >>>> >>  OBJS-$(CONFIG_MPEGVIDEO_PARSER)        += mpegvideo_parser.o mpeg12.o mpeg12data.o
>> >> >>>> mpegvideo.o
>> >> >>>> >> error_resilience.o
>> >> >>>> >> +OBJS-$(CONFIG_PNG_PARSER)              += png_parser.o
>> >> >>>> >>  OBJS-$(CONFIG_PNM_PARSER)              += pnm_parser.o pnm.o
>> OBJS-$(CONFIG_VC1_PARSER)
>> >> >>         += vc1_parser.o
>> >> >>>> >>  OBJS-$(CONFIG_VP3_PARSER)              += vp3_parser.o
>> >> >>>> >> diff --git a/libavcodec/allcodecs.c b/libavcodec/allcodecs.c index e8d5d27..362021f
>> >> 100644
>> >> >>>> >> --- a/libavcodec/allcodecs.c
>> >> >>>> >> +++ b/libavcodec/allcodecs.c
>> >> >>>> >> @@ -334,6 +334,7 @@ void avcodec_register_all(void)
>> >> >>>> >>      REGISTER_PARSER  (MPEG4VIDEO, mpeg4video);
>> >> >>>> >>      REGISTER_PARSER  (MPEGAUDIO, mpegaudio);
>> >> >>>> >>      REGISTER_PARSER  (MPEGVIDEO, mpegvideo);
>> >> >>>> >> +    REGISTER_PARSER  (PNG, png);
>> >> >>>> >>      REGISTER_PARSER  (PNM, pnm);
>> >> >>>> >>      REGISTER_PARSER  (VC1, vc1);
>> >> >>>> >>      REGISTER_PARSER  (VP3, vp3);
>> >> >>>> >> diff --git a/libavcodec/png_parser.c b/libavcodec/png_parser.c new file mode 100644
>> >> >>>> >> index 0000000..64f2356
>> >> >>>> >> --- /dev/null
>> >> >>>> >> +++ b/libavcodec/png_parser.c
>> >> >>>> >> @@ -0,0 +1,124 @@
>> >> >>>> >> +/*
>> >> >>>> >> + * PNG parser
>> >> >>>> >> + * Copyright (c) 2009 Peter Holik
>> >> >>>> >> + *
>> >> >>>> >> + * This file is part of FFmpeg.
>> >> >>>> >> + *
>> >> >>>> >> + * FFmpeg is free software; you can redistribute it and/or
>> >> >>>> >> + * modify it under the terms of the GNU Lesser General Public + * License as
>> published
>> >> by
>> >> >> the Free Software Foundation; either + * version 2.1 of the License, or (at your option)
>> any
>> >> >> later version. + *
>> >> >>>> >> + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY
>> >> WARRANTY;
>> >> >> without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR
>> >> >> PURPOSE.  See the GNU + * Lesser General Public License for more details.
>> >> >>>> >> + *
>> >> >>>> >> + * You should have received a copy of the GNU Lesser General Public + * License
>> along
>> >> with
>> >> >> FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth
>> >> >> Floor, Boston, MA 02110-1301 USA + */
>> >> >>>> >> +
>> >> >>>> >> +/**
>> >> >>>> >> + * @file libavcodec/png_parser.c
>> >> >>>> >> + * PNG parser
>> >> >>>> >> + */
>> >> >>>> >> +
>> >> >>>> >> +#include "parser.h"
>> >> >>>> >> +#include "bytestream.h"
>> >> >>>> >> +#include "png.h"
>> >> >>>> >> +
>> >> >>>> >> +static int png_parse(AVCodecParserContext *s, AVCodecContext *avctx, +
>> >> >> const uint8_t **poutbuf, int *poutbuf_size, +                     const uint8_t *buf, int
>> >> >> buf_size)
>> >> >>>> >> +{
>> >> >>>> >> +    ParseContext *pc = s->priv_data;
>> >> >>>> >> +    uint8_t *bytestream;
>> >> >>>> >> +    uint8_t *bytestream_start;
>> >> >>>> >> +    uint8_t *bytestream_end;
>> >> >>>> >> +
>> >> >>>> >> +    /* EOF considered as end of frame */
>> >> >>>> >> +    if (buf_size == 0)
>> >> >>>> >> +        goto flush;
>> >> >>>> >> +
>> >> >>>> >> +    *poutbuf      = NULL;
>> >> >>>> >> +    *poutbuf_size = 0;
>> >> >>>> >> +
>> >> >>>> >> +    /* first add new buffer to parsebuffer */
>> >> >>>> >> +    bytestream_start = av_fast_realloc(pc->buffer, &pc->buffer_size, +
>> >> >>     pc->index + buf_size + FF_INPUT_BUFFER_PADDING_SIZE); +    if(!bytestream_start)
>> >> >>>> >> +        goto fail;
>> >> >>>> >> +    pc->buffer = bytestream_start;
>> >> >>>> >> +    memcpy(&pc->buffer[pc->index], buf, buf_size);
>> >> >>>> >
>> >> >>>> > mandatory copying is unacceptable
>> >> >>>> > the case where each input is exactly a png frame for example should be free of any
>> >> >> timeconsuming operation
>> >> >>>> >
>> >> >>>> > maybe ff_combine_frame() would be usefull
>> >> >>>>
>> >> >>>> But ff_combine_frame also does memcpy, so there is no difference to my approach.
>> >> >>>
>> >> >>> ff_combine_frame() does not always memcpy, your code does always
>> >> >>
>> >> >> change my patch to only memcpy when input is not exactly a png frame.
>> >> >>
>> >> >> cu Peter
>> >>
>> >> ping
>> >
>> > does your code do something else than ff_combine_frame() ?
>> > if no its duplicate code
>> > if yes its either worse or ff_combine_frame() should be improved
>> >
>>
>> 1. ff_combine_frame is overhead
>
> elaborate

there is more done than i need (pc->overread, pc->state,...)

>> 2. i use a variable from PasteContext also altered by ff_combine_frame
>
> which variable is how used by you so that it conflicts with ff_combine_frame?

* i use pc->last_index to remember where to parse from, because i do not like
to parse every time from the beginning.

* to return a packet starting with a png signature i skip the garbage if any
  and need this information in my combine_frame version.


well, i will try to use  ff_combine_frame()

cu Peter