Opened 9 years ago

Closed 5 months ago

#4609 closed defect (fixed)

webp muxer silently seeks backwards when converting bigger png inputs and produce borked results when trying to write the length in the header

Reported by: ronag Owned by:
Priority: normal Component: avformat
Version: git-master Keywords: webp
Cc: Jorropo Blocked By:
Blocking: Reproduced by developer: no
Analyzed by developer: no

Description

The webp muxer needs to seek backwards.

Currently when e.g. outputting to a pipe the output will be corrupt.

ffmpeg -i file.png -f webp - > test.webp

Should probably throw an error at least?

Though, I believe it should be possible to implement the muxer in such a way that it doesn't write to output until the header has all required information.

Change History (7)

comment:1 by Carl Eugen Hoyos, 9 years ago

Keywords: muxer removed

Please provide the command line that allows to reproduce the issue together with the complete, uncut console output to make this a valid ticket, use ffmpeg -f lavfi -i testsrc -vframes 1

comment:2 by ronag, 9 years ago

$ ffmpeg -f lavfi -i testsrc -frames:v 1 -f webp - > test.webp
ffmpeg version 2.6.3 Copyright (c) 2000-2015 the FFmpeg developers
  built with Apple LLVM version 6.1.0 (clang-602.0.53) (based on LLVM 3.6.0svn)
  configuration: --prefix=/usr/local/Cellar/ffmpeg/2.6.3 --enable-shared --enable-pthreads --enable-gpl --enable-version3 --enable-hardcoded-tables --enable-avresample --cc=clang --host-cflags= --host-ldflags= --enable-libx264 --enable-libmp3lame --enable-libvo-aacenc --enable-libxvid --enable-libfreetype --enable-libvorbis --enable-libvpx --enable-librtmp --enable-ffplay --enable-libfdk-aac --enable-libopus --enable-libwebp --enable-nonfree --enable-vda
  libavutil      54. 20.100 / 54. 20.100
  libavcodec     56. 26.100 / 56. 26.100
  libavformat    56. 25.101 / 56. 25.101
  libavdevice    56.  4.100 / 56.  4.100
  libavfilter     5. 11.102 /  5. 11.102
  libavresample   2.  1.  0 /  2.  1.  0
  libswscale      3.  1.101 /  3.  1.101
  libswresample   1.  1.100 /  1.  1.100
  libpostproc    53.  3.100 / 53.  3.100
Input #0, lavfi, from 'testsrc':
  Duration: N/A, start: 0.000000, bitrate: N/A
    Stream #0:0: Video: rawvideo (RGB[24] / 0x18424752), rgb24, 320x240 [SAR 1:1 DAR 4:3], 25 tbr, 25 tbn, 25 tbc
Output #0, webp, to 'pipe:':
  Metadata:
    encoder         : Lavf56.25.101
    Stream #0:0: Video: webp (libwebp), bgra, 320x240 [SAR 1:1 DAR 4:3], q=2-31, 200 kb/s, 25 fps, 1k tbn, 25 tbc
    Metadata:
      encoder         : Lavc56.26.100 libwebp
Stream mapping:
  Stream #0:0 -> #0:0 (rawvideo (native) -> webp (libwebp))
Press [q] to stop, [?] for help
[libwebp @ 0x7fe3fc806600] Using libwebp for RGB-to-YUV conversion. You may want to consider passing in YUV instead for lossy encoding.
frame=    1 fps=0.0 q=-1.0 Lsize=       3kB time=00:00:00.04 bitrate= 605.6kbits/s    
video:3kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.132275%
Last edited 9 years ago by ronag (previous) (diff)

comment:3 by Carl Eugen Hoyos, 9 years ago

(I cannot reproduce the issue here.)
Please test current FFmpeg git head.

comment:4 by ronag, 9 years ago

Resolution: fixed
Status: newclosed

I can also verify that it works with current head.

comment:5 by Carl Eugen Hoyos, 9 years ago

Resolution: fixedinvalid

Thank you for testing again!

comment:6 by Jorropo, 7 months ago

Cc: Jorropo added
Resolution: invalid
Status: closedreopened
Summary: webp muxer seeks backwardswebp muxer silently seeks backwards when converting bigger png inputs and produce borked results when trying to write the length in the header

Hey, I have a reproduction that fails on all versions of ffmpeg I could get my hand on (4.1.11, 6.0, 2023-10-5 git-master 9078dc0c5202d54160f9a2d4c86c6f1ce756a224).

> curl -sssL https://w3s.link/ipfs/QmXgtZqEkiVuZgXSHJT3Cxmg2FDDqLwhRDRwrF3DipvKmN/t1.png | ffmpeg -i - -vcodec libwebp -preset default -f webp - | cat > t1.webp # gives broken file
ffmpeg version n6.0 Copyright (c) 2000-2023 the FFmpeg developers
  built with gcc 13.2.1 (GCC) 20230801
  configuration: --prefix=/usr --disable-debug --disable-static --disable-stripping --enable-amf --enable-avisynth --enable-cuda-llvm --enable-lto --enable-fontconfig --enable-gmp --enable-gnutls --enable-gpl --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libdav1d --enable-libdrm --enable-libfreetype --enable-libfribidi --enable-libgsm --enable-libiec61883 --enable-libjack --enable-libjxl --enable-libmodplug --enable-libmp3lame --enable-libopencore_amrnb --enable-libopencore_amrwb --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librav1e --enable-librsvg --enable-libsoxr --enable-libspeex --enable-libsrt --enable-libssh --enable-libsvtav1 --enable-libtheora --enable-libv4l2 --enable-libvidstab --enable-libvmaf --enable-libvorbis --enable-libvpl --enable-libvpx --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxcb --enable-libxml2 --enable-libxvid --enable-libzimg --enable-nvdec --enable-nvenc --enable-opencl --enable-opengl --enable-shared --enable-version3 --enable-vulkan
  libavutil      58.  2.100 / 58.  2.100
  libavcodec     60.  3.100 / 60.  3.100
  libavformat    60.  3.100 / 60.  3.100
  libavdevice    60.  1.100 / 60.  1.100
  libavfilter     9.  3.100 /  9.  3.100
  libswscale      7.  1.100 /  7.  1.100
  libswresample   4. 10.100 /  4. 10.100
  libpostproc    57.  1.100 / 57.  1.100
Input #0, png_pipe, from 'fd:':
  Duration: N/A, bitrate: N/A
  Stream #0:0: Video: png, rgba(pc), 2860x1576 [SAR 5669:5669 DAR 715:394], 25 fps, 25 tbr, 25 tbn
Stream mapping:
  Stream #0:0 -> #0:0 (png (native) -> webp (libwebp))
Output #0, webp, to 'pipe:':
  Metadata:
    encoder         : Lavf60.3.100
  Stream #0:0: Video: webp, bgra(pc, gbr/unknown/unknown, progressive), 2860x1576 [SAR 1:1 DAR 715:394], q=2-31, 200 kb/s, 25 fps, 1k tbn
    Metadata:
      encoder         : Lavc60.3.100 libwebp
[libwebp @ 0x55f4e77f54c0] Using libwebp for RGB-to-YUV conversion. You may want to consider passing in YUV instead for lossy encoding.
frame=    1 fps=0.0 q=-0.0 Lsize=     163kB time=00:00:00.00 bitrate=N/A speed=   0x    
video:163kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.002401%
> curl -sssL https://w3s.link/ipfs/QmXgtZqEkiVuZgXSHJT3Cxmg2FDDqLwhRDRwrF3DipvKmN/t2.png | ffmpeg -i - -vcodec libwebp -preset default -f webp - | cat > t2.webp # gives broken file
ffmpeg version n6.0 Copyright (c) 2000-2023 the FFmpeg developers
  built with gcc 13.2.1 (GCC) 20230801
  configuration: --prefix=/usr --disable-debug --disable-static --disable-stripping --enable-amf --enable-avisynth --enable-cuda-llvm --enable-lto --enable-fontconfig --enable-gmp --enable-gnutls --enable-gpl --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libdav1d --enable-libdrm --enable-libfreetype --enable-libfribidi --enable-libgsm --enable-libiec61883 --enable-libjack --enable-libjxl --enable-libmodplug --enable-libmp3lame --enable-libopencore_amrnb --enable-libopencore_amrwb --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librav1e --enable-librsvg --enable-libsoxr --enable-libspeex --enable-libsrt --enable-libssh --enable-libsvtav1 --enable-libtheora --enable-libv4l2 --enable-libvidstab --enable-libvmaf --enable-libvorbis --enable-libvpl --enable-libvpx --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxcb --enable-libxml2 --enable-libxvid --enable-libzimg --enable-nvdec --enable-nvenc --enable-opencl --enable-opengl --enable-shared --enable-version3 --enable-vulkan
  libavutil      58.  2.100 / 58.  2.100
  libavcodec     60.  3.100 / 60.  3.100
  libavformat    60.  3.100 / 60.  3.100
  libavdevice    60.  1.100 / 60.  1.100
  libavfilter     9.  3.100 /  9.  3.100
  libswscale      7.  1.100 /  7.  1.100
  libswresample   4. 10.100 /  4. 10.100
  libpostproc    57.  1.100 / 57.  1.100
Input #0, png_pipe, from 'fd:':
  Duration: N/A, bitrate: N/A
  Stream #0:0: Video: png, rgba(pc), 2530x1430 [SAR 5669:5669 DAR 23:13], 25 fps, 25 tbr, 25 tbn
Stream mapping:
  Stream #0:0 -> #0:0 (png (native) -> webp (libwebp))
Output #0, webp, to 'pipe:':
  Metadata:
    encoder         : Lavf60.3.100
  Stream #0:0: Video: webp, bgra(pc, gbr/unknown/unknown, progressive), 2530x1430 [SAR 1:1 DAR 23:13], q=2-31, 200 kb/s, 25 fps, 1k tbn
    Metadata:
      encoder         : Lavc60.3.100 libwebp
[libwebp @ 0x555b029314c0] Using libwebp for RGB-to-YUV conversion. You may want to consider passing in YUV instead for lossy encoding.
frame=    1 fps=0.0 q=-0.0 Lsize=     147kB time=00:00:00.00 bitrate=N/A speed=   0x    
video:147kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.002663%
> curl -sssL https://w3s.link/ipfs/QmXgtZqEkiVuZgXSHJT3Cxmg2FDDqLwhRDRwrF3DipvKmN/t3.png | ffmpeg -i - -vcodec libwebp -preset default -f webp - | cat > t3.webp # gives broken file
ffmpeg version n6.0 Copyright (c) 2000-2023 the FFmpeg developers
  built with gcc 13.2.1 (GCC) 20230801
  configuration: --prefix=/usr --disable-debug --disable-static --disable-stripping --enable-amf --enable-avisynth --enable-cuda-llvm --enable-lto --enable-fontconfig --enable-gmp --enable-gnutls --enable-gpl --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libdav1d --enable-libdrm --enable-libfreetype --enable-libfribidi --enable-libgsm --enable-libiec61883 --enable-libjack --enable-libjxl --enable-libmodplug --enable-libmp3lame --enable-libopencore_amrnb --enable-libopencore_amrwb --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librav1e --enable-librsvg --enable-libsoxr --enable-libspeex --enable-libsrt --enable-libssh --enable-libsvtav1 --enable-libtheora --enable-libv4l2 --enable-libvidstab --enable-libvmaf --enable-libvorbis --enable-libvpl --enable-libvpx --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxcb --enable-libxml2 --enable-libxvid --enable-libzimg --enable-nvdec --enable-nvenc --enable-opencl --enable-opengl --enable-shared --enable-version3 --enable-vulkan
  libavutil      58.  2.100 / 58.  2.100
  libavcodec     60.  3.100 / 60.  3.100
  libavformat    60.  3.100 / 60.  3.100
  libavdevice    60.  1.100 / 60.  1.100
  libavfilter     9.  3.100 /  9.  3.100
  libswscale      7.  1.100 /  7.  1.100
  libswresample   4. 10.100 /  4. 10.100
  libpostproc    57.  1.100 / 57.  1.100
Input #0, png_pipe, from 'fd:':
  Duration: N/A, bitrate: N/A
  Stream #0:0: Video: png, rgba(pc), 1518x802 [SAR 5669:5669 DAR 759:401], 25 fps, 25 tbr, 25 tbn
Stream mapping:
  Stream #0:0 -> #0:0 (png (native) -> webp (libwebp))
Output #0, webp, to 'pipe:':
  Metadata:
    encoder         : Lavf60.3.100
  Stream #0:0: Video: webp, bgra(pc, gbr/unknown/unknown, progressive), 1518x802 [SAR 1:1 DAR 759:401], q=2-31, 200 kb/s, 25 fps, 1k tbn
    Metadata:
      encoder         : Lavc60.3.100 libwebp
[libwebp @ 0x561307cf74c0] Using libwebp for RGB-to-YUV conversion. You may want to consider passing in YUV instead for lossy encoding.
frame=    1 fps=0.0 q=-0.0 Lsize=      69kB time=00:00:00.00 bitrate=N/A speed=   0x    
video:69kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.005682%

The resulting files really feels like they want to work, firefox flashes the correct image for a frame before switching to an error.

This one smaller than the others works:

> curl -sssL https://w3s.link/ipfs/QmXgtZqEkiVuZgXSHJT3Cxmg2FDDqLwhRDRwrF3DipvKmN/t4.png | ffmpeg -i - -vcodec libwebp -preset default -f webp - | cat > t4.webp # works
ffmpeg version n6.0 Copyright (c) 2000-2023 the FFmpeg developers
  built with gcc 13.2.1 (GCC) 20230801
  configuration: --prefix=/usr --disable-debug --disable-static --disable-stripping --enable-amf --enable-avisynth --enable-cuda-llvm --enable-lto --enable-fontconfig --enable-gmp --enable-gnutls --enable-gpl --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libdav1d --enable-libdrm --enable-libfreetype --enable-libfribidi --enable-libgsm --enable-libiec61883 --enable-libjack --enable-libjxl --enable-libmodplug --enable-libmp3lame --enable-libopencore_amrnb --enable-libopencore_amrwb --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librav1e --enable-librsvg --enable-libsoxr --enable-libspeex --enable-libsrt --enable-libssh --enable-libsvtav1 --enable-libtheora --enable-libv4l2 --enable-libvidstab --enable-libvmaf --enable-libvorbis --enable-libvpl --enable-libvpx --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxcb --enable-libxml2 --enable-libxvid --enable-libzimg --enable-nvdec --enable-nvenc --enable-opencl --enable-opengl --enable-shared --enable-version3 --enable-vulkan
  libavutil      58.  2.100 / 58.  2.100
  libavcodec     60.  3.100 / 60.  3.100
  libavformat    60.  3.100 / 60.  3.100
  libavdevice    60.  1.100 / 60.  1.100
  libavfilter     9.  3.100 /  9.  3.100
  libswscale      7.  1.100 /  7.  1.100
  libswresample   4. 10.100 /  4. 10.100
  libpostproc    57.  1.100 / 57.  1.100
Input #0, png_pipe, from 'fd:':
  Duration: N/A, bitrate: N/A
  Stream #0:0: Video: png, rgba(pc), 672x480 [SAR 5669:5669 DAR 7:5], 25 fps, 25 tbr, 25 tbn
Stream mapping:
  Stream #0:0 -> #0:0 (png (native) -> webp (libwebp))
Output #0, webp, to 'pipe:':
  Metadata:
    encoder         : Lavf60.3.100
  Stream #0:0: Video: webp, bgra(pc, gbr/unknown/unknown, progressive), 672x480 [SAR 1:1 DAR 7:5], q=2-31, 200 kb/s, 25 fps, 1k tbn
    Metadata:
      encoder         : Lavc60.3.100 libwebp
[libwebp @ 0x55cb0b0564c0] Using libwebp for RGB-to-YUV conversion. You may want to consider passing in YUV instead for lossy encoding.
frame=    1 fps=0.0 q=-0.0 Lsize=      27kB time=00:00:00.00 bitrate=N/A speed=   0x    
video:27kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.000000%

Writing to a file directly instead of through a pipe correctly works, diffing the old and bad files the surface bug seems evident:

1c1
< 00000000: 5249 4646 b48a 0200 5745 4250 5650 3820  RIFF....WEBPVP8 
---
> 00000000: 5249 4646 0000 0000 5745 4250 5650 3820  RIFF....WEBPVP8 
10412c10412
< 00028ab0: 0000 0000 0000 0000 0000 0000            ............
---
> 00028ab0: 0000 0000 0000 0000 0000 0000 b48a 0200  ................

We can see that for the bad file the length bytes of the header are blank, and the value that is supposed to be there is instead precariously appended to the end.
This is the first time I've looked at the webp spec so take this with a grain of salt, but it looks like when the input file is too big, some internal processing switches to a streaming mode, it zerofills the length header and then after writing the complete webp it then tries to write the length to the header using seek but errors aren't properly checked and the seek is silently dropped leaving the length written to the tail of the file.

Last edited 7 months ago by Jorropo (previous) (diff)

comment:7 by mkver, 5 months ago

Resolution: fixed
Status: reopenedclosed
Note: See TracTickets for help on using tickets.