lav2yuv converts an MJPEG video sequence described by a sequence
of MJPEG video files and/or edit lists pointing to such files into the
simple uncompressed planar 4:2:0 Y'CbCr format, as used by mpeg2enc(1)
MPEG encoder and image processing filters like
yuvscaler(1) or yuv2dfilter(1) or yuvdenoise(1).
Output is to stdout so that by piping the output of lav2yuv into a
suitable pipeline it is possible to process and then encode or play
back video recorded in any of the mjpegtools MJPEG container formats:
AVI, quicktime or edit lists describing editted versions of
video held in such files.
Mixing different files with different video formats is currently not possible.
The -S -T -D options are used for scene detection,
which is used by linux video studio.
lav2yuv accepts the following options:
Output a scene list with scene detection
Set scene detection threshold to num (default: 4)
Width decimation to use for scene detection (default: 2)
Frame offset skip num frames in the beginning if num is negative,
all but the last num frames are skipped
Only num frames are written to stdout. 0 means that all frames are processed.
Set/override the sample aspect ratio (SAR) emitted in the output stream
header. (Currently, the SAR is only auto-detected for DV files, and
Set the intended/approximate display aspect ratio (DAR) for the stream.
This value is only used to guess the SAR when it cannot be detected
from the input stream. The default DAR is '4:3' (corresponding to a
standard TV screen); another useful value is '16:9' (corresponding to
This man page was originally written by Bernhard Praschinger.
If you have questions, remarks, problems or you just want to contact
the developers, the main mailing list for the MJPEG-tools is: