[FFmpeg-devel] evaluating the experimental status of ffv1 version 3

[David Rice]

Hi all,

	I have been working with ffv1 version 3 for several months. It is much faster than version 1 and incorporates many features that make it well suited for digital audiovisual preservation and long term archival storage in additional to its other predominant use cases (such as an intermediate format); however, ffv1 version 3 is currently labelled as experimental and requires -strict experiment for use.

	Michael, Peter B., and I have shared speed tests and losslessness tests on ffmpeg-devel (http://comments.gmane.org/gmane.comp.video.ffmpeg.devel/143686), Peter B. has started work on an ffv1 reference paper (see: http://permalink.gmane.org/gmane.comp.video.ffmpeg.devel/149292), and Michael has undertaken the ffv1 technical documentation (http://ffmpeg.org/~michael/ffv1.html).

	I am very interested in what may be involved in moving ffv1 version 3 from its current experimental status and to label the ffv1 version 3 bitstream as fixed. I know that a key necessity in this effort is for the version 3 commits to be reviewed by someone(s) besides the original author. I am sadly under-qualified for this task but would be eager to facilitate or perform further testing or description in order to contribute to efforts to evaluate the experimental status and making progress on the reasons that ffv1 version 3 remains in this status.

	So far, in my testing and experience with ffv1, it has been a boon for the particular types of video work that folks like me and Peter B. has been involved with. Last month there was a presentation on the progress of ffv1 version 3 at the Society of American Archivists and the Library of Congress has added documentation on ffv1 to their digitalpreservation.gov website (http://www.digitalpreservation.gov/formats/fdd/fdd000341.shtml). Peter B. helped release a set of archival tools based on ffv1: http://www.dva-profession.mediathek.at/.

	How close are we to an official ffv1 version 3?

Dave Rice