ffmpeg-video

Chicken Scheme egg with bindings for ffmpeg designed for computer vision applications. In essence videos are treated as containers of images.

Note that this egg binds only to a small subset of ffmpeg that deals with video. It is used for computer vision where video files have only a single stream and audio is irrelevant. In addition this egg is designed to have frame-level accuracy, as such it doesn't support seeking and computing the video length is linear in the size of the video and very slow. Because of this it stores the computed video length of each video along with each video in a file named '<video></video>.video-length'.

Note that in keeping with ffmpeg's screwed up convention the first frame is 1, not 0.

Currently the main limitation of this egg is that it cannot encode video. One day it will be able to do so.

API

(with-ffmpeg-video video-path f)

Call f with an open ffmpeg at video-path.

(ffmpeg-open-video filename)

(ffmpeg-close-video video)

Open and close a video.

(ffmpeg-video-finished? video)

(ffmpeg-video-frame-index video)

(ffmpeg-video-width video)

(ffmpeg-video-height video)

(ffmpeg-video-frame-rate video)

Get information about an open video of an open video.

(ffmpeg-next-frame! video)

Advance the frame of a video.

(ffmpeg-video-frame-data video)

(ffmpeg-video-frame-data-as-imlib video)

Returns an image for the current frame as a PPM or an imlib image. See the scheme2c-compatibility or imlib eggs respectively.

(video-length video-pathname)

Compute the length of a video in frames. Note that this is a very slow operation as it must inspect the entire video. Frame-level accuracy cannot be achieved by seeking. See the note at the beginning of this document about caching.

(for-each-frame f v)

(for-each-frame-but-last f v)

(for-each-frame-reversed f v)

(for-each-frame-indexed f v)

(for-each-frame-indexed-but-last f v)

(map-frame f v)

(map-frame-but-last f v)

(map-frame-indexed f v)

(map-frame-indexed-but-last f v)

Walk over the frames of a video. Note that this does not fetch any data, even though decoding must happen for frame-level accuracy, it simply calls the function with one argument, the frame index. The indexed versions of these functions also pass an index (guaranteed to be 0-based) into the current movie.

(for-each-frame-pair individual-f pair-f video-name)

(map-frame-pair individual-f pair-f video-name)

As above this only provides frame indices not data. individual-f is called once for each frame and pair-f is called on every pair of frames.

(for-each-imlib-frame-from-video-indexed f video)

(for-each-imlib-frame-from-video-indexed-but-last f video)

(for-each-imlib-frame-pair-from-video individual-f pair-f video)

(for-each-imlib-frame-pair-from-video-indexed individual-f pair-f video)

(map-imlib-frame-from-video-indexed f video)

(map-imlib-frame-pair-from-video individual-f pair-f video)

Iterate over the video as above but the functions are called with both a frame number and an imlib image of the current frame. The indexed versions of these functions also pass an index (as the second arugment) (guaranteed to be 0-based) into the current movie.

License

Copyright 2012, 2013 Andrei Barbu. All rights reserved. This program 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 3 of the License, or (at your option) any later version. This program 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 this program. If not, see <http://www.gnu.org/licenses></http://www.gnu.org/licenses>.