Author, manage, and run multitrack audio/video compositions.
The engine of a non-linear video editor that can be used in all
sorts of apps, not just desktop video editors.
MLT is an open source multimedia framework, designed and developed
for television broadcasting. It provides a toolkit for broadcasters, video editors,
media players, transcoders, web streamers and many more types of applications.
The functionality of the system is provided via an assortment of ready to use tools,
XML authoring components, and an extensible plug-in based API.
You can download it from SourceForge or GitHub. The major announcement is the introduction of support for libavfilter! This is still a work-in-progress. It is limited to FFmpeg 2.3 and up, and there are a number of filters that are black-listed because they are known to not integrate with MLT, which is not a full libav* environment or simple wrapper for it. There are likey avfilters that are not yet black-listed but might not work because they have not been completely tested. Also, they do not support MLT's keyframable property animation nor its frame-threaded parallelism due to architectural or integration limitations. However, some avfilters are slice-threaded (internal parallelism), and that works. Finally, libavfilter filtergraph syntax is not supported either.
All of the supported libavfilters are exposed as MLT filters beginning with the prefix "avfilter." All of the avfilter parameters are exposed as MLT properties with the "av." prefix to prevent clashes with MLT properties. You can run
melt -query filters to see the new avfilters, and
melt -query filter=avfilter.rotate, for example, to view generated documentation for an individual filter.
Here is a list of notable fixes and enhancements in this release:
Added support for libavfilter to avformat module.
Added auto-rotate support to avformat producer.
Added animated GIF preset for avformat consumer.
Prevent serializing and deserializing mlt_type property to xml module.
Fixed relative paths for WebVfx "plain:" resources in xml module.
Updated libebur128 to v1.1.0 in plus module.
Added dynamic_loudness filter to plus module.
Added loudness_meter filter to plus module.
Qt 5 fixes for kdenlivetitle producer.
Added gradients and text shadows to kdenlivetitle producer.
Added support for building rtaudio against external build of lib.
Upgraded bundled RtAudio to v4.1.2.
Added status parameters to ladspa producer and filters.
Added 5.1 surround to stereo downmix to audiochannels filter in core module.
This is a bugfix and minor enhancement release. Note that our release versioning scheme has changed. We were approaching 1.0 but decided to synchronize release version with the C library ABI version, which is currently at v6. You can download it from SourceForge or GitHub. Here are some of the notable changes and enhancements:
Added unit tests for tractor, multitrack, and field.
Added drop_count readable property to mlt_consumer.
Define MIN, MAX, CLAMP in mlt_types.h in not already defined.
Switched to __APPLE__ and _WIN32 defines throughout codebase.
Added UDP and SMPTE 2022-2 support to cbrts consumer.
Fixed build against latest FFmpeg versions - now requires v1.1 and up.
Added audiospectrum filter to qt module.
Added meta.media.0.codec.rotate property to avformat producer to let apps and other services get the media orientation.
Make the avformat producer handle animated images.
Added style property to dynamictext filter.
Added timewarp producer to core module.
Fixed slowly accumulating A/V sync drift in mix audio transition.
Added width_crop and width_fit properties to pango producer.
Added -abort option to simply exit without full cleanup.
Thanks to the sponsorship of BCE (Broadcast Centre Europe), the git version of the cbrts consumer now includes SMPTE 2022-2 RTP/UDP output. It has been successfully tested for 36 hours of continuous playback with an ateme DR8400 IRD and the DecTek StreamXpert performing PCR analysis and TR 101 290 monitoring.
In addition to the exciting new audio/visualization filters mentioned in the previous blog post, this version features some exciting API improvements for manipulating multitracks and property animations. There are still many services, however, that need updating to support animated properties. Also, since SourceForge is still recovering from a storage problem preventing the upload of new files, you can download the release source archive from GitHub.
Added mlt_service_disconnect_producer() and Mlt::Service::disconnect_producer().
Added mlt_multitrack_disconnect() and Mlt::Multitrack::disconnect().
Added mlt_tractor_remove_track() and Mlt::Tractor::remove_track().
Added mlt_service_insert_producer() and Mlt::Service::insert_producer().
Added mlt_multitrack_insert() and Mlt::Multitrack::insert().
Added mlt_tractor_insert_track() and Mlt::Tractor::insert_track().
Added mlt_transition_set_tracks() and Mlt::Transition::set_tracks().
Added Mlt:Animation class with methods:
Added mlt_animation_key_count() and Mlt::Animation::key_count().
Added mlt_animation_key_get() and Mlt::Animation::key_get().
Added audiowaveform video filter.
Added fft audio filter.
Added dance video filter (uses fft).
Added lighshow video filter (uses fft).
Added distort property to movit.rect video filter.
Added rotate property to pango video producer.
Added 2K DCI and 4K modes to decklink producer and consumer.
Added audiomap (channel remapping) filter.
Added property animation to all LADSPA audio filters and producers.
MLT developer, Brian, developed 4 new filters, available now in git master. The fft filter uses the fast Fourier transform to produce data for other filters. The dance filter moves, scales, and rotates an image or video in reaction to the audio characteristics. The lightshow filter alters the color of the image based on the audio. Lastly, the audiowaveform filter draws a waveform over the image.
These were all sponsored by the "Learn Your Lyrics" YouTube channel. They generate the videos using MLT through shell scripts that generate long melt command lines. You can see the fft, dance, and audiowaveform filters in action in this video: