The glorious history of FLOSS video editors

In case you have been living under a rock and have not seen my talk at LGM this year, you have probably missed this important piece of information:

This image is now published in the contributing section of PiTiVi’s website (as if you did not have enough arguments already! What are you waiting for? Come here and help us rock the world!).

Oh, and gnuclad is awesome.

24 thoughts on “The glorious history of FLOSS video editors

  1. Just a small clarification, Novacut is not a fork of dmedia. dmedia is the base of Novacut. So you basically need to replace dmedia with Novacut and remove the Novacut fork

    Just for information

  2. @Daniel: yes, see http://sourceforge.net/projects/lightscameramon/

    @Enrico: yes. This is based on research I did on sourceforge, ohloh, etc. You can go take a look for yourself. I based myself upon version control history to determine the lifespans of projects. If you know of other video editors for Linux that I have not found, let me know.

    @Manish: yeah, I know the relation between Novacut and D-Media, but at this time, Novacut still does not exist (and I can’t represent D-Media as “Novacut”) yet I “have” to represent it somewhere, somehow (thus the dot, just like Lightworks that has still not been released as open-source for Linux to this day).

    Currently there just is no other way to represent Novacut’s situation easily with a cladogram. If I only put the “unreleased dot” of Novacut on the diagram, people would accuse me of not having included D-Media in it (even though this library is *not* a video editor per se and should not be included in the diagram, just like gnonlin or ges or MLT are not included in this diagram).

    P.s.: this cladogram concentrates on apps available for Linux. If you start including all those failed projects to make open source video editors on Windows, the graph becomes a huge mess. Also, some apps in there do not fit exactly the “typical” definition of a video editing program (such as Blender, Ingex or Lives).

  3. Oh good. You being in this field for long will surely have an eye on what all editors are there. Yeah Novacut is not there right now, it will come slowly. BTW best of luck of the PiTiVi hackfest.

  4. I believe OpenShot is still an active project, unlike what the graph implies!

  5. @Vadim: we’ll see, for now there has been no significant commit activity in it since the end of March.

  6. @Brian: I knew that, but I’m not seeing any public code. I can only track activity that is done in the open, so when/if they release it I shall update the graph.

    @Luc: it’s color-coded by “engine” or core technology (ex: GStreamer+GNonlin, ffmpeg+MLT, VLC, mplayer, etc.). The first version of this cladogram, instead of using color codes for engines, made the various editors “fork” from engines, but it was reportedly more confusing. Of course, even color-coding by engine is not entirely accurate as it does not represent editors “switching engines” along the way…

  7. @drago01: Avidemux (just like vanilla VLC, mencoder, ffmpeg or gst-launch) didn’t fit my “typical definition” of a video editor. Avidemux is more like a transcoder/encoding tool than a generic video editor. Yes, you can apply filters and trim/cut out some portions of the video and whatnot, but you can’t (at least, not that I know of) stitch clips together, do compositing and transitions, etc. I hope this clears things up, correct me if I’m wrong.

  8. Well it is more than an encoding tool, you cannot add transitions but you can stich clips together (even though the interface is akward).

    I use it to quickly cut parts of a video and save it again which seems to work pretty well.

    I just wondered why it isn’t listed here.

  9. Avidemux is not a NLE (non-linear editor). That means you can not combine more tracks (subtitle track, chroma keying, compositing, transitions etc.) Avidemux is “just” a simple video joiner and cutter, just like VirtualDub. A really interesting opensource NLE is Avisynth, but it’s Linux port was never finished.

  10. Not sure where you’re looking, but the last code commit for Openshot 1.3.1 was on 3rd May, according to Launchpad. If you want to go by activity, then you should add some major gaps to that Pitivi time line :-)

  11. @af: I did look at the last 3 commits in openshot, and they were either just translation updates or changes to the version number file or updates to the copyright notice/contributors.

  12. Yes, final changes prior to a release – honestly, if you think a project is dead because the last release was in May (and it’s now July) then you have to wonder how much of the rest of this chart of yours can be relied upon.

  13. As I said, I will update the chart again once new commits appear. The last 2-3 commits were really really punctual and spaced and, in my humble opinion, did not reflect properly the level of activity; since early March, there has been no *significant* amount of commit activity. If this can help you visualize my argument, take a look at https://ohloh.net/p/openshot-video-editor and change the graph mode to “commits” and look at the huge drop.

    Call this unfair if you want, but I’m waiting for the new big code drop from Openshot to update my graph, and everything will be peachy again.

  14. I’m interested in code development activity. A release is often just a snapshot in time, a tag on a commit to indicate a particular epoch. You may notice that my cladogram only calculates commits/visible active development, it does not take releases into account. If that graph was supposed to represent releases, at *least* half of the projects displayed in the graph would go away completely, and it would also do injustice to the projects that are in “development hell” (hello Lumiera).

    Look at http://bazaar.launchpad.net/~jonoomph/openshot/main/revision/485 for example. That’s the last “release” commit you are referring to. Look at the diffs for that commit. I’m sorry, to me, that doesn’t count. I’m not trying to say my graph is perfect, but at the time being I’m trying to reflect publicly visible activity and changing a version number and committing translation updates is not a good indicator IMHO.

    Again, I’ll be glad to update the graph when a new bunch of code development commits suddenly appear. I should also mentiont that I still haven’t added the editor “Bauxite” from 2004 yet (I need to find the time to do this).

    Please do not see malice in this. It’s not exactly easy to try to synthetize the history of all those projects with a one-size-fits-all metric.

  15. Hello,

    Could you perhaps provide the script which you used to create this neat graph?

    It is amazingly helpful at figuring out which video editors one probably should not spend time on trying to learn. It’s a bit hard to figure this out without your help, as projects usually don’t tell “we’re on ice for now” and one would have to check out the main repo to be totally sure about presence or absence of development activity…

  16. Thanks for the heads up Andy, hadn’t noticed the activity that restarted during the last month. Should be up to date now!

Comments are closed.