Hi there,
As the last project (co-)maintainer, I'd like to announce a few
decisions I made in the last months of hiatus regarding the project.
Quick summary for the impatient
-------------------------------
* transcode 1.1.6 will be released. It will be the last release from
mine. berlios.de is closing. Moving on bitbucket.
* the development of transcode will NOT restart as it was.
* there is a plan to not waste what was good in transcode.
Let me give some details in the remainder of the email.
Transcode 1.1.6 FINAL
---------------------
Transcode 1.1.6 it is brewing from LONG time. It will be the last bugfix
feature.
This round will be the last. I'm collecting all the new patches on
transcode-devel (vid.stab optimization, FFMpeg API update et. al.) and
I'll apply them.
I'll also made a last full scan on the major bug-trackers in order to
collect as much bugfixes as possible. Anyone willing to collaborate can
send patches here on transcode-devel until october the 15 (GMT+1, Rome
Timezone), included. Please tag them with something like
"[PATCH]"
After October the 16, I'll freeze and tag transcode 1.1.6...
... on bitbucket.org, because berlios.de is shutting down (see the
announcement on the homepage).
The whole mercurial repository is already cloned and publicly avalaible:
https://bitbucket.org/france/xtranscode
It was originally planned as experimental branch, but I then choose
another route, see below for more. Bitbucket is kind enough to provide
infinite space for repositories, so I'll likely add another clone of the
tree, transcode-legacy or something like that.
the exit1.org repo, if still active (as I remember/hope :)) will be
updated too, of course.
The official announcement of transcode 1.1.6 will be presented here on
transcode-devel of course pointing to the final repo. It will either be
https://bitbucket.org/france/xtranscode
or (something like)
https://bitbucket.org/france/transcode-legacy (not yet created!)
I got mixed feelings about extracting the project data from berlios.de,
most notably the old release tarballs and the bugtracker data.
I think the only really worth old tarball is the last from the 1.0.x
series, but a tag on mercurial should replace it quite efficiently.
The bugtracker is another story, and it'll probably deserve to be
extracted and put somewhere else, maybe imported on the bitbucket one.
Farewall to transcode (so long, and thanks for all the fish)?
-------------------------------------------------------------
Yes, and no.
Yes, because the situation hasn't changed, and I cannot restart the
development as it was.
Of course I hereby renew my offer to fully support anyone wanting to
step out to continue the project.
No, because of after a lot of thinking I found some ways to save at
least partially what was good in transcode.
tycho-plugins: the transcode filters for everyone
--------------------------------------------------
There is an effort in progress to port all the filters of transcode
which are generic enough to allow that to the frei0r interface, in order
to make them avalaible to a broad range of applications and no longer
confine them into transcode itself.
Doing like that, an important and appreciated part of transcode will
still be developed and be useful for someone out there.
You can follow the progress of this sub-project here:
https://github.com/mojaves/tycho-plugins
* Why frei0r?
In order to maximize the potential usage of the filters. A bunch of
video-related software already support frei0r, most notably
MLT (http://www.mltframework.org/twiki/bin/view/MLT/)
Kdenlive (http://www.kdenlive.org/)
GStreamer (http://gstreamer.org)
FFmpeg | libAV (http://ffmpeg.org/ | http://libav.org/)
Moreover, the filter API of transcode match quite nicely with the one of
frei0r, so the porting is really quite cheap.
* Why not to aim to GStreamer/FFmpeg directly?
Because I don't have the time/resources to match the requirements of
those two projects :)
* What will be ported? Which filter will be excluded?
Only the filter which are very specific to the transcode architecture
are excluded from the port. This roughly excludes from the port
just the frame-rate changing plugins. Unfortunately, this also includes
the telecine-related filters.
* Why git?
To maximize the upstream merging probability. frei0r filters are
developed using git.
mercurial has a git interface. A mercurial mirror will be made avalaible
(on bitbucket).
To be honest, I hardly understand all the git hype and I feel much more
confortable with mercurial. My fault, probably.
* Why github?
I liked it. :)
That's all!
bests,
--
Francesco Romani // Ikitt
