One of the difficulties of having a long release cycle for a small project is that a lot of activity might be taking place behind the scenes, but a casual observer might not notice. Of course, in the case of Schrödinger, it didn’t help that diracvideo.org was lacking CSS and important links for over a month, looking rather bit rotten. So it’s not surprising that there have been people wandering into the IRC channel wondering if the project is dead. Um, no. We’re just quiet. And there’s a new release.
Partly the reason for the long release cycle is that it took more time than expected merging several of the encoding tools from dirac-research. But now there are fewer loose threads and development and releases can proceed at a more even pace. I’m hoping to do new releases at the pace of about one a month. (But I’ve said that before…)
Schrödinger now requires Orc to build. Switching from liboil to Orc has made decoding a lot faster, sometimes as much as 4 times faster.
Look forward to separate blog posts about some of the new features. Encoding quality has improved quite a bit for typical cases, and hugely in cases where there were bugs that were being ignored.
Awesome, I’ll have to try it out. Can you elaborate a bit about the new encoder changes, what kind of input do they give improvements for and why?
This isn’t the place for bug reports, but I couldn’t find where was, so:
A couple of recent tests show both Dirac and Schro having issues with colors, possibly adding a slight red/purple cast.
http://keyj.s2000.ws/?p=356
and
http://forum.doom9.org/showthread.php?t=135176&page=2
which my limited knowledge of codecs seems to suggest is a bug, rather than a compression artefact, though it’s odd that both codebases seem to have a similar issue.
[...] latest version of the Schrodinger encoder/decoder from diracvideo.org. If you saw David Schleefs blogpost about Dirac you would have seen him mentioning it is much [...]