dcpomatic/src/lib/player.cc, branch v2.16.44 DCP-o-matic DCP tools https://git.carlh.net/cgit/dcpomatic/atom?h=v2.16.44 2023-02-10T09:31:15Z Ignore overlapping ATMOS data in the same way that we ignore video (#2429). 2023-02-10T09:31:15Z Carl Hetherington cth@carlh.net 2023-02-09T23:38:33Z urn:sha1:3a9265b3f824bde609ec3f16a853b7c69c11877d Fix multiple video overlaps. 2023-02-10T09:31:15Z Carl Hetherington cth@carlh.net 2023-02-09T23:25:06Z urn:sha1:46a1e43f9e3f03565536419afc8137c64d5681c0 Cleanup: add Content::period(). 2023-02-10T09:31:15Z Carl Hetherington cth@carlh.net 2023-02-09T21:49:47Z urn:sha1:fbdfc901498605f89f28a57d595df966bddc4eb1 Cleanup: use some better variable names. 2023-02-10T09:31:15Z Carl Hetherington cth@carlh.net 2023-02-09T21:42:11Z urn:sha1:ead33f1db9657b3af93ec966d7bbc24218b6fba0 Cleanup: remove unnecessary check. 2023-02-10T09:31:15Z Carl Hetherington cth@carlh.net 2023-02-09T21:07:39Z urn:sha1:11bc225507bbd6f2942b7b867389d9571a09c717 Handle 2D/3D mismatches in the player (#2409). 2023-02-01T00:12:38Z Carl Hetherington cth@carlh.net 2023-01-11T22:00:25Z urn:sha1:2f12058c535045cecc226fe47b3d60da8851a862 Previously if there was 2D content in a 3D project (or vice versa) the player would emit the video anyway and the encoder would have to deal with it. Since previously the FFmpeg encoder did not deal with this quite right, it seems to make sense to move the logic into the player so that neither encoder has to worry about it. Fix build on Ubuntu 16.04. 2022-12-15T15:39:53Z Carl Hetherington cth@carlh.net 2022-12-15T15:39:53Z urn:sha1:192ba9fe562ccda0b888bf0d3c8f71d4c67be309 Move some methods out of util.{cc,h} 2022-12-10T00:16:06Z Carl Hetherington cth@carlh.net 2022-12-10T00:08:56Z urn:sha1:02b74112721d13a27b0bbaece714d5c8ea743d43 _stream_states must be reset on seek otherwise #2368. 2022-11-25T22:42:55Z Carl Hetherington cth@carlh.net 2022-11-20T21:03:38Z urn:sha1:9e49b5f78d1150f75a6016b4f01e2645a7c023b0 Fix sound corruption with multiple sound files (#2368). 2022-11-25T22:42:55Z Carl Hetherington cth@carlh.net 2022-11-19T16:16:38Z urn:sha1:7d0cd227fa8cf13b04504af22301521dc231c0c7 last_push_end would be initialised to the position of the content, meaning that content at (for example) 60s into the project would be treated as the leader, and content much earlier would then be considered finished by the code which checks ignore_streams_behind (as its last push was more then 5s before the leader). This caused audio corruption, shown up by the test here.