dcpomatic/src/lib/audio_decoder.cc, branch v2.16.97 DCP-o-matic DCP tools https://git.carlh.net/cgit/dcpomatic/atom?h=v2.16.97 2023-08-15T16:27:09Z Cope with unexpected channel counts in data coming from audio decoders. 2023-08-15T16:27:09Z Carl Hetherington cth@carlh.net 2023-08-15T13:46:41Z urn:sha1:df7697962ef1049077643636918fba3cd3457e71 Improve flushing behaviour when there is a lot of space to fill (#2364). 2022-11-15T21:42:43Z Carl Hetherington cth@carlh.net 2022-11-14T23:44:37Z urn:sha1:d42bc731c125170efb1bb7b8c9f990a3e9fa5b57 Previously a call to flush() could result in a lot of audio being emitted from the decoder (if there is a big gap between the end of the audio and the video). This would end up being emitted in one chunk from the player, crashing the audio analyser with an OOM in some cases. Use the resampled rate to calculate the slack threshold rather than just 48kHz (#2275). 2022-06-29T22:53:20Z Carl Hetherington cth@carlh.net 2022-06-19T20:45:18Z urn:sha1:f49a724918ad3d1082384576960b1098d7f15822 Fix incorrectly-timed emission of silence padding causing buffer fill (#2217). 2022-03-15T08:20:17Z Carl Hetherington cth@carlh.net 2022-03-14T19:29:44Z urn:sha1:0c846af54055b9915c6c68617cd28176d5f84351 On initialisation or after seek we insert silence corresponding to a positive delay in an audio stream. Previously this inserted silence was done at time 0, so that after a seek to time T the silent frames would come out of the audio merger at time 0 and then the player would fill the space up to time T with silence. If T was far enough along this would fill the audio buffers without there being any video. Assorted tidying. 2021-09-19T23:44:30Z Carl Hetherington cth@carlh.net 2021-09-19T23:44:30Z urn:sha1:2e2f11b29651cffe37c64275dbd45c7563310020 Assorted C++11/formatting cleanups. 2021-04-08T19:32:44Z Carl Hetherington cth@carlh.net 2021-04-08T19:32:44Z urn:sha1:3339d3bce70afe9ae2ca10e9fcfc4b54b748fbf4 More c++ tidying. 2021-01-09T08:27:29Z Carl Hetherington cth@carlh.net 2021-01-09T08:27:29Z urn:sha1:0241df1707c7ea5658f471828ff6dc944e21af42 BOOST_FOREACH. 2021-01-07T22:27:56Z Carl Hetherington cth@carlh.net 2021-01-04T20:32:13Z urn:sha1:a5d004b0773f633401528392fc28e66d70e13ac8 std::shared_ptr 2021-01-07T21:48:29Z Carl Hetherington cth@carlh.net 2021-01-04T20:16:53Z urn:sha1:dd9be86db6cde0afa5da0d1d1ac43b42e05dca26 Obey audio timestamps if they don't deviate by more than some threshold. 2020-12-02T23:23:38Z Carl Hetherington cth@carlh.net 2020-11-28T23:51:38Z urn:sha1:8798bfab1539c3ac6031bb325e1b181c58b13fc5 Previously we would ignore audio timestamps because they are not contiguous in a sample-accurate way. However with bugs like #1833 we do need to obey large discontinuities in audio timestamps, otherwise we get large sync errors. Here we change timestamp handling to ignore small discontinuities in timestamps but not larger ones.