libdcp/src/sound_asset.cc, branch python310 DCP handling library https://git.carlh.net/cgit/libdcp/atom?h=python310 2021-04-29T07:36:46Z Slightly more details message when sound data differs in ::equals(). 2021-04-29T07:36:46Z Carl Hetherington cth@carlh.net 2021-04-29T07:36:46Z urn:sha1:6836df3cafad335bfe9aaf0fa2a03d3b49638281 Fix/hide some warnings. 2021-04-13T23:00:05Z Carl Hetherington cth@carlh.net 2021-04-13T23:00:05Z urn:sha1:274dd30f2b11fe8ea563a2ac7436c9d463865b0b Make sound asset language optional. 2021-04-08T22:02:09Z Carl Hetherington cth@carlh.net 2021-04-08T22:02:09Z urn:sha1:45525875fef67ec1885968c0c5692307b16a279b Write MCA tags based on the specified sound field. 2021-03-25T23:09:15Z Carl Hetherington cth@carlh.net 2021-03-25T23:09:15Z urn:sha1:b9c5e3f74550917676e14324791d6ba4d0c07cac I had previously assumed that one should write MCA information into the sound MXF based on what channels actually contain sound rather than silence. However a previous example of a stereo DCP gives a verification error in EasyDCP (see DoM bug #1935) which is solved by adding MCA tags for each channel in the specified sound field (e.g. at least 6 tags for a file which is marked as 5.1) even if the audio is really stereo. This commit removes the "active channels" stuff and makes sure that if we say a file is 5.1 we write at least 6 MCA tags (and similarly for 7.1). Tidying. 2021-01-24T03:15:26Z Carl Hetherington cth@carlh.net 2021-01-24T03:15:26Z urn:sha1:ceaf7bc52712cb60708ed5eb5c62c5e463dd8e89 Use enum class for the things in types.h 2021-01-21T00:57:06Z Carl Hetherington cth@carlh.net 2021-01-21T00:57:06Z urn:sha1:76e3325a16cdf6d7220a61e2b5cfdb9c804cc32c Store sound asset language as a string, but set it as a 2021-01-17T19:13:22Z Carl Hetherington cth@carlh.net 2020-12-09T12:12:21Z urn:sha1:0b39df958f534068950669a763be7b4602740107 dcp::LanguageTag. This makes it work the same as other language tags, and tolerates bad values while trying not to allow them to be created. std::shared_ptr 2021-01-07T23:35:29Z Carl Hetherington cth@carlh.net 2020-12-20T13:14:07Z urn:sha1:d39880eef211a296fa8ef4712cdef5945d08527c Fix terrible SoundAsset::equals() implementation. 2020-11-29T19:57:57Z Carl Hetherington cth@carlh.net 2020-11-29T19:57:57Z urn:sha1:e116649cb93e7784601c1c55b2f3c8cf32099abf It would check individual bytes of samples to see if they differed by more than the threshold. Not only is this almost useless, but the default threshold is 256 so with the default settings it would always say that two assets of the same length (and channels, etc.) were the same, even if the sample data was different. Fix crashes when static linking due to tricky lifetime 2020-09-21T22:06:21Z Carl Hetherington cth@carlh.net 2020-09-21T20:33:43Z urn:sha1:e40f5c5923d72b42d7b1c617a6e459569d584e8c management of ASDCP::DefaultSMPTEDict().