For review, a fix for #5727. Please test and comment. I've made new versions of the processing functions that process stereo channels together so that the rule regarding target_gain can be implemented. The target_gain used for both channels will now be the last valid target_gain that matched between channels.
The old single-channel functions are still there, so that testing is possible, and for compatibility with the original fb2k code. They can be used by: -af hdcd=process_stereo=0 I've tested several HDCD flac and the majority come out exactly the same using either path, but there have been a few differences where control codes do not appear at exactly the same sample, so there is a few samples of delay before they match up. One strange case is John Mellencamp - [1994] Mr. Happy Go Lucky 06. Emotional Love, where there is one extra HDCD packet in one channel, and that target_gain value ends up being ignored. This patch is against master, so it also includes the changes in an earlier (not yet applied) patch set for HDCD detection, PE mode, cdt counter, and code comments. _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
