Hi all, We're running into problems with leading and trailing silence due to pre-roll and remainder with aac and mp3 encodes. To illustrate here's a demo showing the decoded output of a couple mp3 and aac files:
http://rtoy.github.com/webaudio/codec-tests/plot-audio.html (requires Chrome or very recent Safari). Compare the aac and mp3 results with ogg. We're wondering what the best approach is here. Should ffmpeg/libav be cutting the codec specific preroll information before sending it on? Or should we simply expose this information through say AVCodecContext->metadata for downstream consumers to do what they wish with? See the following for the AAC specific information: http://developer.apple.com/library/mac/#documentation/QuickTime/qtff/QTFFAppenG/QTFFAppenG.html The reason this is necessary is for sample accurate playback of clips which is important for applications like WebAudio. Has anyone looked into this before? Thanks in advance! - dale
_______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
