James Mead <ffmpeg <at> floehopper.org> writes: > On 1 February 2016 at 13:39, Carl Eugen Hoyos <cehoyos wrote: > > James Mead <ffmpeg <at> floehopper.org> writes: > >> I want to make the MP3 available on a web page as the source of an > >> HTML5 audio element and for it to be playable in Google Chrome. > > > > There was no encoder change, I don't think we can fix this. > > OK. In that case, can you explain a bit more about > scenarios in which the bug is triggered?
It is triggered if the size of the attachment is slightly smaller than 1048576 bytes and if the beginning of the mp3 stream has similarities with aac (or another codec), the similarities can happen, if the attachment is small enough, more bytes can be checked (to rule aac etc. out). After the change, the case when the attachment is only slightly smaller than 1M is handled more similar to the case when the attachment is bigger (mp3 is assumed in this case iirc). All this is related to the issue that some users (like you) need / want precise detection, others want fast detection... (But for the given sample, the behaviour was definitely a bug.) Carl Eugen _______________________________________________ ffmpeg-user mailing list [email protected] http://ffmpeg.org/mailman/listinfo/ffmpeg-user
