Thanks for your reply. On 23 Jun 2009, at 18:06, Romain Beauxis wrote: > Currently, the MP3 decoder is tried last because for detection > purpose, mp3 > format is very bad: it had no header, hence it can appear that > random files, > such as JPEG images are taken as mp3.. Hence, we try to test all > decoders > before it, since it might accept files that were not for it. >
I understand. >> Moreover, a little suggestion, could you first try the encoder linked >> to the extension of the file as in the most of the case it is the >> right one (.ogg => vorbis, .mp3 => mp3, etc ...) ? > > We have always wanted to avoid using file extension for detecting > the audio > format. This is rather weak. Instead we try to decode audio data > from the file. > In the case of mp3, there is also a mime check done when ocaml-magic > support > has been compiled. I was not saying that it should use the file extension to choose the decoder but it could use it to optimize the tests list (and put the try of mp3 decoder first if the mp3 extension is there). > All in all, the decoder checks for formats other than mp3 should not > be very > expensive in terms of load.. The only case is decoders using ocaml- > natty, due > to some internal limitation in ocaml-bitmatch, and that is why these > decoders > are disabled by default.. I know it's not so expensive in a single liquidsoap process. But we are running 80 liquidsoap per server :-) Best regards, Jef ------------------------------------------------------------------------------ Are you an open source citizen? Join us for the Open Source Bridge conference! Portland, OR, June 17-19. Two days of sessions, one day of unconference: $250. Need another reason to go? 24-hour hacker lounge. Register today! http://ad.doubleclick.net/clk;215844324;13503038;v?http://opensourcebridge.org _______________________________________________ Savonet-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/savonet-users
