Good day, I want to use Speex codec for audio translation and now port it to AVR32. But during my experiments with Speex I see one trouble: very often I will not can to start stream reception from 1st frame. Therefore I have ugly sound quality at the beginning. I have idea to send some more information for data recovery in translation packets (I have 110 free bytes in each frame). What data from SBDecState structure decoder needs for each next frame? Or it need all data in this structure (in this case I will can not pack it in one packet). -- Best regards, Fedor Bobkov
See the speexenc/speexdec code. You need some container that allows resync (like Ogg). Jean-Marc Quoting Fedor Bobkov <bobk-off at yandex.ru>:> Good day, > > I want to use Speex codec for audio translation and now port it to > AVR32. But during my experiments with Speex I see one trouble: very > often I will not can to start stream reception from 1st frame. > Therefore I have ugly sound quality at the beginning. I have idea to > send some more information for data recovery in translation packets > (I have 110 free bytes in each frame). > > What data from SBDecState structure decoder needs for each next > frame? Or it need all data in this structure (in this case I will > can not pack it in one packet). > > -- > Best regards, > Fedor Bobkov > > _______________________________________________ > Speex-dev mailing list > Speex-dev at xiph.org > http://lists.xiph.org/mailman/listinfo/speex-dev > >