Ilya Basin
2013-Nov-29 18:49 UTC
[opus] Please help me decode this webrtc chrome conversation
Hi. I made a webrtc relay with recording and dumped the SDP requests and RTP packets into files. Then I made a java decoder based on jitsi. Although the files contain all the needed info: encription keys, codec info, timestamps, etc., I could only decode one side in one of 2 conversations. For example, the RTP payload is decrypted successfully, but opus_packet_get_nb_samples() or opus_decode() return OPUS_INVALID_PACKET. payload-0.dat , contains my real voice. Decoded successfully. https://docs.google.com/file/d/0B40Hp1Tx6I6eaHJaRjBKY1M0WnM/edit payload-1.dat , contains looped beep recorded with a virtual microphone. While recording I was hearing it, but I can't decode the file. https://docs.google.com/file/d/0B40Hp1Tx6I6eR25ybGh2MnEzRk0/edit -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: test.c Url: http://lists.xiph.org/pipermail/opus/attachments/20131129/62bfcc74/attachment.txt
Jean-Marc Valin
2013-Nov-29 19:39 UTC
[opus] Please help me decode this webrtc chrome conversation
On 29/11/13 01:49 PM, Ilya Basin wrote:> For example, the RTP payload is decrypted successfully, but > opus_packet_get_nb_samples() or opus_decode() return > OPUS_INVALID_PACKET.Are you sure you didn't forget to strip the RTP header before decoding it? Jean-Marc
Ilya Basin
2013-Nov-29 20:31 UTC
[opus] Please help me decode this webrtc chrome conversation
Yes, I parse the RTP header, decrypt the payload and then feed the decrypted data to opus. Besides, opus_decoder_get_nb_samples() on 1st packet returns reasonable number 960, but then opus_decode() fails. On Fri, Nov 29, 2013 at 11:39 PM, Jean-Marc Valin <jmvalin at jmvalin.ca>wrote:> On 29/11/13 01:49 PM, Ilya Basin wrote: > > For example, the RTP payload is decrypted successfully, but > > opus_packet_get_nb_samples() or opus_decode() return > > OPUS_INVALID_PACKET. > > Are you sure you didn't forget to strip the RTP header before decoding it? > > Jean-Marc > >-------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.xiph.org/pipermail/opus/attachments/20131130/b5b2050c/attachment.htm