We have captured RTP streams between 2 endpoints, saving RTP header and payload.
For encrypted data, we have managed to decrypt and this has been tested successfully with Siren payload types - enabling a recording of a VOIP conversation to be obtained.
Our interest in IPP is specifically to use the IPP_RTA rtaudio library.
We have produced a tool that saves RTP data with RTAudio payload to rtp_dump format after decryption.
When this is provided as input to umc_speech_rtp_codec, the application crashes in ippsp8-6.0.dll with an Unhandled Exception at 0x01cd1244 in umc_speech_rtp_codec.exe: 0xC0000005: Access violation writing location 0x00940000.
Should the sample application work with RTAUdio, and/or, is there anything we might be doing wrong. It typically works through around 20 packets.
Can you give any hints as to how RTAudio bitstreams are packetized in RTP packets, such that we can call the codec directly with an input buffer? Does it use anything similar to RFC3267 "RTP Payload Format for AMR and AMR-WB". Typically, we observe payloads from 92 bytes up to 105 bytes.How is the bitrate determined?