I’ve finished working on supporting audio calls in Telegram. The latest version is available already in HaikuDepot. I tested the performance with my sound cards: HD Audio (Azalia codec) and Creative SB Live 5.1 (emu10k) - everything works fine. I need your help to test this functionality on different sound cards because different cards have different bit rate, sampling and audio capture format.
Does this mean audio output/input in QtMultimedia has been implemented? Or does Telegram use something else for its audio handling core?
Telegram uses the libopenal library to play audio (audio and video files). Which we have openal, but small problems with channel mapping when playing sound: for example I hear only the left channel in the front speakers. But nevertheless the sound when watching the video is available.
For calls used libtgvoip library which has its own modules to work with sound.
I implemented a playback module through BSoundPlayer and a sound recording module through BMediaRecorder. Where you had to solve a lot of problems with mixing multichannel sound to a mono stream, and perform a sample rate conversion to 48000.
Tested (Telegram 1.3.5, Haiku x86_64):
- Creative Sound Blaster Audigy (Model #: SB0090)
- Creative Sound Blaster Audigy 2 (Model #: SB0240)
- HT Omega Claro Halo II
- VIA VT1708S 8-Channel High Definition Audio CODEC
Note: Using them gave good results.
Needs work (OSS 4.2-related):
- Creative Sound Blaster X-fi Elite Pro (Model#: SB0550)
Glad to hear it. Thnx!