Add commands for user audio control: mute, unmute, and start with user audio muted. #23
Conversation
Signed-off-by: Dario Pellegrino <dario.pellegrino@voismart.it>
dariopellegrino00
force-pushed
the
feature/user-audio-flow-control
branch
from
df59cb4 to
7162b55
Compare
|
Require refactor of the stream api command function, too many nested ifs |
There was a problem hiding this comment.
Pull Request Overview
This PR adds mute functionality to the mod_openai_audio_stream module, allowing selective control of audio flow between the caller and OpenAI Realtime API. It introduces the ability to start with the user muted and provides runtime controls to mute/unmute either the user's audio going to OpenAI, OpenAI's audio playing to the user, or both.
Key changes:
- Added
user_audio_mutedandopenai_audio_mutedflags to control audio flow independently - Implemented
muteandunmuteAPI commands with support for targeting specific audio directions - Introduced
start_mutedparameter to allow initializing sessions with user audio muted - Changed default sampling rate from 8kHz to 24kHz to align with OpenAI Realtime API requirements
- Replaced C-style casts with
static_castandreinterpret_castthroughout for better type safety
Reviewed Changes
Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.
Show a summary per file
| File | Description |
|---|---|
| openai_audio_streamer_glue.h | Added function declarations for mute control and updated stream_session_init signature to accept start_muted parameter |
| openai_audio_streamer_glue.cpp | Implemented mute logic, replaced C-style casts with C++ casts, added handling for speech_stopped events, removed unused variable, and integrated mute state into audio streaming logic |
| mod_openai_audio_stream.h | Added user_audio_muted and openai_audio_muted bit fields to private_data struct |
| mod_openai_audio_stream.c | Added do_audio_mute function, mute/unmute API commands, updated default sampling to 24kHz, and added mute_user initialization flag handling |
| README.md | Updated documentation to describe new mute/unmute commands and clarified behavior of pause/resume commands |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Signed-off-by: Dario Pellegrino <dario.pellegrino@voismart.it>
dariopellegrino00
force-pushed
the
feature/user-audio-flow-control
branch
from
a0ec65a to
eba0a45
Compare
There was a problem hiding this comment.
Pull Request Overview
Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
dariopellegrino00
force-pushed
the
feature/user-audio-flow-control
branch
from
eba0a45 to
f362a43
Compare
There was a problem hiding this comment.
Pull Request Overview
Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
dariopellegrino00
force-pushed
the
feature/user-audio-flow-control
branch
from
f362a43 to
239ddf2
Compare
There was a problem hiding this comment.
Pull Request Overview
Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
) Not flushing audio buffer caused a delayed response upon unmute (caused by releasing old buffered user audio). Not sending an empty chunk of audio would not trigger model inference when user speech gets interrupted by mute. Also check pr #23 for more details Signed-off-by: Dario Pellegrino <dario.pellegrino@voismart.it>
dariopellegrino00
force-pushed
the
feature/user-audio-flow-control
branch
from
239ddf2 to
747c123
Compare
There was a problem hiding this comment.
Pull Request Overview
Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.
Comments suppressed due to low confidence (1)
mod_openai_audio_stream.c:1
- Double semicolons at the end of these declarations should be single semicolons.
/*
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Signed-off-by: Dario Pellegrino <dario.pellegrino@voismart.it>
Signed-off-by: Dario Pellegrino <dario.pellegrino@voismart.it>
Signed-off-by: Dario Pellegrino <dario.pellegrino@voismart.it>
Signed-off-by: Dario Pellegrino <dario.pellegrino@voismart.it>
dariopellegrino00
force-pushed
the
feature/user-audio-flow-control
branch
from
747c123 to
468fe2c
Compare
) Not flushing audio buffer caused a delayed response upon unmute (caused by releasing old buffered user audio). Not sending an empty chunk of audio would not trigger model inference when user speech gets interrupted by mute. Also check pr #23 for more details Signed-off-by: Dario Pellegrino <dario.pellegrino@voismart.it>
) Not flushing audio buffer caused a delayed response upon unmute (caused by releasing old buffered user audio). Not sending an empty chunk of audio would not trigger model inference when user speech gets interrupted by mute. Also check pr #23 for more details Signed-off-by: Dario Pellegrino <dario.pellegrino@voismart.it>
dariopellegrino00
force-pushed
the
feature/user-audio-flow-control
branch
from
468fe2c to
129d7cd
Compare
No description provided.