Add fluid_voice_set_callback() for voice lifecycle notifications#1778
Merged
Conversation
Implement a new public API function fluid_voice_set_callback() that allows applications to register a callback to be notified when: - A true noteoff is processed (FLUID_VOICE_CALLBACK_NOTEOFF) - The voice finishes playing and is removed from DSP (FLUID_VOICE_CALLBACK_FINISHED) Changes: - Add fluid_voice_callback_reason enum and fluid_voice_callback_t typedef to include/fluidsynth/voice.h - Add fluid_voice_set_callback() declaration to public header - Add callback/callback_data fields to struct _fluid_voice_t - Implement callback invocation in fluid_voice_release() and fluid_voice_stop() - Clear callback fields on voice re-initialization in fluid_voice_init() - Add unit test test_voice_callback.c Agent-Logs-Url: https://github.com/derselbst/fluidsynth/sessions/3c1d7aa5-1096-44d4-ae88-631b25ff90c3 Co-authored-by: derselbst <8152480+derselbst@users.noreply.github.com>
…ext API call Move finished voice processing into fluid_synth_render_blocks() right after fluid_rvoice_mixer_render() completes. This ensures the FINISHED callback fires as soon as the voice finishes rendering, regardless of whether any subsequent synth API call is made. Agent-Logs-Url: https://github.com/derselbst/fluidsynth/sessions/735fc152-b315-4555-a553-672d551927b3 Co-authored-by: derselbst <8152480+derselbst@users.noreply.github.com>
… thread The previous approach called fluid_synth_check_finished_voices() from fluid_synth_render_blocks(), which created two consumers on the SPSC finished_voices ringbuffer — a data race between the render thread and the API thread (in fluid_synth_api_enter). New approach: - Add finished callback fields to fluid_rvoice_t (function ptr, voice ptr, data) - Add fluid_rvoice_set_finished_callback() event function dispatched via the event queue from fluid_voice_set_callback() - Fire the FINISHED callback directly from the mixer's fluid_mixer_buffer_process_finished_voices() in the render thread, before pushing to the ringbuffer - Remove FINISHED callback from fluid_voice_stop() (now redundant) This maintains the SPSC invariant: the render thread is both the producer of finished voice notifications AND the invoker of the callback, while the API thread remains the sole consumer of the ringbuffer. Agent-Logs-Url: https://github.com/derselbst/fluidsynth/sessions/735fc152-b315-4555-a553-672d551927b3 Co-authored-by: derselbst <8152480+derselbst@users.noreply.github.com>
…safety Eliminate the intermediate trampoline that read voice->callback from the API-thread-owned fluid_voice_t. Instead, store the user's callback function pointer directly on the rvoice (ABI-compatible: void* vs fluid_voice_t*, int vs enum). This ensures the render thread only reads data from the rvoice (set via the event queue), never from the API-thread's fluid_voice_t. Agent-Logs-Url: https://github.com/derselbst/fluidsynth/sessions/735fc152-b315-4555-a553-672d551927b3 Co-authored-by: derselbst <8152480+derselbst@users.noreply.github.com>
Member
Author
|
@copilot SonarQube complains about too much code duplication in the newly added test/test_voice_callback.c. Pls. fix that for me, thanks. |
Agent-Logs-Url: https://github.com/FluidSynth/fluidsynth/sessions/8f0c89cb-66e9-4e4e-97bc-870c80534532 Co-authored-by: derselbst <8152480+derselbst@users.noreply.github.com>
Contributor
Fixed in 68271c8. The three near-identical voicelist-scan-and-configure blocks are now consolidated into a |
|
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Currently, apps using
fluid_synth_start()/fluid_synth_start_voice()have no way to know when afluid_voice_tor its ID becomes invalid — i.e. the voice may finish playing before even receiving a noteoff (e.g. short unlooped percussion samples). This adds a per-voice callback for two lifecycle events: true noteoff (release phase entered, not held by sustain/sostenuto) and voice finished (removed from DSP loop).Example Usage