2 Commits

Author	SHA1	Message	Date
sl-jetson	14164089dc	feat: Audio pipeline end-to-end (Issue #503) ... - Add VoskSTT class to audio_utils.py: offline Vosk STT backend as low-latency CPU alternative to Whisper for Jetson deployments - Update audio_pipeline_node.py: stt_backend param ("whisper"/"vosk"), Vosk loading with Whisper fallback, CPU auto-detection for Whisper, dual-backend _process_utterance dispatch, STT/<backend> log prefix - Update audio_pipeline_params.yaml: add stt_backend and vosk_model_path - Add test/test_audio_pipeline.py: 40 unit tests covering EnergyVAD, PCM conversion, AudioBuffer, UtteranceSegmenter, VoskSTT, JabraAudioDevice, AudioMetrics, AudioState - Integrate into full_stack.launch.py: audio_pipeline at t=5s with enable_audio_pipeline and audio_stt_backend args Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-03-07 10:03:31 -05:00
sl-firmware	6f3dd46285	feat: Add Issue #503 - Audio pipeline with Jabra SPEAK 810 ... Implement full audio pipeline with: - Jabra SPEAK 810 USB audio I/O (mic + speaker) - openwakeword 'Hey Salty' wake word detection - whisper.cpp GPU-accelerated STT (small/base/medium/large models) - piper TTS synthesis and playback - Audio state machine: listening → processing → speaking - MQTT status and state reporting - Real-time latency metrics tracking ROS2 Topics Published: - /saltybot/speech/transcribed_text: STT output for voice router - /saltybot/audio/state: Current audio state - /saltybot/audio/status: JSON metrics with latencies MQTT Topics: - saltybot/audio/state: Current state (listening/processing/speaking) - saltybot/audio/status: Complete status JSON Configuration parameters in yaml: - device_name: Jabra device pattern - wake_word_threshold: 0.5 (tunable) - whisper_model: small/base/medium/large - mqtt_enabled: true/false with broker config Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>	2026-03-06 10:30:58 -05:00