v1.2.0

Added

• @steelbrain/media-speech-detection-web: Add preloadModel() function for faster speech detection initialization
  • Preloads the Silero VAD ONNX model by fetching it into browser cache during app initialization
  • Optional but recommended for optimal user experience in production applications

v1.1.0

Added

• @steelbrain/media-ingest-audio: Add channelId option to select specific audio channel from multi-channel streams

  • Enables processing of individual channels from stereo/multi-channel audio sources
  • Useful for stereo interview recordings, multi-channel audio interfaces, and professional audio setups
  • Defaults to channel 0 (first channel) for backward compatibility

• @steelbrain/media-ingest-audio: Add sampleRate option to configure AudioContext sample rate

  • Allows custom sample rates from 8kHz to 96kHz+ for different use cases
  • Supports telephony (8kHz), voice (16kHz default), music (44.1kHz), and professional audio (48kHz+)
  • Automatically handles sample rate conversion from input to target rate
  • Useful for optimizing quality vs. bandwidth based on application needs

Fixed

• @steelbrain/media-ingest-audio: Fix audio worker continuing to send messages after stream closure
  • Added try/catch around controller.enqueue to handle closed stream errors
  • Properly cleanup resources when stream is cancelled while audio worker is still processing

v1.0.0

Initial release 🎉

Includes the following packages:

• @steelbrain/media-ingest-audio
• @steelbrain/media-speech-detection-web
• @steelbrain/media-buffer-speech