Audio Format Specifications

​

Individual Streams (user.raw, agent.raw)

• Encoding: PCM signed 16-bit little-endian (pcm_s16le)
• Sample Rate: 16000 Hz (16 kHz)
• Channels: Mono (1 channel)
• Bit Depth: 16-bit

​

Interleaved Stream (interleave.raw)

• Encoding: PCM signed 16-bit little-endian (pcm_s16le)
• Sample Rate: 16000 Hz (16 kHz)
• Channels: Stereo (2 channels)
• Channel Layout: Both channels contain the same mixed audio (user + agent)
• Bit Depth: 16-bit

​

Converting Recordings to WAV

# Convert user audio
ffmpeg -f s16le -ar 16000 -ac 1 -i user.raw user.wav

# Convert agent audio
ffmpeg -f s16le -ar 16000 -ac 1 -i agent.raw agent.wav

# Convert interleaved audio
ffmpeg -f s16le -ar 16000 -ac 2 -i interleave.raw interleave.wav