What is codec degradation?
Codec degradation is quality loss introduced by audio compression, heard in music streaming, broadcast and telephony alike. In a voice AI pipeline, that same loss can confuse downstream automatic speech recognition (ASR) or voice activity detection (VAD).
What is an example of codec degradation?
A call routed over a landline or a low-bandwidth VoIP connection sounds thin or metallic, and a transcription engine mishears words as a result.
How does codec degradation work?
Compression discards audio information to shrink file size or bandwidth; below a certain bitrate, what gets discarded includes detail that ASR and VAD models depend on, beyond what a human ear would notice missing.
How does ai-coustics address codec degradation?
Codec degradation is one of the six Tyto Dimensions. Tyto detects when it’s happening and surfaces it as a score, so a team can act – reroute the call, flag it, or adjust downstream expectations – instead of finding out only after a transcript comes back garbled.
