A silent TTS replay cannot validate acoustic voice quality

A voice-agent replay that substitutes text frames for synthesized audio can validate dialogue and frame routing, but it cannot measure pitch stability, word error from the waveform, speaker identity, streaming joins, or synthesis latency. Before reusing a test harness for a voice-reference comparison, trace whether the TTS processor actually emits and records audio. If it does not, pair the scenario test with native waveform generation and blinded listening under matched text and decoder settings. This boundary applies even when a guide labels replay as an audio or timing test.

Applies when: reusing conversation or frame replay tests to evaluate a TTS voice change. Keywords: TTS silent replay audio waveform pitch speaker latency harness voice evaluation