# lipsync-degraded

> Talos 0.7 · Since 0.7

the face cannot make every viseme as written: its model lacks some viseme or ARKit mouth shapes (or has only a jaw, or no mouth shape at all, or no morph targets), or the `track` did not load, did not parse, or was made for another sound (the mouth follows loudness instead), or the voice streams so its loudness is not read at load. The message says which

| | |
|---|---|
| Severity | `warning` |

## Meaning

the face cannot make every viseme as written: its model lacks some viseme or ARKit mouth shapes (or has only a jaw, or no mouth shape at all, or no morph targets), or the `track` did not load, did not parse, or was made for another sound (the mouth follows loudness instead), or the voice streams so its loudness is not read at load. The message says which

## Fix

give the model the 15 `viseme_*` shapes (or the ARKit mouth set), fix the track file the message names (`argus-visemes` writes valid ones), or give a long voice a visemes track