GameCult/AquaSynth
Watch
9
0
Fork
You've already forked AquaSynth
 0
Code Issues Pull requests 4 Projects Releases Packages Wiki Activity Actions

Add speech timing receipts

Browse source
This commit is contained in:
Meta Vens 2026-05-20 22:25:32 +01:00
commit eb2156ab37
8 changed files with 128 additions and 13 deletions
Download patch file Download diff file Expand all files Collapse all files

6
docs/ipa-vocal-tract-roadmap.md
View file

@ -497,6 +497,12 @@ Ground truth means "reference pressure," not "the truth about throats." eSpeak
can teach early intelligibility, timing, and rough spectral targets. It cannot
decide AquaSynth's anatomy, coarticulation, or Weksa morphology.
Training receipts must keep time as a witness. Every request, result, and
checkpoint should record when each stage made its decision, how much latency it
spent, what latency budget it was meant to respect, and how confident the
mapping was. A tiny audible witness that sounds right but arrives from a
clockless swamp is already teaching the model to lie with a lovely voice.
## Optional eSpeak NG Workout
The first IPA-adjacent parity lane uses eSpeak NG as a development reference.