mARTin-B78/tts-voice-creator-clone-and-design-2
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

docs(audiobook): record the rejected scene-cast experiment

Browse Source 
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
This commit is contained in:
mARTin-B78 2026-09-19 20:54:22 +02:00
parent f34266c0d2
commit ded70c3853
2 changed files with 5 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
TODO.md
View File

@ -345,8 +345,9 @@ Still open (each needs a real decision or new evidence, not more prompt tuning):
until the book itself ties it to a name. until the book itself ties it to a name.
- [ ] Look-alike main characters in one scene (Marcian/Zerwas/Lancorian) remain the biggest real error - [ ] Look-alike main characters in one scene (Marcian/Zerwas/Lancorian) remain the biggest real error
class (~1/5 of spoken words). Tried and rejected: reasoning mode (25× slower, no gain), larger/smaller class (~1/5 of spoken words). Tried and rejected: reasoning mode (25× slower, no gain), larger/smaller
windows, tag-based overrides, reasoning adjudication of disputes (made disputed lines worse). Untried: windows, tag-based overrides, reasoning adjudication of disputes (made disputed lines worse), an explicit
an explicit scene-cast constraint (characters present in the scene) built from the text. scene-cast hint (names mentioned in the passage: +0.3 points, noise). What remains is mostly model
capacity plus reference noise; the useful lever left is human review guided by the agreed/disputed flags.
--- ---

3
architecture/audiobook-attribution.md
View File

@ -132,7 +132,8 @@ lines) but compared against a biased baseline; on the real pipeline it made the
so the adjudication step (and the `candidates` request field) was removed. The only reliable signal is so the adjudication step (and the `candidates` request field) was removed. The only reliable signal is
agreement itself: agreed lines are right 82 % of the time, contested first answers 67 %. agreement itself: agreed lines are right 82 % of the time, contested first answers 67 %.
Sampling matters: windows of ~1 200 characters were clearly worse (−4 points) and ~5 000 about equal to Sampling matters: windows of ~1 200 characters were clearly worse (−4 points) and ~5 000 about equal to
the default, so the default stays. the default, so the default stays. A scene-cast hint (which known characters are named in the passage or
just before it) gave +0.3 points — noise — and was not adopted.
Request additions: `attribute_only` `[ids]` — the other spans are context; unrequested lines are never Request additions: `attribute_only` `[ids]` — the other spans are context; unrequested lines are never
reported missing. reported missing.