floravox: warn when the G2P stack degenerates to letter spelling - #33
Merged
Conversation
Without a language (no 'lang' credential resolving a bundle, no
lexicon/Phonetisaurus/ByT5), the OOV chain ends at RuleFallback and,
with no lexicon, every word is OOV - the voice reads words spelled out
('a r e ...') with no diagnostic. This bit VoiceGarden-SAPI with
coqui-en-ljspeech (fixed client-side, but the engine should say so).
build_phonemizer now reports whether any real G2P stage loaded; when
none did and no character frontend covers the voice (MMS-style char
tables are exempt), synthesize logs a one-time warning naming the
remedy. Tested: empty config and unopenable lexicon paths both flag
real_g2p=false.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Warns (once per synthesizer build) when the floravox G2P stack degenerates to letter spelling — no lexicon/Phonetisaurus/ByT5 loaded and no character frontend covers the voice. Every word is then OOV and gets spelled out ("a r e ...") with no diagnostic.
Found in the wild: VoiceGarden-SAPI routed
coqui-en-ljspeechto floravox without alangcredential (fixed client-side in AACTools/VoiceGarden-SAPI#16, but the engine should say so).build_phonemizerreturns aPhonemizerStack { phonemizer, real_g2p }—real_g2pis true only when a lexicon actually opened, Phonetisaurus opened, or ByT5 loaded (configured-but-unopenable paths do not count)synthesize_forwarns when!real_g2pand no char frontend is active (explicitcharscredential or auto-detected char table — CharFrontend voices are exempt by design)real_g2p=false(11 floravox lib tests pass withsapi,cloud,sherpaonnx,floravox-lexicons)