{{ message }}
Repository navigation
fix(omnivoice): normalize Arabic numerals before synthesis - #1203
Open
wei668899-luo wants to merge 1 commit into
Open
wei668899-luo wants to merge 1 commit into
wei668899-luo wants to merge 1 commit into
Conversation
OmniVoice reads bare digits poorly in the languages that fall back to num2words, and the damage is not subtle: the Vietnamese sentence "Tính đến ngày 11 tháng 10, tổng doanh thu phòng vé ... đã đạt 31 tỷ nhân dân tệ ..." comes out as unintelligible audio (Whisper-large-v3 can only hallucinate a stock Vietnamese outro line over it), while the identical sentence with the numbers spelled out transcribes back verbatim. Pass normalize_text=True so OmniVoice converts the numerals before synthesis. zh/en route through the optional WeTextProcessing extra, which raises when it is missing and has no macOS arm64 wheel because pynini is conda-only there, so those two languages only normalize when that package is importable. Every other language uses the num2words fallback, which leaves unsupported values alone instead of raising.
This branch has not been deployed
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
OmniVoice reads bare digits poorly in the languages that fall back to
num2words, and the damage is not subtle: the Vietnamese sentence "Tính đến
ngày 11 tháng 10, tổng doanh thu phòng vé ... đã đạt 31 tỷ nhân dân tệ ..."
comes out as unintelligible audio (Whisper-large-v3 can only hallucinate a
stock Vietnamese outro line over it), while the identical sentence with the
numbers spelled out transcribes back verbatim. Pass normalize_text=True so
OmniVoice converts the numerals before synthesis.
zh/en route through the optional WeTextProcessing extra, which raises when it
is missing and has no macOS arm64 wheel because pynini is conda-only there, so
those two languages only normalize when that package is importable. Every other
language uses the num2words fallback, which leaves unsupported values alone
instead of raising.