Repository navigation
Conversation
The hosted glados.c-net.org endpoint used by voice-gen-glados.py had become unreliable and was producing badly truncated audio. After trying several local alternatives (glados-tts, Piper+RVC voice conversion, Style-Bert-VITS2 - each hit either audio-quality or dependency issues), a community-trained GLaDOS voice model on the hosted Fish Audio API (https://fish.audio/) produced the best results. generate.py now calls the new voice-gen-glados-fishaudio.py for the en_gb-glados voice pack instead of the old hosted-endpoint script, and the SOUNDS/en_gb-glados audio has been regenerated with it. Also fixes a few phrases (e.g. "3D Mode", "timer 1/2/3 reset") whose embedded digits were being mispronounced, by spelling them out in the Translation column. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01EwwCu3hzZwiDPjsd6HusQW
git-lfs was only needed for the now-removed local glados-tts checkout. Also updates voice-gen-glados-fishaudio.py's module docstring now that it's the script generate.py actually calls, not one of several experimental options. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01EwwCu3hzZwiDPjsd6HusQW
1 task done
1 task done
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
glados.c-net.orgendpoint used byvoice-gen-glados.pyhad become unreliable and was producing badly truncated audio.glados-ttsForward Tacotron+HiFiGAN, Piper+RVC voice conversion, Style-Bert-VITS2) but each hit audio-quality or dependency issues.voice-gen-glados-fishaudio.pyscript.generate.pynow calls it instead of the old hosted-endpoint script.SOUNDS/en_gb-gladoswith the new pipeline."3D Mode","timer 1/2/3 reset") whose embedded digits were being mispronounced, by spelling them out in theTranslationcolumn ofvoices/en-GB.csv.Test plan
SOUNDS/en_gb-gladosaudio (SYSTEM + SCRIPTS/BETAFLIGHT/INAV/YAAPU) sounds correct on-deviceuv run generate.pystill runs the Azure/ElevenLabs jobs unaffectedFISHAUDIO_API_KEYsetup instructions in the README are accurate🤖 Generated with Claude Code
https://claude.ai/code/session_01EwwCu3hzZwiDPjsd6HusQW