Assess text-to-speech audio quality for intelligibility and acceptability.
npx skills add https://github.com/mattmireles/kokoro-coreml --skill audio-judge
Please help me install this Agent Skill. Skill: audio-judge Source: https://github.com/mattmireles/kokoro-coreml/tree/main/.claude/skills/audio-judge Command: npx skills add https://github.com/mattmireles/kokoro-coreml --skill audio-judge
The Skill addresses the issue of evaluating the audio quality of synthesized speech outputs, enabling efficient decision-making regarding TTS acceptability, intelligibility, and comparison with reference clips.
To judge the quality of the generated TTS audio, run 'node scripts/run-audio-judge.mjs --clip coreml=/path/to/synthesized_audio.wav --baseline-label reference'.
High-intent search queries and answers about installing and using this skill.
You can assess TTS audio quality by running automated scripts that evaluate synthesized speech for intelligibility and acceptability. This approach enables rapid decision-making regarding text-to-speech acceptability and comparison with reference clips.
You compare synthesized speech against a reference clip by running an audio assessment script with parameters for the synthesized audio path and a baseline label. This enables side-by-side comparison to check if the TTS output matches expectations.
Yes, you can automate text-to-speech evaluation for new software versions. The automated audio assessment checks if the generated audio output matches expectations compared to the standard or previous versions.
Yes, you need Gemini to evaluate generated speech for intelligibility. The audio quality assessment requires Gemini and specific scripting to perform the evaluation of text-to-speech synthesis outputs.
Automated audio assessment for text-to-speech pipelines is a process that quickly evaluates the quality of audio clips generated from TTS synthesis. It applies to pipelines requiring rapid evaluation of generated speech for intelligibility and acceptability.
pavan53732Generate podcast scripts and WAV audio from text or web search results.
GeneralReasoningClean, normalize, and time TTS audio segments for delivery with FFmpeg.
akrindevGenerates speech from text using Google Gemini TTS models via Python scripts.
michaelboedingConvert text into speech using Gemini, ElevenLabs, or OpenAI TTS backends.
