In this video I look at Coqui’s new XTTS v1 text to speech model, and complain about licensing. Then I look at a couple tools, pyannote and Speechbrain, and use a model to generate and compare audio embeddings. This can be used to identify mismatching audio clips in your datasets. Remove poor quality clips, and […]
Read More… from A look at XTTS v1 and Tools for Comparing Audio Embeddings
In this video, I look at text cleaners, and how they could be potentially causing issues with training your TTS models. I refer to cleaners in the Tortoise TTS AI Voice Cloning WebUI (MRQ) and Coqui TTS Messy and unfinished LJSpeech-format dataset markup/processing script: […]
Read More… from Are Text Cleaners Making Your TTS Models Sound Bad? | TTS Model Training Tips
Narration of the short story ‘Selling Point’ by Norman Arkawy using a Tortoise TTS model generating a familiar-sounding, expressive, British voice. Originally published in ‘Imagination Stories of Science and Fantasy’, December 1955.One of my favorite short stories. Full story text:https://www.gutenberg.org/cache/epub/66713/pg66713.txt Training: 10 epochs total. Epochs 1-4 LR 1e-5, 5-6 LR 1e-6, 7-10 LR 1e-7 Mel/Text: […]
First test of the new Tortoise model. 4 voices, which also can be found in the YourTTS model I posted recently. LJS, John, and Tom rendered without any stammers or repeats, Lah has some stutters if I recall. No cherry-picked examples. Gen settings same as in my Tortoise fine tuning video, except denoise set to […]
I’ve been looking at multispeaker VITS TTS models lately, so thought I’d share the Google Colab notebook. Its similar to the others posted, but this is using precomputed vectors; the configuration is similar to the YourTTS model, however this seems a little easier to fine tune. As always, this stuff is experimental, but this should […]
This is about as close to automated as I can make things. I’ve put together a Colab notebook that uses a bunch of spaghetti code, rnnoise, OpenAI’s Whisper Speech to Text, and Coqui Text to Speech to train a VITS model. Upload audio files, split and process clips, denoise clips, transcribe clips with Whisper, then […]