Mastering AI Voices Technologies

Discover offline text-to-speech with a Python package in Visual Studio Code, switching between Windows SAPI5, macOS, and Linux Espeak engines by installing the package and running a simple App.py script.

Learn to convert text to speech with gtts in Visual Studio Code, install the Python package, and generate audio via command line or Python, noting internet requirements and text limits.

Demonstrate running gTTS in Google Colab, install the package, convert text to speech, save the audio as output.mp3, and play it with IPython, including language changes from English to Vietnamese.

Explore 11 labs text-to-speech capabilities, including voice selection, voice changer, dubbing, text-to-sfx, and voice cloning in 32 languages. Learn how open source tools offer full control.

Explore Speechify's text-to-speech and voice cloning tools to convert documents into audiobooks, use transcription, switch languages, and generate custom audio with uploaded voices.

Learn how to generate audio from text with OpenAI text-to-speech via the ChatGPT API, selecting a model and voice and saving output as speech.mp3.

Explore VITS, an open source text-to-speech model, and learn to run inference on Google Colab by cloning the repo, installing libraries, loading the model, and generating natural audio.

Discover how Vitesse combines a one-stage approach to waveform generation with spectrogram-to-waveform steps and GAN-based adversarial training with a discriminator and generator to deliver high-quality, low-cost text-to-speech.

this lecture introduces piper, an open source text-to-speech project with 33 languages and multiple voices for local device use, optimized by onnx runtime.

Learn to run Piper TTS on Google Colab, install pip, download ONNX and JSON voice models in multiple languages, perform inference, and generate wav outputs.

Learn to run Piper TTS on Ubuntu by installing the Piper test, downloading Onnx and JSON voice models, and running a Python app to produce audio.

Discover how to clone a voice with Piper in Google Colab, from building a Vietnamese dataset and phoneme processing to fine-tuning, checkpointing, and Onnx inference.

Evaluate voice cloning quality across epochs, noting noisy outputs at ten epochs and clear words at 500–1000 epochs, prioritizing data quality and high-quality WAV audio.

Learn the structure of the LJ Speech dataset for voice training, including the CSV metadata with audio paths and transcriptions, and the importance of 50–300 diverse sentences for quality.

Automatically build a high-quality voice dataset by cloning a repo, splitting audio into 3–5 second chunks, transcribing with Google speech recognition, and exporting a metadata CSV for training.

Train an English voice cloning model in Piper using Google Colab, preparing a single-speaker dataset, resampling audio, converting text to phonemes, and exporting checkpoints to Onnx for inference.

Explore running Open Voice in Google Colab to clone English voices, install Aria, download checkpoints, and use the Open Voice Radio GUI for text to speech with a reference voice.

Clone voices in Google Colab using Open voice, extract tone color embeddings from source and reference voices, convert audio with tone color converter, and generate cross-lingual outputs with OpenAI text-to-speech.

Discover TorToise TTS, an auto regressive, high-quality text-to-speech model; explore voice cloning, open-source tools, and multilingual training across Coco AI, X test, and Google Colab.

Learn to fine tune a tts voice for English on Google Colab, covering dataset preparation, vocabulary expansion, checkpointed training, and running inference.

Fine-tune the tortoise tts model for Vietnamese in Google Colab, from cloning the repo and installing dependencies to training with a Vietnamese checkpoint and performing voice-conditioned inference.