Official repository of DailyTalk: Spoken Dialogue Dataset for Conversational Text-to-Speech, ICASSP 2023
Extract phoneme-level timestamps from speeh audio.
ManaTTS is the largest open Persian speech dataset with 114+ hours of transcribed audio. Includes data collection pipeline and tools. Suitable for Persian text-to-speech models.
A python library to generate speech dataset from Youtube videos
Extract a target speaker’s clean, non-overlapped speech from multi-speaker audio and export word-safe LJSpeech-style TTS datasets.
ParsiGoo is a Persian multispeaker dataset for text-to-speech purposes. It includes recordings from different speakers and is designed to be used for training and evaluating text-to-speech models.
A free licensed Persian TTS dataset including 6+ hours of audio-text pairs with subject
LibriVox dataset for Bulgarian language TTS
VoxPolish - a TTS dataset cleaner. Remove noise, trim silence, normalize loudness for text-to-speech training data in one command.
Persian to Finglish dataset with all the sentences voice for TTS dataset used to train tacotron2
Blind people face several challenges when reading books, but the main problem is a lack of accessibility to printed materials. This can limit the opportunities for blind people to gain knowledge and improve their education and employment prospects. The mobile-based Sinhala book reader for visually impaired individuals can provide a solution.
Construct voice from web-videos and then clone it!
top dataset for voice conversion models
Python pipeline that turns raw audio/video into structured speech-training data: acquisition, speaker diarization, and emotion tagging, with RTTM/JSON/CSV outputs.
Turn raw audio into clean speech datasets. Auto-transcribe with Faster-Whisper, review & edit transcripts with audio preview, and export paired files + metadata in one click.
Turn any audio file into a TTS training dataset
কথার রং Kothar Rong — Open Bangla Slang Dataset: 3,318 entries / 20 categories + 505-row cross-lingual slang→Bangla reverse-mapping layer (EN/HI/UR/AR/KO/JA/ES, process-labeled) + 239K-comment usage corpus + 2,303 audio clips + usage evidence. CC BY-SA 4.0
Automated voice dataset factory for Speech AI with VAD silence trimming, loudness normalization, and LJSpeech metadata generation
To associate your repository with the tts-dataset topic, visit your repo's landing page and select "manage topics."