NTIA/alignnet
Train no-reference speech quality estimators with multiple datasets via learned, per-dataset alignments.
This tool helps speech and audio researchers develop more accurate algorithms for automatically assessing speech quality. It takes multiple audio datasets, which might use different scoring scales, and trains a 'no-reference' model to produce a consistent quality score, even if those datasets weren't originally designed to work together. The output is a robust speech quality estimator.
No commits in the last 6 months.
Use this if you need to combine several independent datasets of speech audio with subjective quality ratings to train a single, reliable speech quality estimation model.
Not ideal if you only have a single, perfectly consistent dataset for training, or if you are not working with no-reference speech quality estimation.
Stars
18
Forks
—
Language
Python
License
—
Category
Last pushed
Aug 01, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/NTIA/alignnet"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
voicepaw/so-vits-svc-fork
so-vits-svc fork with realtime support, improved interface and more features.
sarulab-speech/UTMOSv2
UTokyo-SaruLab MOS Prediction System
ssmall256/mlx-audio-io
Native audio I/O for MLX on macOS and Linux
ssmall256/mlx-spectro
High-performance STFT/iSTFT for Apple MLX with fused Metal kernels and autograd support
daniilrobnikov/vits
VITS: Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech