MUSAN is a Music, Speech, and Noise Corpus.

The NOISE dataset is a semi-synthetic dataset constructed from the matrix A∗, where the data is generated from y = A∗x + ζ, where ζ is the noise.

The dataset is used for investigating the capabilities of a neural spatiospectral filter to represent and exploit spatial information.

Deep Matrix Approximately Nonlinear Decomposition (DEMAND) employs Adam and an alternative optimization strategy that is well suited to optimize convex/alternative convex...

The Demand database [30] used to create noisy conditions in the training data.

Speech enhancement dataset

The TIMIT corpus is a widely used benchmark for speech recognition tasks. It contains 3,696 training utterances from 462 speakers, excluding the SA sentences. The core test set...

MNIST-ADD dataset is a dataset of handwritten digits with added noise.

The BrainWeb database is a collection of MRI images with known noise levels.

The Rician distribution of noisy MRI data.

The starting set consists of noise (random tiles) that is produced from a predefined distribution such as a uniform distribution.

The WHAM! dataset is used for testing the proposed Bayesian factorised speaker-environment adaptive training and test time adaptation approach for Conformer models.

The dataset used for training the CNN in cough detection is composed of various modified audio clips gathered from open-source online sources. Each of these audio files...

Gaussian Noise, Uniform Noise, Tin (C), Tin (R)

The VoiceBank-DEMAND dataset is a standard benchmark for speech denoising systems. It consists of 28 speakers with 4 signal-to-noise ratios (SNR) (15, 10, 5, and 0 dB) and...

The DNS challenge provides clean speech, noise, impulse responses, and a training data synthesizer for both non-personalized and personalized DNS models.

A dataset for exploring the effect of noise and architecture on neural machine translation.