Train a simple deep learning model that detects the presence of speech commands in audio. The example uses the Speech Commands Dataset [1] to train a convolutional neural network to recognize a given set of commands.

Demonstrates a machine learning approach to identify people based on features extracted from recorded speech. The features used to train the classifier are: pitch of the voiced segments of the speech, and the Mel-Frequency Cepstrum Coefficients (MFCC). This is a closed-set speaker identification - the audio of the speaker under test is compared against all the available speaker models (a finite set) and the closest match is returned.

Isolate a speech signal using a deep learning network.

Identify a keyword in noisy speech using a deep learning network. In particular, the example uses a Bidirectional Long Short-Term Memory (BiLSTM) network and mel-frequency cepstral coefficients (MFCC).

Denoise speech signals using deep learning networks. The example compares two types of networks applied to the same task: fully connected, and convolutional.

Detect regions of speech in a low signal-to-noise environment using deep learning. The example uses the Speech Commands Dataset to train a Bidirectional Long Short-Term Memory (BiLSTM) network to detect voice activity.

Classify the gender of a speaker using deep learning. In particular, the example uses a Bidirectional Long Short-Term Memory (BiLSTM) network and Gammatone Cepstral Coefficients (gtcc), pitch, harmonic ratio, and several spectral shape descriptors.

Classify spoken digits using wavelet time scattering paired with a support vector machine and a deep convolutional network based on mel-frequency spectrograms.

Classify the genre of a musical excerpt using wavelet time scattering and the audio datastore. In wavelet scattering, data is propagated through a series of wavelet transforms, nonlinearities, and averaging to produce low-variance representations of the data. These low-variance representations are then used as inputs to a classifier.

A typical workflow for feature selection applied to the task of speech emotion recognition. You begin by creating a baseline accuracy using common audio features (MFCC). You then augment your data set to diminish overfitting. Finally, you perform sequential feature selection to select a better feature set.

Create a multi-model late fusion system for acoustic scene recognition. The example trains a convolutional neural network (CNN) using mel spectrograms and an ensemble classifier using wavelet scattering. The example uses the TUT dataset for training and evaluation [1].

Perform pitch tracking using multiple pitch estimations, octave and median smoothing, and a Hidden Markov Model (HMM).

Time-frequency masking is the process of applying weights to the bins of a time-frequency representation to enhance, diminish, or isolate portions of audio.

Use the Levinson-Durbin and Time-Varying Lattice Filter blocks for low-bandwidth transmission of speech using linear predictive coding.