Microsoft WAVLM and TDNN: Speaker Recognition Algorithms Explained

Microsoft WAVLMs use a variety of algorithms, such as Hidden Markov Models (HMM), Gaussian Mixture Models (GMM), and Support Vector Machines (SVM). They extract features from the speech signal such as mel-frequency cepstral coefficients (MFCCs) and form an acoustic model which is then used for speaker recognition.

TDNNs (time-delay neural networks) are used for speaker recognition in the same way as WAVLMs. They extract features from the speech signal and form an acoustic model which is then used for speaker recognition. TDNNs are able to capture the temporal context of speech, which can be used to improve the accuracy of the speaker recognition system.