Visualization of shared subspace representation to discriminate 30 speech commands invariant to the speaker. The visualizations are generated using t-SNE; Click anywhere to begin audio