Using temporal information for improving articulatory-acoustic feature classification

This paper combines acoustic features with a high temporal and a high frequency resolution to reliably classify articulatory events of short duration, such as bursts in plosives. SVM classification experiments on TIMIT and SVArticulatory showed that articulatory-acoustic features (AFs) based on a combination of MFCCs derived from a long window of 25ms and a short window of 5ms that are both shifted with 2.5ms steps (Both) outperform standard MFCCs derived with a window of 25 ms and a shift of 10 ms (Baseline). Finally, comparison of the TIMIT and SVArticulatory results showed that for classifiers trained on data that allows for asynchronously changing AFs (SVArticulatory) the improvement from Baseline to Both is larger than for classifiers trained on data where AFs change simultaneously with the phone boundaries (TIMIT).