← Back to the chapter · Previous detail · Next detail →

Technical detail · Audio Feature Extraction for IoT

Read the pre-emphasis, overlap, and taper plots

Three formulas have now been introduced, and each of them does something to the signal that a plot makes obvious in a second. Inspect the three panels of Figure 5 in order; the numbers marked on each plot are the ones from this chapter’s own 16 kHz pipeline.

Three stacked panels. The first plots the pre-emphasis magnitude response for alpha 0.97, marked with a gain of 0.03 at DC and 1.97 at the 8 kHz Nyquist frequency. The second shows a speech-like waveform with three overlapping analysis bars, a 25 ms frame and a 10 ms hop marked by a double-headed arrow. The third compares a dashed rectangular window against a Hamming taper falling to 0.08 at both ends.
Pre-emphasis changes relative spectral gain, overlapping frames preserve repeated views of events, and a Hamming taper reduces sidelobe leakage. These choices change what the model can measure.

Panel one of Figure 5 is the whole argument for α\alpha: the response climbs from a gain of 0.03 at DC to 1.97 at Nyquist, so slow components are suppressed relative to rapid change — which is why the same curve can amplify high-frequency sensor noise on a machine that is not a talker. Panel two makes the overlap concrete: the 25 ms frame bar and its two successors are offset by the 10 ms hop, leaving 240 shared samples, so an event landing near one frame edge is analysed again inside the next. Panel three sets the Hamming taper against the rectangular window drawn as a dashed outline; the taper’s ends fall to 0.08 instead of stopping abruptly, which is what trades sidelobe leakage for a wider main lobe. Read together, the three panels are three measurement decisions, and none of them is recoverable from a model file that only says “MFCC”.

← Back to the chapter · Previous detail · Next detail →