Poem in October

In class, there was some interest in a frequency analysis of "Poem in October". The image below is the spectral analysis of the mp3 file. Time is the horizontal axis, frequency is vertical axis, and intensity is color (blue -no amplitude to red - peak amplitude).

For sake of comparison, here is the poem represented in the time domain:

While conceptually intriguing, a spectograph offers little insight into the feel or inflection of the speech. What is more noticeable is the amount of noise (the shades of light blue and green) in the recording, which is a combination of several factors: microphone selection, recording medium, ambient sound, and mpeg compression. Of note, such distortion lends itself to certain adjectives, some of which arose during our phenomenological discussion of the piece- warm, brown, etc.

The recording is also frequency limited. Human hearing (and hi-fidelity recordings) extend up to 20 kHz, it is the range responsible for adding 'air' or 'presence' to a sound. The bulk of Thomas' voice lies within the 1-3 kHz, which is average for a human voice (see below).

Better illustrated is the dynamics, or changes in volume, of Thomas' performance. As seen in the waveform above (Figure 2), his narrative arch rises and falls very naturally.