INFRASONICON: Time-Compressing Infrasonic Recordings to Discover New Sounds, by Clark Huckaby
This transducer example derives an infrasonic signal from audio amplitude envelopes. This is a bit abstract, because such signals don't really accompany sound in nature. (Unless it's the infrasound radiated by analog VU meters. Imagine a giant one's motor driving a diaphragm instead of a pointer.) But music has a beat, and I decided that its time-compressed envelope might therefore have a tone. Its pitch should depend on the tempo, and its timbre on the groove.
|Figure 13: Audio Amplitude Detector. A separate Technical Page has the schematic and circuit description.|
I designed the Audio Amplitude Detector (Figure 13) for line-level audio (such as from a CD deck). It mixes stereo inputs to a single channel which feeds a precision full-wave active rectifier. This is followed by a low-pass filter with a 20-Hz cut-off. (Further 20-Hz low-pass filtering in the Data Converter makes sure audio frequencies can't be recorded--envelopes only.) An AC-coupling with a 1.6-second time constant gives the detector a consistent settling level (zero reference for the infrasonic signal). This is required since all audio amplitudes are positive. Offset and attenuation trimmers match the unit's output to the Data Converter's voltage input. Please refer to the Amplitude Detector's technical page for a schematic diagram and discussion of the unit's theory.
|Figure 14: Audio amplitude envelopes aligned with source waveforms. Drum machine and speech are from Sound Examples 8 and 9, respectively. The speech is a sound-bite from President Obama, part of a radio "talk" show on August 3, 2009.|
In Figure 14, two examples of amplitude envelopes are compared with their audio source waveforms. These examples are at opposite poles of the audio world's rhythmic continuum. A drum machine can generate pure rhythm; depending on program, its time-compressed envelope could pass for sonorous tones (Sound Example 8). However, human speech is not precisely rhythmic, and different syllables have a great diversity of envelope shapes; time-compressed 165-fold, it is heard as noise (Sound Example 9).
Most music envelopes are between these two extremes (Sound Examples 10-12). A tone is discernible for most songs, lasting about one second per three-minute track (most noticeable on Sound Example 11). But features that make songs interesting--syncopation, vocals, change-ups, fills, solos, etc.--add noise or break up the tones. Let me emphasize that high-speed envelope playback is not the same as fast-forwarding audio. Well beyond Chipmunk-scale, a simple 165-fold speed-up would make all but the bass tones ultrasonic, and the bass tones dog-disturbingly shrill.
Without upsetting dogs, I'm aware of one other method that can compress the listening experience to about one song per second. In 2006, R. Luke DuBois released a CD entitled Timelapse (Ref. 9). Using his "time-lapse phonography" digital algorithm, DuBois averaged spectral content across entire songs, summarizing each as a steady tone (or a rolling average for longer compositions). In a unique historical study spanning 41 years, he chronologically played the averaged tones of Billboard hits for one second per chart-topping week. While each method compresses time, high-speed envelope playback otherwise completely differs from time-lapse phonography: The former retains only the amplitude information (rhythmic or not), discarding all tone content. The latter keeps the tone while ignoring all rhythm and dynamics.
Overview concludes in Part 6==>
Go back to <==Part 4
Go to Home Page Go to Sound Example Page Audio Amplitude Detector Tech Page Contact Clark Huckaby
Go to clarkhuckaby.com
References Cited in Overview Part 5:
9. Time-Lapse Phonography:
Copyright 2011-2013 Clark Huckaby. All rights reserved. Commercial use of this content (text, images, sounds, or source code) without first obtaining the expressed permission of Clark Huckaby is prohibited by law. Distribution of copies beyond personal use and/or re-posting any or all of this page is prohibited by law.