Music 171 homework 3

Music 171 homework 3
Psychoacousticians have long been interested in the phenomenon of streaming, in which a succession of sounds is heard as a single stream, like a melody or a person speaking. Cues that might contribute to hearing sounds as belonging to a stream might be proximity in time, pitch, amplitude, timbre, or physical location, This assignment is to reproduce a famous experiment (I believe due to David Wessel) in which a loop of six sounds is heard as a single stream if cycled slowly, but as two separate ones if cycled more quickly. The sounds in the loop are as follows:

PITCH TIMBRE
 65   flute
 69   oboe
 72   flute
 65   oboe
 69   flute
 72   oboe

Since we're studying sampling this week, the assignment uses two recordings: flute (pitch G4 or MIDI 67) and oboe (pitch A4 or MIDI 69); these are from freesound.org, by Carlos_Vaquero and acclivity respectively, and have a creative commons license (attribution; non-commercial): https://creativecommons.org/licenses/by-nc/3.0/.

At low speed (two notes per second) your successful patch should sound like this, and faster (10 per second) it should sound like this; notice that in the first case you hear an ascending three-note chord but in th esecond you hear two interlocked descending three-note chords.

To make the patch (each of the following steps is worth 2 points for a total of 10):

make a sampler (just one voice is fine; it doesnt' have to be polyphonic), of the "line~" type as described in class 1/17 and 1/22. The book doesn't go into this until Chapter 3 (example C05.sampler.oneshot.pd). Because the the recordings both don't attack quickly enough for this example to work well, the playback should start one second into the recording.
Using pack or otherwise, fix the sampler so that it has three controllable parameters: which recording to play, what transposition, and how loud. (The third is necessary only because the oboe recording is louder than the flute one and you will want to correct that to make them more even.)
To get the transposition working it will be fine just to play a fixed amount of time (10 seconds, say, which is longer than you'll ever need) and calculate a variable "distance" in the recording to play in the fixed time. This way, a transposition of 1:1 should "play" ten seconds of recording, or 2:1 should play twenty, and so on.
For ease of use, add conversion so that you can type "60" to ask for no transposition, "61" for "up one half-step", etc. If you do this the pitches above would be played if you specify transpositions in the loop of
```
58, 60, 65, 56, 62, 63
```
(the flute transpositions are two higher than the corresponding oboe ones because the original flute recording is two half-steps lower than the oboe one.)
Finally, make a looper controlled by metro (I'll show this in class 1/22). It should cycle though siz messages, each of which has three numbers to specify a transposition, which recording (1 for flute and 2 for oboe), and how loud to play it (I chose 1 for the flute and 0.5 for the oboe). The sample player will have to unpack this. To change the recording you can use the "set" message to tabread4~, which will need the name of the table holding the recording; a "select" and two message boxes will do this fine.

Warning: this one is harder and more complicated than the first two were - you might want to start early on this!

back to music 171 main page