Electronic Interlude: Part 3 of 3

An image I created using MetaSynth's "Image Synth."  Each pixel represents a note in time.

In this third and final part of "Electronic Interlude," I will share my most recent, completed electronic project, which I have called, "In Hindsight, the Treat Would Have Been Better." As was the case with the piece I shared in part 1 of 3 of this "miniseries," this most recent work fulfilled an assignment, viewable by clicking here.

Like "Invitation to Groove" (discussed in part 1 of 3), "In Hindsight, the Treat Would Have Been Better" was built from a database of collected and modified sounds. But in this case, each sound was generated with a specific tool in MetaSynth – the Image Synth. With the Image Synth, any picture you draw (or import) is translated into a sound file.  The rules are simple, but the possibilities are endless.

Like the Effects Room (also discussed in part 1), the (horizontal) x-axis represents time. In other words, the Image Synth reads and translates the image into sound going from left to right. (There's no word yet as to whether or not the Hebrew version of MetaSynth synthesizes the image from right to left.) If the image is 500 pixels wide, then each column of pixels corresponds to 1/500 of the resulting sound. Exactly how long this is depends on the total duration set by the user; if the whole picture represented 10 seconds, then 1 column of pixels would be 0.02 seconds.

The (vertical) y-axis represents frequency – in general, the higher up in the image you go, the higher pitched the sounds will be. Every row of pixels represents a single frequency (or pitch).  By default, going from one row to the next yields a sound one semitone higher, like moving up the notes on a piano in a chromatic scale. However, the tuning system can be altered in an infinite number of ways, yielding the aforementioned endless possibilities. (Think of it like opening up a piano and retuning all the strings however you see fit.) For instance, instead of having each successive row of pixels be a semitone higher, each could be a whole tone higher, 1/8 of a tone higher, etc. And the rows of pixels need not be equally spaced to correspond with the equal temperament system widely used today – you could have a Pythagorean tuning, for example. Putting these two axes together is just like reading sheet music: at any given moment you could have one note sounding or several (up to 1024); and there is always the possibility of silence as well.

So what about the varying brightness and color? The brighter a pixel, the louder the corresponding note will be.  As for color, red represents the left channel, and green represents the right channel. Yellow, the 50-50 mix of red and green using the RGB color model, subsequently corresponds to the center. (Observant readers will notice that in the sample image above, there is no blue — the blue channel is used for the grid, in case the user wishes to line up particular elements within the picture. Blue, like black, corresponds to silence.)

Below are two examples of sound files created with the Image Synth. Each image is played twice (the second image is a variation on the first; notice the differences).

The instruments used in the examples above are entirely computer generated.  I combined three different waveforms (shown below) which are played by the computer at the appropriate frequency.

If you think everything mentioned so far is pretty awesome, then be careful your mind doesn't get blown by what I will demonstrate next. (And if you say "Meh." to the aforementioned demonstrations, hopefully what follows will be more intriguing.)

First, there are more interesting artistic opportunities – not every image has to be a series of lines. Consider the image below, which is notably more abstract:

So that's one aspect of the Image Synth that helps illustrate how endless the possibilities really are. But suppose you got really technical and argued that because there are a finite number of pixels and a finite number of colors available to the user that the possibilities aren't really endless. This brings us to the second really remarkable feature of the Image Synth. Even if you buy the preceding argument (which is flawed because of all the ways one can design the waveform that acts as the instrument), the case will still fall apart because you can build your own instruments. In other words, you can take any sound you can get your hands on (or microphone near) and use that as an instrument. And the sound does not even have to be all that complex.

Consider a simple finger snap. Using a recording of my own fingers snapping, and another one of the images I "painted" (shown below), I created a rather unique sound. Here is the image:

And here is the resulting sound (played twice), using the finger snap as the instrument:

With a little reverb added (which you can hear in the final piece), this sound takes on a character of its own.

One other image I want to share can be found below:

For the corresponding sound, I went for a couple extremes. First, I slowed the tempo down considerably, to 5.33 beats per minute (at this rate, a measure in 4/4 time would take 45 seconds). The picture, 512 pixels wide, takes 3 minutes to play. (This became the "sound bed" for my piece, on top of which I lay all the other sounds.)

Second, I changed the tuning system to Micro32. Here, every row of pixels is only 1/32 of a whole step; the frequency range of the image is about 135 Hz to 338 Hz, roughly C#2 to E3 (i.e., approximately a minor tenth). Even though I used a simple, pure sine wave as my instrument, the resulting sound is pretty dramatic:

Now that you have seen part of the artistic project and a few of the building blocks of the piece, I present to you now "In Hindsight, the Treat Would Have Been Better." (You can view the program notes here.)