Time Lag Accumulator
Time lag accumulation was the fitting name that Terry Riley used to call his techniques for using tape loops and longer tape delays back in the times where ‘looping’ was yet to become a well-established practice.
Inspired by his and others (such as Pauline Oliveros) virtuosic use of longer tape delays I wanted to apply similar methods to melodies and fragments of melodies.
I choose my first experiment to be on a ‘vallåt’ (insert link here) because of the limited harmonic content together with it’s rubato tempo would be well suited for a kind of uneven musical canon.
The scale I derived from the recording.
9/8 (ren sekund)
32/27 (litt låg moll ters)
27/20 (litt hög kvart)
15/11 (litt högare kvart men ikke helt opp till rein 11)
3/2 (ren kvint)
64/35 (neutral 7)
21/11 (hög dur 7)
2/1
On a more poetic level I also pictured this herding call being echoed back to an imaginary horn-player standing in a valley surrounded by mountains to the left & right.
My version of the time lag accumulator I used in this example is composed of two digital buffers of 14s and 15s that are read continually. The two buffers are hard panned left & right. The two buffers are continually being recorded to (from the melody we hear being played) without overdub, thus deleting the previous material. The two channels are hard panned left & right. The long and uneven times were employed for the effect of an uneven and unpredictable musical call and response.
Apart from the stereo panning I wanted to distinguish the sound of the echoes further subtly by emulating some qualities of analog tape (again inspired by Rileys take on the subject).
This was achieved by the use a tanh function (modulated by white noise) applied to the audio to add the grainy and compressing qualities of tape without the, for my purposes, distracting pitch wavering associated with tape delays.
Using digital buffers instead of analog tape opens up possibilities for playing back from arbitrary points in the recording and in doing so breaking away from the 1:1 recording / playback relationship. Later on, I activate a function which at random intervals of about 3-15s shifts the buffer to read and write from a new random point. My goal with introducing this randomness was to further dissolve the melody into an unrecognizable state as the piece went on.
To further conceal the looping points I gradually faded in and out the phrases (using my left hand as seen in the video) in order for the phrases to overlap more smoothly. I also was mindful to take longer pauses between the phrases as the time lag accumulator would later fill those in.
My aim was to have a gradual dissolving and abstraction of the original melody as the piece went on by introducing the randomization as well as introducing other improvised phrases that served to fill out the sonic space with a focus on establishing two harmonic distinct states. One with the steady fundamental and the other with the dissonant neutral seventh and high fourth.
This time lag accumulator patch I have since been used in concerts.