DavidCecchetto.net

Simple Explanation of SRMP (Skewed Remote Musical Performance)

Accomplished through the open source software SuperCollider, SRMP consists of two remotely situated musical performers who collaborate in real time. The key parameter of SRMP, though, is that the sound heard in each of the two locations must be markedly different, and must differ such that the specificity of the differences is not anticipatable or captured. For example, SRMP's initial performance took place simultaneously in San Diego, CA and Victoria, BC, and used sound-sample players that accessed a bank of approximately 100 sound files. At any given time, each performer could articulate any of the sound files using a musical keyboard interface, which allowed the performers to select the sample, select an effect chosen from a bank of nine signal processors, select the parameters of the effect, and indicate the articulation's duration/rhythm.

(For those who are unfamiliar with digital signal manipulation, these signal processors behave analogously to the 'distortion pedal' that is frequently heard in popular guitar music. For example, the signal processors that were used in the premiere performance included standard digital tools such as granular processing, reversal, and pitch and time shifting. Thus, a pre-recorded sample of a sound that is stored in the computer as a digital file is subjected to an algorithm that alters it in a predetermined way. In the case of granular processing, for example, the algorithm divides the sample into small 'pieces' and 'scatters' them according to certain principles; the resulting sound contains the same sounds as the original sample, but cut-up and re-ordered in a (seemingly) chaotic fashion. Further explanation of this technique can be found in the Wikipedia entries for granular synthesis and Barry Truax.)

In order to introduce a significant difference between the sounds heard at the two locations, the computer randomly turned on and off a skewing mechanism that altered which sound file was played at the remote location, but did not alter the effect applied to the sample (see Fig. 1). For example, the Victoria performer may play a sound file of a motorcyle processed by pitch and time shifting so that it sounds like a scratchy violin; while this sound is heard in Victoria, in San Diego the audience might hear a different sound produced by applying the same effect to a different sound file (such as a fast chattering, if the effect is applied to a sound file of a deep human voice).

Theoretically, the skewing that is used in SRMP could be applied to any sonic parameter (or visual parameter, for that matter) that the software can recognize (which is to say, any parameter that can be digitally manipulated). In practice, the skewing used in SRMP has tended to combine a series of binary choices (i.e. whether or not to apply a sound effect at all; whether or not to substitute a randomly chosen effect for that which was remotely specified; whether to apply the effect uniformly or only to certain sound samples; whether to articulate the effected in addition to the original sample or as a substitute for it; etc.) with a continuum of dynamic  specifications (i.e. how small will the 'grains' in the granular processing be; to what extent is pitch-shifting applied; what are the relative volumes of the processed samples to the originals; etc). This is to say that the skewing does not introduce new parameters to the performance, but rather alters the sound of the parameters already included in the performance interface. Similarly, skewing has not been applied to formal compositional structures of relation in these performances, but functions instead at the level of the performance itself (though I am currently experimenting with algorithmically skewed formal relations in another context). With all this said, the most important things to note are that skewing happens between both locations, the sound is never skewed locally, and the local performer never knows if or how the sound is skewed remotely.

SRMP is typically presented as structured improvisation, with the specific structural parameters varied from performance to performance. Limited audio documentation of SRMP, titled Two excerpts from SRMP, is available at www.davidcecchetto.net/artistic.html.

The specific networking protocols for SRMP were developed by William Brent as part of his doctoral research at the University of California at San Diego.

----------------