further down the spiral

First off, I absolutely love this blog. I've spent many hours up at night going through it and reading all about Radiohead and their awesome noises. I have some quick questions about the Max/MSP stutter patch that Jonny uses in the "Go To Sleep" solo; does he activate the effect via a footswitch? Also, after watching countless live videos of Radiohead's performance of this song, it seems as if when the effect is active, his guitar is muted. Is that true? Thanks for creating this blog!

Hello there, glad to hear that you’ve found the site informative!

Jonny’s laptop is essentially in an “effects-loop”. Whether his signal goes through his laptop is controlled by a Boss LS-2, which is the very first pedal in his guitar “signal chain”. He generally keeps it on the floor to the left of his pedalboard. When the Boss LS-2 is inactive, Jonny’s guitar signal goes straight through it to the rest of his pedals. When the Boss LS-2 is active, Jonny’s guitar signal goes through his computer (via an audio interface) before going to his other effects.

In this photo from a 2003 Radiohead performance (Camden, NJ, USA), one can see Jonny’s “laptop” Boss LS-2 to the left of his pedalboard. The LS-2 on top of the power supply at the top left corner of his left pedalboard switches between his guitar and his Korg Kaoss Pad. The LS-2 on the right pedalboard switches between two signal chains – one going to his Vox amp (SD1, RV3, RE201, Headrush) and one going to his Fender amp (Shredmaster, FV300).

Once Jonny opens the Go To Sleep Max/MSP patch on his computer, it is always “active”. Whether or not we hear its glitching, rather than Jonny’s dry signal, is determined randomly by the patch itself. He just has a random generator running, and when it hits a certain value the patch loops some audio for “x” amount of time. All that Jonny does is hit his Boss LS-2 to incorporate the patch into his signal flow, then plays a lot of notes to ensure that the patch always has information to loop.

In addition to whether or not the patch loops audio, both the length of time that the loop is repeated and the loop length are random. The patch determines if there will be 1 second of a 50ms loop, or 500ms of a 250ms loop. I don’t know for sure whether the loop lengths are always an (random) integer subdivision of the length of time that the loop is repeated, but it would make sense for it to be in order to avoid “clicks” (although there are ways around that).

The Max patch that Jonny uses on Go To Sleep can essentially be thought of as a delay with 100% feedback. It records a bit of audio, then plays it back, feeding it both to the output and back into itself. But on a normal delay at 100% feedback, the delay circuit would also be receiving the dry guitar’s signal, and the combination of the new and old signals would build up and push the delay to oscillation. Part of what differentiates Jonny’s patch from a conventional delay is that when the patch decides to loop a clip of audio, it stops receiving any of the dry signal. As such, the initial audio clip is played over and over again at full volume (the 100% feedback means that it never gets quieter), but there is nothing new being added to the delay which could push it into oscillation.

A photo of what is very likely a version of the Go To Sleep patch, from the 2012 Tour.

You are correct that when audio is being looped, Jonny’s “dry” signal is muted. It is probably muted at the very start of the patch, which is also why the delay doesn’t receive its signal and start to oscillate. There are a few ways that it could be implemented in the patch, however, so it’s hard to say.

By the way, I’m working on a way to create this effect without the need for Max or computers. Stay tuned!