Alva Noto + Ryuichi Sakamoto - Insen Tour Part 1

Setup, Design and Programming

Carsten Nicolai aka Alva Noto asked me if i’d like to do the visuals for the upcoming tour of him and Ryuichi Sakamoto and to be on tour with them. That sounded exciting to me. I agreed immediately. The project started around May 2005.

Custom Instrument Design

At first Carsten wanted to have a custom designed music instrument that could compete, in terms of presence, with the piano that Mr. Sakamoto would be playing on stage. I spent some weeks working out the design, the hardware including touch panels and the overall system setup only to find out that the deadline was approaching too quickly. We finally agreed on cancelling the custom hardware because of this. The fallback solution was two Powerbooks with four of the wonderful, now unfortunately discontinued, Faderfox LV-1 controllers, which i was able to order in a custom version with only white, grey and black buttons.

Rebuilding the tracks for a live performance

Carsten used software like for example Logic Audio to produce and arrange the original tracks on Vrioon and Insen. He explained to me how the tracks were made. On Vrioon it was roughly speaking a technique of analyzing a frequency of the piano samples that he got from Mr. Sakamoto and using this frequency for sinewaves he layered on top. On Insen he sampled small bits of the piano and arranged them in a special way. I won’t go more into detail here. He suggested to use Ableton Live for the basic arrangement, which i agreed, since the software i would build to complete the setup would be able to sync with it. Carsten wanted me to rebuild the sequencial steps it took him to produce the original tracks in the form of a realtime instrument. So i got his original Logic Audio tracks and transfered some of them to Live. I figured out which parts would be played back with Live and which parts were the ones that the custom software would take care of.

The system setup

In parallel i thought about the general system setup. Carsten was used to a Mackie VLZ 1402 Mixer that he wanted to use on stage, so i bought one for myself, as i was working on the project in Frankfurt and he worked in Berlin. I made a layout of the setup of instuments on stage and the way they are connected. A Powerbook for Ableton Live, another for a custom build MaxMSP based instrument that Nibo and i would develop, the Yamaha Disklavier, a PowerMac that runs the visuals on the LED screen. And all the Mixers, Midi and Firewire gear in between. I ordered the equipment, connected everything in my office and tested it. In Berlin David Lettelier, Nibo and Carsten rebuilt the setup.

The Insen Step Sequencer “inseq”

I started with the software part by trying out all the various possibilities to sync Ableton Live and MaxMSP, which i programmed the step sequencer with. In the end Ableton Live was the master and MaxMSP the slave, synced via MIDI Clock. I showed the first version of the software to Carsten when we were installing another Installation for him at the Akademie der Künste in Berlin. That was when i could convince him that the Lemur interface wasn’t the right thing to use, because you can’t build custom user interface objects for it, like we could in Max.

The sequencer consists of five tracks. The BPM can be synced to the incoming MIDI Clock or run offline. An incoming audio signal can be recorded in realtime and played back with the step sequencer immediately. You can alter the amount of steps, the panning of each step, the overall volume of a track, the attack and release and the part within the sample that is being triggered by the sequencer. Everything can be controlled with one of the Faderfox LV-1 controllers, including cueing before the signal is sent to the main mix.

Jhno helped me to get some issues with clicks at the samplestart resolved and Jeremy Bernstein released a special version of his pattrstorage objects in MaxMSP for me.

The Vrioon Nibosen Instrument

As i still had to do all the visuals and we ran out of time, Carstens assistant Nibo was commissioned to program the Vrioon Instrument, also in MaxMSP. It is used for the first track “Impro” at the concerts. He also programmed Carstens website, so he’s biased to design as well. He made a very nice and clean layout that Carsten liked, although the white background was the exact opposite of what Carsten suggested in the first place. Finally both instruments were packed into one MaxMSP patch in order to reduce the burden on the computers CPU.

The visuals

In the beginning of the project Carsten thought about CRT monitors for the stage, but when he was presented some Barco LED screens he was convinced that they would be much better suited for the concerts. He came up with the idea of several LED panels that would be positioned about one and a half meters above the stage floor, but seperate from another, so i programmed a patch in Jitter that would allow to animate a continuous movement along the panels and automatically took care of the gaps in between. Soon after the layout evolved to a narrow but broad band of LED elements without the gaps.

Because Carsten relied on my aesthetic expertise i was completely free in developing the visuals. I got the original Logic Audio files of the tracks and started transcribing two of the piano scores into MIDI files. I installed a Grand Piano VST plugin on my IBM notebook to simulate the piano play of Ryuichi Sakamoto. I would use the MIDI output of his Yamaha Disklavier to trigger the visuals in some of the tracks. Then i rearranged Carstens parts from the Logic Files into Ableton Live clips for some of the tracks.

For controlling the visuals in realtime with the music i could use the MIDI from the piano, MIDI Clock from Ableton Live and 4 separate audio channels for an FFT analysis. Each of the 8 channels of the Faderfox controller had a patch connected. Moving the Fader upwards would start the connected patch and fade it in smoothly, moving it down would fade out and stop the patch respectively. I prepared different patches for about eight music pieces. For the track “noon” Sebastian Gregor, the main mastermind behind VVVV, allowed me to use his pool shader which Jhno implemented in Jitter.

Then about five days before we were going to have the rehearsals in Berlin, Carsten gave me a call and said that Forma, the production company, supposedly had forgotten me in the budget for the tour! He asked me to prepare everything that he could control the visuals himself on stage. Until then i had taken it for granted that i would control the visuals at all the concerts myself. He said i would only be on the tour for the first two or three concerts to make sure everything runs perfectly.

Rehearsals

I came to Berlin with my own PowerMac, Display and all of my equipment. I slept at a friends home, because there was no budget for a hotel for me. The PowerMac for the concerts was ordered and would arrive four days before the first concert.

When i arrived i met the Insen crew for the first time. Ryuichi Sakamoto sitting at his Powerbook writing for his blog, greeting friendly. He’s very cautious and attentive and was very nice to me whenever we met. Nibo, Carstens young assistant from Tokyo, who does a lot of graphic design for him and also built his website. David Lettelier, who just recovered from a cancer therapy. He released the beautiful Stabil record under the pseudonym Kangding Ray a few month later. Kamal, technical director at Forma, took care of the LED Wall. Nigel Edwards, the chummy lighting designer, and last but not least Dominic Nieper, the smart and always organized tour manager.

The LED Screen was supposed to be installed in its entirety, but instead there was only the leftmost part of it, about 2 meters of 11 meters all in all. I had two major problems now. The visuals looked completely different from what i had prepared and seen only on my TFT Display until now. They were programmed with gradients and smooth looking. The LED Screens were older models from Lighthouse with a very low resolution. The video rental company didn’t have the promised Barcos in stock any more. And i could see almost none of my visuals, because they were generated algorithmically and most of the time they were in the middle of the virtual 11 meter screen, not on the far left, which was all that could be seen on the small section of the LED Screen.

Carsten and Ryuichi were very concentrated on the music, rehearsing piece by piece. Carsten had prepared most of the tracks and they had to find out how to sync Ryuichi’s live piano play with the clips in Ableton Live. They almost didn’t care about the visuals. Most of the time nothing could be seen on the far left screen anyway. But also i still had a lot to do and continued programming and adjusting till late at night while everyone else had dinner and went to bed after that.

Almost the only time when Ryuichi and Carsten decided on the visuals was when i presented the japanese typography which was later used for Uoon. Some of the characters (that i couldn’t read) had a meaning that Ryuichi didn’t like and some other characters’ shape Carsten didn’t like, so we took those out. Carsten also suggested to use some variation of a patch that i had used for another track of him.

In the evenings and in the early mornings the piano tuner (forgot his name, sorry) came in. He had to adjust the Disklavier because it sounded to harsh in Sakamotos ears. This took about three nights. In the morning the Piano had to be tuned before Ryuichi would arrive. While tuning it had to be absolutely quiet. Fortunately i had modified my mouse to be quiet some time before (actually i always wanted to make a business out of selling quiet mice).

Two days before we left for the first concert the visuals still weren’t finished and Carsten got a bit impatient. The software wasn’t stable. Sometimes the visuals would freeze and i didn’t know why. Just in time for leaving i found the bug and at the last rehearsal Carsten and i were quite confident with the visuals.

If i had seen the way the visuals look on the LEDs some time before and if we had more time i would have done the visuals completely different. The gradients just didn’t look good on the LEDs. The LEDs also had a very steep brightness curve. If you send 20% white you will get 0% white on the LED. It starts about 35% and ends about 65% i guess.

I made one new patch from scratch in about half an hour which considered the LEDs resolution. Thats the grid-patch that you can see at the last encore (Ax Mr. L.). It fits exactly on the LEDs and doesn’t have any blurring to the next LEDs, giving a very sharp and precise look.

5 October 2005 | Concert | Comments