CRASH TEST

Pierre-Yves Macé's concert at the Festival "Octobre en Normandie" - October 2003

 

 

 

TRIVIA

October 11th, 2003, Orangerie du Jardin des Plantes, Rouen, France.

Music, Pierre-Yves Macé
Trumpet, Xavier Bornens
Cello, Ingrid Kuntzmann
Piano, Beata Suranyi
Oboe, Antoine Lazennec
I took care of the live processing & spatialization aspects.
One hour.

Pictures from the show

 

 

 

1. THE PLACE

 

This concert took place in a room which was not exactly "normal", since it was a greenhouse.
It was a choice PYM & I made - typical classical concerts in typical concert halls are not always exciting, to say the least.

The greenhouse was quite large and extremely reverberant - something like a 6 second RT60 reverb, very rich in frequencies around 100 - 200Hz.
There were no attempts to damp the reverberation by any means. The palm trees we asked for helped reduce the reverb's harmonic content, though.

 

 

 

2. THE PLACE VS THE MUSIC

 

There were four instruments, cello, trumpet / flugelhorn, oboe / barytone oboe, and piano.
From the beginning, one of the principles we decided to follow is that the score and the musicians had to adapt to the acoustics.
Thus, would Crash Test be performed in another room, it would be played in a different way.

Nowadays, in the classical music milieu, where the score is some kind of a holy bible, this principle seems to be a bit sacrilegious, but it's very natural - small baroque orchestras were small because at this time, music was played in "salons" for the aristocracy. Church music is very often slow, being played in very reverberant churches. Nowadays, rock bands use hugely powerful PA systems when they play in large venues - and any jazz player would tell you how differently they play whether they're giving a concert in a very small jazz club or in a larger venue.

 

 

 

3. TIMBRE ISSUES

 


* When playing in a place with such a reverberation, it's no use trying to ignore it.

In such a case, one has two choices - to reduce it or to take advantage of it. In accordance to the principles we were following, we chose the second solution.
To take advantage of the reverb, one must naturally consider it as part of the instrument timbre - then it becomes a musical parameter.
To get the reverb to be a proper musical parameter, one has to be able to play with it - to transform it, modulate it....


* Several methods were used to achieve this goal properly.

First, and maybe a bit unexpectedly, the right method was to add more reverb, using reverb processors. Thus it becomes possible to change the reverb's color at will. More specifically, in order to get a reverberated sound which was very different from the room's natural reverb, we ended up mainly using delays with a lot of feedback.

What's more, the added reverb was usually played on a single speaker, preferably far from the instrument - thus, everything was easily distinguishable : the instrument sound from one point ( the instrument itself ), the added reverb from another point ( the speaker ), and the room reverb from everywhere.

The third method is more obvious : the instument player were not always seated at the same place, but they actually moved a lot in the room, following "paths". Thus the natural reverb amount on the instruments was varying all the time.

And by setting up mikes in remote corners, the relationship between the instrument and the instrument's reverb could be inverted : instead of having the direct sound close to the audience and the reverberated sound far from it, you get the contrary - and it can become another musical parameter.

All those elements can be seen on the stage setup :

 
Stage setup

 

 

* A more difficult problem was the piano.

You don't want to get a highly reverberated piano, which will become completely blurred. In our case even the natural reverb was a problem, not to speak about actually adding reverb to it. Furthermore, you can't have the piano player -and the piano ! - moving around the room.

Thus the piano was treated the "normal" way : a lot of EQed direct sound in the speakers close to the audience - dry, no low freqs, boosts on hi med freqs, to make it remain understandable despite the difficult conditions.

 

 

 

4. FOLLOWING THE MUSIC

 

The Crash Test piece is made of 19 little pieces, each one having its own style and personality.
Thus it was necessary to give each one of the pieces a distinct sound personality, as if each piece was a distinct sound design object.

Some pieces were left dry ( no sound modifying or reinforcement except for the piano ), others were processed as explained previously, more or less heavily - in some cases, the natural instrument sound was completely covered by the reinforced / modified sound, which can be very spectacular when the instrument player is way right of the room and the modified sound is played by a speaker way left.

In other cases, especially solo or duet pieces, the sound was more radically transformed, using VST plug-ins which can modify the melody itself.
In such cases, the music becomes more electronic than acoustic.

Using those two methods, the whole concert became some kind of a game in which the distance between what's seen and what's heard varies constantly, and the limit between reality and virtual reality is always thin.

To be able to achieve this "sound flexibility", one naturally requires a couple of digital devices, as seen in the following diagram of the technical setup :

 

 
Technical setup

 

 

The O2R scene memories were extensively used, and rapidly appeared to be life savers !
... and to say the least, avoiding feedback when using such methods with acoustic instruments, without any prerecorded parts or feedback killers was no laughing matter...

 

 

 

5. THE CUBE



Additionally, mainly during solos / duets, we used the Cube, a particular speaker cluster I helped develop at Ircam .
To sum it up, the Cube basically acts as a variable directivity speaker.

For Crash Test, we used mainly a mobile figure of eight directivity, with a lot of negative phase impressions - "phase effect" use, as described in the following page on the Ircam web site :
http://www.ircam.fr/equipes/design/activites/cube/pres_cube_juin2002/index.html.

This effect, used in a very mobile way ( the figure of eight axis is moving constantly ), gave in this case a really "sparkling" result - bursts of sound coming sometimes from the Cube, sometimes from another place, sometimes from nowhere... quite entertaining.

The Cube was driven using Max MSP running on a PowerBook G4 with a Motu 828.



Additionally, when the sound of an instrument was directed towards the Cube, its sound was almost systematically heavily processed, mainly using a VST plug-in called "reorder", which "shuffled" the instrument part, creating a similar yet distinct part - this process being sometimes known as "hétérophonie" in french.

This method lead to a curious result : the Cube was almost like an instrument by itself, dialoguing live with the real instrument.
Additionnally, the use of the Reorder plug-in enabled us to play the Cube as loud as the original instrument without having to fear any feedback, though the instrument and the Cube were quite close to each other.

Here is a screen capture of the Max patch used for this concert - with "scene memory like" recalls, plug ins, and several trackballs, corresponding to several Cube directivity behaviours between which it was easy to switch.

 
MSP patch