posted 06 August 2005 12:28            

sitting_in_a_room.zip

I am having fun playing with the CrossFilter.
I was trying to simulate the effect done in the Alvin Lucier's piece "I am sitting in a room". I am feeding a CrossFilter with the sentence "I am sitting in a room" and I use the response of a room with a n impulse response of about 800 ms. The result goes in a MemoryWriter and is fed back into the CrossFilter. Again, and again.

But that doesn't work.

I have attached an archive of my sound.
One is not working, and another one which has an attenuation instead of the CrossFilter, demonstrates ok.

Please could you tell me what's wrong with that sound?

By the way, I forgot to mention I was using 3 expansion cards, which makes a sum of 10 DSPs.

Thanks !

posted 06 August 2005 14:24            

sitting2.kym

try using the FeedbackLoopInput/Output pair for these kind of things.

attached is a modified example, note that you'll have to drop the feedback input level otherwise it will overload the feedback network.

posted 18 August 2005 14:17            

Sitting3.kym

Love the sound . I've added a compressor in the feedback path to regulate it (stop it overloading or dieing away).

posted 20 August 2005 04:53            

Thanks for your contribution!
Pete, I would like to know something about the CrossFilter algorithm : is it an implementation of convolution? I am aksing this to myself because I found that feeding the CrossFilter with impulse responses gave failry good results. Anyway, being convolution or not, it can be used to make creative reverbs, and thanks to SSC and you for this new prototype!

About the sound: I heard about Alvin Lucier's "I am sitting in a room" in the Tweaki. So I bought the record, being curious to know what effectively would give the idea of recording multiple generations of a recorded voice in a room.
At one point in this recording, the voice of Alvin Lucier sounds like it is vocoded ; it is something VERY strong and natural. I guess I feel it this way because the voice has been recorded on an analog tape, and most important, the room resonated nicely with complex resonant and absorbant modes.

I wondered which reverb or processing could actually produce the same vocoding effect in the digital realm. I believe the CrossFilter a start, but I need to fine the correct impulse response in order to have to correct "vocode-alike" resonances ...

posted 22 August 2005 12:33            

With regard to the vocoder sound, I wonder if you could cut out a bit of the original recording and post it so that I can hear what you mean. One thing to note with your sitting in a crossfilter is that if you leave the sound running for half an hour or so you end up with the sound of just a few continuous sine waves. These are all the sine waves that are common to both the impulse and the signal (of course averaged out over each signals duration), so maybe an impulse with a much fuller frequency responce ( more white-noise-like) may make the input signal more transparent and give this vocoder type effect.