Nanda Vishnu - PureData

Holy crap ok. The patch referred to in this thread finally got completed to a workable point. You have no idea how happy I am.

I'm posting screenshots of each page so you have an idea of what's going on.

Well, I'm gonna need to double post anyway so......

Ok, so here's what happens in this patch.

This patch was based on a load of research I did on Indian music. While the sources for the raga and tala that the piece is based on might not be totally legit (kksongs.org), its existence gave me a building block to start from.

The raga provides three major things to this piece. firstly, it provides the notes to be used. It also provides the probabilities for the markov chain to determine which note to choose. Thirdly, I treat the Swarupa as the gat, a recurring composed section, occuring at 80 seconds, 100 seconds, 120 seconds, 200 seconds, and 340 seconds.

The tala provided an initial basis for structure, but later only the length resembled the tala - 17*2*10sec=340, the length of the piece. However, the tala provides to the piece the probabilities for bol (stroke) choice.

The final piece, the overall structure, came from a book which went in great detail into Hindustani rhythm and structure.

pd structure controls the patch. it uses a 10-second (10,000 ms) timer to act as a general metronome for the piece, ending at the 35th tick. At the same time, a 3000ms metronome acts as the true tempo marking for the piece. This later become 750ms and 428ms. pd structure controls most parts of the patch, such as envelope, gate threshold, and tala speed.

The rhythm of the "melody" is derived from a hard gate with a changing envelope on a pink noise signal, which changes to give the illusion of speeding up. The gate bangs (triggers) a complex sine wave using info from this thread to simulate a pizz violin, a Bb clarinet, an alto flute with vibrato, and a martele stroke on a contrabass simultaneously and with timbre changes at various points of frequency.

Both the pink noise and the complex tone go through a quick reverb and EQ. The EQ page has a real-time effect over each step, cuz it looks cool.

Going on at all times is a drone of the tonic and perfect fifth of the scale, as well as a drone of both a M3 and a m3, since I found the inclusion of both highly unusual about the raga.

Starting at the first tempo change (to a 750ms pulse), the drums kick in and continue relatively constantly throughout.

The frontend is still being worked on, and will be a visualizer of sorts.

I don't think it is clear from the output, neither do I think it is important, but the website claimed that the raga I used is meant to signify joy entering into night.

The sound file is in mono, but the actual patch is in stereo. Space starts and stops, 1/2 starts and stops the drone.

I think that's it.

LINK! http://www.zshare.net/audio/17683562aae4e07b/

Well I have no clue whats going on in those pictures, but it seems very complicated and you obviously took a lot of time on it. So, why don't you tell us the basics of what I'm seeing.

It's some sort of synthesizing program, am I right?

Oh sorry, I was typing while you were typing the explanation.

.wav link - zSHARE - nandavisnu.wav

This is fantastic! I love it. (And kind of makes me wish I had a subwoofer )

Nice way of combining algorithmic composition with traditional elements very uniquely, but without sounding arbitrary (which algorithmic composition sometimes tends to). It has a very strong personal colour and works fine formally. And the sounds are much less cliché than in much other electronic music. Really nice job!

I'm too lazy to study the path in detail, particularly in screenshot form, plus I'm not too familiar with pD, only Max/MSP (which is of course very similar, but still somewhat different), but I think I get more or less how it works.

Thanks for the reply!

I actually may have cut off the end of this or something. at any rate, it is the general idea.

Yeah, I'm at a pretty conservative composition program, so I have to make sure it "sounds good" before I claim something like this is done. But that's what's so awesome about the Markov chain - its pretty effective at modelling note choice. I'd actually like the model to be tighter, but I'd need to define things very well and that's a lot of obscure reading or assertion that I don't want to do.

What do you do with Max?

Well, I haven't done any serious composing with Max so far, I've more used Csound for that. (Just a matter of personal preference. I prefer programming the music in text to dragging lots of tiny boxes around the screen and connecting them with cables. For lots of people it's just the other way round.)

What I love Max for however are live electronics, control mechanisms etc., which all are quite impractical in Csound. Right now I'm programming a control algorithm that turns 6 tape recorders on and off via a microcontroller for an installation in about two weeks. Fun stuff!

I've often felt a lot of music created with Max (or pD) sounds a bit similar, which makes it even better that your piece -doesn't- just sound like that. And you're probably right that one can still optimise some things about the formal structure there, even if it already works quite fine. I guess you can always improve things in any composition though

Yeah, text programming is a lot less tedious for sequencing and stuff. I just can't visualize the patch in text. I was taught about 2 seconds of ChucK - wish I learned more

Yeah I feel the same way about pD music - I went to the ICMC last year when it was at my school and saw the Max and Pd concert, twas less than impressive. Other parts were cool, though.

This patch actually could really easily be based from an input, come to think of it.

Ferkungamabooboo, since you prefer graphic programming why don't you try Reaktor? It's a NI software and as I can see from the wikipedia entry and NI homepage I think it works exactly like the program you use. I would try the demo version, but it's too complicated for my level and with the knowledge I have (I don't even understand how you created music from these graphs. All I can see is boxes containing numbers and connected with lines; lots of lines. ).

Alexandros