New Ideas for Pulsar Modular
-
- Posts: 1139
- Joined: Tue Apr 03, 2001 4:00 pm
- Location: Tennessee, USA
- Contact:
-
- Posts: 1139
- Joined: Tue Apr 03, 2001 4:00 pm
- Location: Tennessee, USA
- Contact:
-
- Posts: 1139
- Joined: Tue Apr 03, 2001 4:00 pm
- Location: Tennessee, USA
- Contact:
-
- Posts: 1139
- Joined: Tue Apr 03, 2001 4:00 pm
- Location: Tennessee, USA
- Contact:
-
- Posts: 1139
- Joined: Tue Apr 03, 2001 4:00 pm
- Location: Tennessee, USA
- Contact:
Great idea A.
I'd be happy to see almost ANY development for ModV2. But first up I'd like a better sequencer - maybe a display with 32 steps. And an "export to audio" function would be superb. It's so good in Rebirth (and various other programs) to be able to have a drum or synth loop and export to a perfectly cut wav file.
I'd be happy to see almost ANY development for ModV2. But first up I'd like a better sequencer - maybe a display with 32 steps. And an "export to audio" function would be superb. It's so good in Rebirth (and various other programs) to be able to have a drum or synth loop and export to a perfectly cut wav file.
-
- Posts: 1139
- Joined: Tue Apr 03, 2001 4:00 pm
- Location: Tennessee, USA
- Contact:
a quad crossfade module. a la the tg33. with a visual x, y joystick and cv inputs to control the x and y axis.
i'm trying to work out how to do this with the crossfade modules, but as i've not dealt with a modular synthesizer before i'm having a little trouble susing out if this is possible/how to do.
i'd also appreciate a oscilloscope module so as to be able understand what some of the more esoteric mod routings/signal chains do to a signal.
being able to enclose a group of modules into a macro of one object would be nice. i have been playing with a graphics synth i discovered lately that has this option (onadime composer).
i'm trying to work out how to do this with the crossfade modules, but as i've not dealt with a modular synthesizer before i'm having a little trouble susing out if this is possible/how to do.
i'd also appreciate a oscilloscope module so as to be able understand what some of the more esoteric mod routings/signal chains do to a signal.
being able to enclose a group of modules into a macro of one object would be nice. i have been playing with a graphics synth i discovered lately that has this option (onadime composer).
- kensuguro
- Posts: 4434
- Joined: Sun Jul 08, 2001 4:00 pm
- Location: BPM 60 to somewhere around 150
- Contact:
How about going in a totally different direction? I've seen some requests on alternate synthesis techniques like physical modeling, granular etc.. I think it would be logical to have some sort of fft or stft analyzers with minimal support for algorhythmic
control.
The general "lean" towards subtractive and FM is actually sort of getting old... and I think there's only so much you can do with these methods.. What's interesting is that the STS5000's time stretching and pitch/formant shifting ability relies heavily on phase vocoding and heterodine analysis/resynthesis... So it means that Pulsar does have a way to do FFT related tranformations in real time.. Perhaps a spectral morph synth/effect shouldn't be too hard to implement since STS5000 seems to handle phase vocoding to a pretty high level. (though it probably won't reach kyma's level in terms of quality)
And perhaps a small scripting or programming package to tranform the analyzed data would be nice. As you all know, sts5000 requires a short analysis before you do the pitch/formant shifting abilities but what if you would alter the data with certain codes? What if there was a linear predictive or some sort of predictive module or perhaps even a 4 pole pitch detecting code and you could control read pitch information off an FFT analysis? Then you could control an EP's note number with your voice... and then perhaps modulate the EP/strings morph ratio with the amplitude of your voice.. you see where I'm heading? With some more heavy R&D it could even be possible to only use the odd harmonics of strings blended with the even harmonics of a guitar.. the possibilities are endless..
Sorry for my blabbering, I've had too much csound lately and I'm practically going berserks trying to help my friend finish up his graduation thesis... sheesh.. I still don't get why I'm feeding him with vital ideas..
But anyway, a little of juice from this field of dsp research could help in building some new wacky synths.
How about analyzing a specific audio, chopping it up into specific time segments(which could be changed according to how much amplitude change there is in the next segment), look for the average amplitude action there are, and then giving that as a label for each segment... that would give us a matrix right? (probably amplitude action for one axis and perhaps some other derivative as another axis) Then, that matrix, or a 2D plane could be used in conjunction with another axis.. say the cut off of a simple resonant lowpass filter.. that would give us a 3 axis 3 dimensional "space" sort of. Then... this is the punch line. You could end up with a 3 dimensional vector synthesizer with FFT/granular characteristics! Wow! Now that's a monster I'd like to work with.
Oh well, I'm getting a bit personal with this FFT thing.. but, you get the idea.. it can open new horizons, and with a little developement from creamware, it seems pretty realistic to me. errr... maybe JUST to me.
Ah ha! How about colaborations with IRCAM or ICMC people? They've got some really crazy/strange ideas.. I just read a '98 paper about having artificial lips made of latex blow a trumpet to see how real it can get... physical modelling? I don't know but it sure is a wacky synthesizer! haha:lol: So I think we could use a little of their craziness.
<font size=-1>[ This Message was edited by: kensuguro on 2001-10-08 15:56 ]</font>

The general "lean" towards subtractive and FM is actually sort of getting old... and I think there's only so much you can do with these methods.. What's interesting is that the STS5000's time stretching and pitch/formant shifting ability relies heavily on phase vocoding and heterodine analysis/resynthesis... So it means that Pulsar does have a way to do FFT related tranformations in real time.. Perhaps a spectral morph synth/effect shouldn't be too hard to implement since STS5000 seems to handle phase vocoding to a pretty high level. (though it probably won't reach kyma's level in terms of quality)
And perhaps a small scripting or programming package to tranform the analyzed data would be nice. As you all know, sts5000 requires a short analysis before you do the pitch/formant shifting abilities but what if you would alter the data with certain codes? What if there was a linear predictive or some sort of predictive module or perhaps even a 4 pole pitch detecting code and you could control read pitch information off an FFT analysis? Then you could control an EP's note number with your voice... and then perhaps modulate the EP/strings morph ratio with the amplitude of your voice.. you see where I'm heading? With some more heavy R&D it could even be possible to only use the odd harmonics of strings blended with the even harmonics of a guitar.. the possibilities are endless..
Sorry for my blabbering, I've had too much csound lately and I'm practically going berserks trying to help my friend finish up his graduation thesis... sheesh.. I still don't get why I'm feeding him with vital ideas..
But anyway, a little of juice from this field of dsp research could help in building some new wacky synths.
How about analyzing a specific audio, chopping it up into specific time segments(which could be changed according to how much amplitude change there is in the next segment), look for the average amplitude action there are, and then giving that as a label for each segment... that would give us a matrix right? (probably amplitude action for one axis and perhaps some other derivative as another axis) Then, that matrix, or a 2D plane could be used in conjunction with another axis.. say the cut off of a simple resonant lowpass filter.. that would give us a 3 axis 3 dimensional "space" sort of. Then... this is the punch line. You could end up with a 3 dimensional vector synthesizer with FFT/granular characteristics! Wow! Now that's a monster I'd like to work with.
Oh well, I'm getting a bit personal with this FFT thing.. but, you get the idea.. it can open new horizons, and with a little developement from creamware, it seems pretty realistic to me. errr... maybe JUST to me.

Ah ha! How about colaborations with IRCAM or ICMC people? They've got some really crazy/strange ideas.. I just read a '98 paper about having artificial lips made of latex blow a trumpet to see how real it can get... physical modelling? I don't know but it sure is a wacky synthesizer! haha:lol: So I think we could use a little of their craziness.
<font size=-1>[ This Message was edited by: kensuguro on 2001-10-08 15:56 ]</font>
That actually might be interesting, even if it did generate latency -- perhaps with the newer 8x8 modular, you could simply route outputs to ASIO, effect it with whatever ASIO/VST instrument you had in mind, and return it back to your modular patch for further manipulation, MUAHAHAOn 2001-10-06 16:50, algorhythm wrote:what do you mean? I doubt it is possible, whatever you mean. Pulsar only runs on SHARCs . . .On 2001-10-06 13:23, yade wrote:
A module which can be interfaced with non sharc-dsp (intel/motorola) software?

-
- Posts: 1139
- Joined: Tue Apr 03, 2001 4:00 pm
- Location: Tennessee, USA
- Contact:
not sure what you have in mind, but run something in one or more of the EIGHT the audio ins! - TV, Radio, a Mic out of the window, the output of "Dope" or the "Filther" a HW synth or another pulsar synth, etc. etc. Process it beyond recognition in the Modular = ambient or industrial sounds depending on extent and type of processing and original input signal . . .Spirit wrote:
How about some different sound sources - not just standard oscillators, but something a bit more environmental and uncontrolled. I'm not sure exactly what, but I'm sure you know what I mean - something for industrial and ambient.