Experimenting with Sound Design - Experimental & Generative
Archive · 201026 old posts
From the original SynthMaker forum (2004–2014), with its members’ names and dates.
attic ·
I would like to start a group project for a couple of items. I'm interested in re-purposing information for Sound Design.
I would like to get some feedback on how you would see this done. I'm looking for the most musical sounding results or at least unique sounds.
1. Translating the human Genome to pitch or frequency. This would give us a unique way to generate semi random midi note sequences and other such uses. As found here http://www.gutenberg.org/browse/authors/h#a856
The text files are large so I would just parse tiny chunks of it, as reading 250.60 MB for Chromosome Number 01 would be to much for Synthmaker. The characters consist of 5 letters ACGT and N. At different intervals throughout the file you find long sequences of N's.
The way I see it being done is for the A character we could assign the pitch for the notes C E and G
The character C could be the notes G B D and E (B and E could be either or depending on a variable you could use the N sequence as a variable to choose what the note generated ended up being. ..... and so on.
It really doesn't matter how you divide up the 12 notes in the scale as long as there is one or more for all the characters so ACGT fits 3 times into 12 notes and you could use the N's as some kind of variable for triggering other notes its really up to you. I'm curious what would end up being the most musical so any thoughts or examples on this would be appreciated.
2. It would also be interesting to create an osm that could translate chunks of the human genome to a wavetable (IE text to wav) for feeding your synths generated random sounds.
3. Another thing that would be interesting is to create an osm that converted the Mandelbrot set, the Fibonacci numbers or the golden ratio to a pulse clock for creating delays. http://en.wikipedia.org/wiki/Golden_ratio
4. Parsing a bitmaps pixels to a wav file in whatever fashion you see fit, but try for interesting sounds.
I will post what I build to this thread and it will probably take me some time as I'm new to this and don't have a lot of free time for Synthmaker right now.
All suggestion's and osm examples of how you see any of these ideas working would be appreciated. Feel free to add your own experimental or generative ideas to this thread.
Here is an Osm with a tiny chunk of genome in a text module to get you experimenting. This is better than downloading the whole 250.60 MB.
oddson ·
Seriously? 😕
Wadaya smoking? 😉
Unless you focus on 'junk' or on known allele sites your going to get the same standard strings if you go looking in the same places each time...
And if you can get enough variation to be interesting; what makes you think it will be more musical than any other quasi-random patterns?
Unless you can interpret meaning form the sequences you might as well use 'real' (i.e. pseudo) random numbers.
attic ·
oddson wrote: Seriously? 😕
Wadaya smoking? 😉
Unless you focus on 'junk' or on known allele sites your going to get the same standard strings if you go looking in the same places each time...
And if you can get enough variation to be interesting; what makes you think it will be more musical than any other quasi-random patterns?
Unless you can interpret meaning form the sequences you might as well use 'real' (i.e. pseudo) random numbers.
Hi Oddson
I gave up smoking along time ago, so the answer is nothing at all. My reason for having interest in this would be that I like the idea of starting with something that already has complex meaning, even if I don't understand the underlying content. I know these things have been done before and I appreciate your help from around the forum, but I was hoping for something maybe a little more constructive than the statement "what are you smoking".
I can understand if you have zero interest in this kind of application. I think the delay based on the Fibonacci sequence would be musical, also I think that if you used the genome seq for something as simple as a forward and back scratch Dj thing it would be musical (triggering direction repetition and maybe length on a wav sample). I could come up with a lot of ways to interpret these patterns into usable musical data. What would you suggest other than random as a starting point? Pseudo random number interpretation as you put it, gives you a starting point that already has attached meaning, much like a poet might start a poem with some random words like cloud blue and sky. There is also the Kitsch value in that you can tell people my Vst generates its output from the Human Genome.
And if you can get enough variation to be interesting
The human Genome should be enough data for a lot of variation, seriously its very large.
what makes you think it will be more musical than any other quasi-random patterns?
Well one of the number sequences I mentioned every spiral in nature is based on and I'm sure it has some great potential for delays and trance style sequences. Those where just suggestions how about you insert your own quasi-random pattern or even random if you like and lets see how musical it could be.
Unless you focus on 'junk' or on known allele sites your going to get the same standard strings if you go looking in the same places each time...
This was interesting and I do find this statement useful so thanks for that. Alleles are alternative DNA sequences at the same physical locus, which may or may not result in different phenotypic traits. http://en.wikipedia.org/wiki/Allele
My interest in this sort of thing started when I bought a Korg Karma keyboard. I now use a Korg M3 and it has Karma built in but just isn't as good in the generative department as the original Karma keyboard so I am looking to build some Vsts with generative capabilities. You could spin the knobs for hours and get great new sequences. I intend to create some stuff along these lines useful or not. I would like to use it for generating random arps and triggering chord sequences. I was hoping there might be other people around here with similar interests that might have some input. I am intending to build a small library of files as a starting point for things of this nature. If the projects section is the wrong place for this, feel free to move the post.
oddson ·
attic wrote: ...I was hoping for something maybe a little more constructive than the statement "what are you smoking".
Sorry... but I did give that a winky face.
But I'd say look for pattern rather than randomness if you're looking for musicality.
Maybe in gene sequences - assign the sequences that code for amino acids to specific musical segments... something like that.
Good luck.
MichaelBenjamin ·
i think thats a very worthwhile direction trying to incorporate algorithmic or even just general abstract or physical structures, formulas and data into music dsp to borderline physical modelling into something more abstract. in case of the human genome it would be a seemingly quite random abcd (4states?) pattern. as far as mapping that to pitch and note length like in algorithmic sequencers, this for my tastes does not produce very interesting musical results because the machine is not able to compose (unless you can somehow put drama and emotion into a formula), just to interpret and perform composition/code thats written by a human. so i would look to put that data rather into the articulation/waveform whatever, at a place that defines the playback state of the instrument.
I gave up smoking along time ago, so the answer is nothing at all.
Comeon oddson, he is from San Francisco, like you had to ask that question! 😉 :P
I have a friend who chatted with a quantum physics prof about the parallels with musical instruments:
JW: It was that everything you said about the harmonics generated from the path interferences which you were studying seem to apply to the acoustics of music as well.
IW: There are some very similar analogues there, having to do with the fact that the underlying entities in all of these physical theories is a field, which is a weird sort of concept. Once you have that, interferences and oscillations play a big role in quantum mechanics as they do in music.
JW: When you have a stopped pipe, one gets that same effect of the odd harmonics as is observed in perturbative harmonic generation in quantum systems based on n-photon absorption.
IW: As in an organ pipe.
JW: Or a clarinet .. the 1st, 3rd, 5th, 7th, etc. harmonic generation being especially strong because the second normal mode is 4/3 the length of the pipe.
I personally think there is music everywere, & a master score that you can choose to either harmonize with or be out of tune/step with.
This is an interesting project, you might want to help out with the chord-tool, since we are building a text/GUI interface for building chords.
MichaelBenjamin ·
infuzion wrote: I personally think there is music everywere, & a master score that you can choose to either harmonize with or be out of tune/step with.
i like that, maybe this is also what the free will boils down to.
Acrobat ·
4. Parsing a bitmaps pixels to a wav file in whatever fashion you see fit, but try for interesting sounds.
This already exist and by a decade or more, it's called Metasynth by U&I softwares, runs exclusively on the Mac.
I worked a lot with it in the past, even arranged a lesson at university based on the software as it was a very promising piece of code. I suspect the authors didn't add any new feature to it recently, but it's still very interesting.
http://www.uisoftware.com/PAGES/uimain.html
attic ·
Sorry... but I did give that a winky face.
Your right .. You did do the winky face and your thoughts on the subject are good.
look for pattern rather than randomness
Agreed that's good advice.
I would look to put that data rather into the articulation/waveform whatever
Yes that's a good call the Korg Karma did this.
Comeon oddson, he is from San Francisco, like you had to ask that question!
This area is known for a few things. I'm north of the city by 30 miles not quit Humboldt county also my birthday is 4,20 same as hitlers amongst other things. I haven't celebrated my Birthday in that manner for many years now. I do know some Berkly music professors that just love the local products and are all about the Generative music thing.
I personally think there is music everywere, & a master score that you can choose to either harmonize with or be out of tune/step with.
Maybe this is also what the free will boils down to
I carry a pocket recorder with me everywhere so I can grab samples and the random sound events around me, its the soundtrack to my life. I use a little olympus usb recorder it needs to be a new Yamaha so the recordings could be better. I did a check on how many samples I collected last year and have almost 2000 of them. Ive been doing this for about five years now so its time to create a romplayer. It takes alot of time to tag cut and normalize them, but I think when I release a vst with them, it will be unique and have been worth it.
This is an interesting project, you might want to help out with the chord-tool, since we are building a text/GUI interface for building chords.
I noticed that project yesterday and the fact that you are building a text/Gui. This would be applicable to some of the things I have in mind. I am however still struggling with some of the basics of Synthmaker and have only owned it for a week. I will advance fast and am very efficient in Synthedit. Having learned Synthedit to the core is helpful in grasping Synthmaker quickly but there are many differences. I almost went straight to C++ but Synthmaker was to tempting. If I see an opportunity that I may offer some help I will jump in. I do wonder if i could create a sieve for sorting the genome data that would trigger those chords in a usefull manner.
This already exist and by a decade or more, it's called Metasynth by U&I
Yeah I wish I had a Mac for that that one piece of software. I want it on the pc in a vst.
I found this software interesting and well beyond my capabilities of recreating, but that doesn't mean I wont get there at some point. I find this software more fascinating than the new Melodyne algo's, I want what he's smoking. It is expensive software, http://www.cognitone.com/support/tutorials/index.stml?p=1. Also another expensive product that I would love to get my hands on is this http://www.symbolicsound.com/cgi-bin/bin/view/Company/WebHome. The Kyma system can do the things I mentioned but it is out of my price range so I need to build my own.
The first thing I did when Oddson posted a guitar tuner here Examples of a guitar tuner. was put a midi output on it so it triggered notes on my M3 and then ran a recording of crickets through it to see what it would sound like. If you want the crickets wav file let me know. The divide between Science and Art is thin, only separated by curiosity and creative use on either side. I don't do drugs but I can be somewhat Odd at times, its just my curiosity. In case you want to try this its all Oddson's files and not musical at all but I attached the osm for any curious individuals around here.
Edit ... It wasn't crickets it was frogs a whole field of them going apeshit and by the way I think they are fairly musical. Those are San Francisco frogs so they have got to be smoking something. Do what you will with the samples.
MegaHurtz ·
I would be more interested is a system that evolves around the user.
Where im thinking the human genome is a very expensive way of making random algorithms
that nobody does understands anyway.
Ive always considered it all to be a balance between function and weight.
Big things die off if smaller ones offer the same results.
Fibbonachi is nice I think becouse theres a logic in it, and is a very natural series.
what you said about delays I think it can work,
although mapping it onto a musical space is nice too, mapping it onto four beats for instance.
attic ·
MegaHurtz wrote: Fibbonachi is nice I think becouse theres a logic in it, and is a very natural series.
what you said about delays I think it can work,
although mapping it onto a musical space is nice too, mapping it onto four beats for instance.
Ok this is one I would like to try where would you start on this? I suppose you need to set up bpm sync and an array, it would be nice to make it a dual delay so we could offset it against itself or a separate running Fibonacci sequence. Any thoughts on this? Depending on how short the delay would be it could end up sounding very granular.
MegaHurtz ·
I was yust wiki-ing the closed form but it looks kinda demanding on the cpu.
Im thinking storing these into an array would be a bit less expensive.
Excuse me for the breef response, but I have to get cracking.
have to get some stuff done.. 😍
attic ·
Quick question, there isn't any problems with having a sample rate module in multiple spots in your schematic is there?
Tzarls ·
I don´t think so, but I think it´s better to have just one attached to a wireless transmitter and then setup a wireless receiver whenever to want to have a look at the current samplerate.
attic ·
I think it´s better to have just one attached to a wireless transmitter and then setup a wireless receiver whenever to want to have a look at the current samplerate.
Thanks for that. 🙂
oddson ·
Tzarls wrote: I don´t think so, but I think it´s better to have just one attached to a wireless transmitter and then setup a wireless receiver whenever to want to have a look at the current samplerate.
I recall suggestions that the wireless overhead was the same or more as multiple instances of the SR primitive.
Andrew J ·
Cool idea about the scratching - and I liked the apeshit frogs :P
attic ·
I recall suggestions that the wireless overhead was the same or more as multiple instances of the SR primitive.
Thanks Oddson, I will have to check cpu usage. When I get further along on this delay I will post some files.
About those frogs.
I was walking past a field after it had rained, those frogs can be pretty loud. The olympus recording doesn't do it justice, they do that once a year for a month straight. If I lived next to that field I would have to move to a new home.
sunsynth ·
btw: Here in our region we also have Frog-months 😈
Even walls and/or tunnels are build 4 them so they can cross the streets!
It's really dangerous to drive across this 'frog-mud' 😝
attic ·
Here in our region we also have Frog-months
You know you picked the wrong project to work on when the frog topic ends up being more interesting then the Synthmaker topic. 😉 They have that kind of thing in the southern states! I was driving in a cloud burst and thousands of them where crossing the road in one spot, they are very slick your car slides like its on ice. San Francisco is well known for its Pita members I hope they don't burn a cross on my lawn because I just posted that. 😈
I just checked out http://www.justmix.biz nice looking site by the way.
Tzarls ·
oddson wrote:
Tzarls wrote: I don´t think so, but I think it´s better to have just one attached to a wireless transmitter and then setup a wireless receiver whenever to want to have a look at the current samplerate.
I recall suggestions that the wireless overhead was the same or more as multiple instances of the SR primitive.
AFAIK thereshouldn´t be a wireless overhead, since internally wireless are treated as wired by SM. OTOH, the SR primitive must be calling the samplerate function, and ,in theory at least, function calls do add overhead, so, in theory again, it´s better to call that function once and store it´s returned value than to call it repeatedly.
oddson ·
Tzarls wrote:
oddson wrote:
Tzarls wrote: I don´t think so, but I think it´s better to have just one attached to a wireless transmitter and then setup a wireless receiver whenever to want to have a look at the current samplerate.
I recall suggestions that the wireless overhead was the same or more as multiple instances of the SR primitive.
AFAIK thereshouldn´t be a wireless overhead, since internally wireless are treated as wired by SM. OTOH, the SR primitive must be calling the samplerate function, and ,in theory at least, function calls do add overhead, so, in theory again, it´s better to call that function once and store it´s returned value than to call it repeatedly.
I thought so too... here's the thread I was recalling.
Parametric EQ
trogluddite ·
Attic, you have some interesting ideas there.
After years of obsessing over formal structures and trying to 'perfect' my music in every possible way (which is of course, a wild goose chase), I've gotten a lot more interested in using a bit of controlled randomness and serendipity.
It amazes me that with the incredible technology we have at our fingertips, just how much music is trammelled into 'genres' and the 'tyranny of four' (four beats to a bar, 16 bar long verses, middle 8 etc. etc.). I've been experimenting a lot recently with polyrhythms. I'm amazed at how complex the music this makes can sound - having many very simple musical phrases but none of them sychronised to the same time signature.
This is a very common idea in much Asian and Far-Eastern music; yet in the west, both 'high brow' and 'popular' music only seem to have ears for having a 'riff' or 'motif' , and a fixed time signature. I'm convinced this is partly because the cut and paste features in our DAWs make life so much easier if you just line everything up together to the nice neat bar lines.
Polyrhythmic music strikes me as being similar to fractals - the component parts and the 'rules' are very basic, yet the result is perceived as complex and non-repetitive, and that's before you even consider adding randomness. It doesn't all have to sound like hippy 'smoking music' or whale noises either - some of the thai and vietnamese music I've been played is more brutal than any hardcore punk band!
Apologies for the blatant plug, but I'm hoping to post up the latest version of my Soopa Loopa project soon. With this you can have several loops running at once - the lengths can all be different, and changed at any time. They can also be offset against each other easily, and even randomly. Parameter moves can also be looped with independent lengths and random offsets. You can grab an older version with manual if you search for the old 'Soopa Loopa' thread here on the forum.
When I've got that finished, I'd love to join you in your project to find the beauty in unpredictable sounds - as I write this, I'm enjoying a wonderful symphony played by several blackbirds, song thrushes, finches, crows etc. with a fantastic robin soloist!
attic ·
Tzarls wrote:
oddson wrote:
Tzarls wrote:I don´t think so, but I think it´s better to have just one attached to a wireless transmitter and then setup a wireless receiver whenever to want to have a look at the current samplerate.
I recall suggestions that the wireless overhead was the same or more as multiple instances of the SR primitive.
AFAIK thereshouldn´t be a wireless overhead, since internally wireless are treated as wired by SM. OTOH, the SR primitive must be calling the samplerate function, and ,in theory at least, function calls do add overhead, so, in theory again, it´s better to call that function once and store it´s returned value than to call it repeatedly.
I thought so too... here's the thread I was recalling.
Ok I appreciate this and just to be clear what I think I'm hearing from you is that this is a single variable shared and readable from anywhere in the code. I also noticed nobody answered Oddson's question on the link he posted Parametric EQ it was this
(What about memory use, does SM allocate memory for each Timing Info component and would it be more than each wireless connection?). Oddson did you ever find the answer on this one?
attic ·
After years of obsessing over formal structures and trying to 'perfect' my music in every possible way (which is of course, a wild goose chase), I've gotten a lot more interested in using a bit of controlled randomness and serendipity.
Yes I so agree with this statement. I couldn't have said it better "Controlled Randomness and serendipity" .
It amazes me that with the incredible technology we have at our fingertips, just how much music is trammelled into 'genres' and the 'tyranny of four' (four beats to a bar, 16 bar long verses, middle 8 etc. etc.). I've been experimenting a lot recently with polyrhythms. I'm amazed at how complex the music this makes can sound - having many very simple musical phrases but none of them sychronised to the same time signature.
This is a very common idea in much Asian and Far-Eastern music; yet in the west, both 'high brow' and 'popular' music only seem to have ears for having a 'riff' or 'motif' , and a fixed time signature. I'm convinced this is partly because the cut and paste features in our DAWs make life so much easier if you just line everything up together to the nice neat bar lines.
I am very interested in polyrhythms also. 🙂
Apologies for the blatant plug, but I'm hoping to post up the latest version of my Soopa Loopa project soon. With this you can have several loops running at once - the lengths can all be different, and changed at any time. They can also be offset against each other easily, and even randomly. Parameter moves can also be looped with independent lengths and random offsets. You can grab an older version with manual if you search for the old 'Soopa Loopa' thread here on the forum.
When I've got that finished, I'd love to join you in your project to find the beauty in unpredictable sounds - as I write this, I'm enjoying a wonderful symphony played by several blackbirds, song thrushes, finches, crows etc. with a fantastic robin soloist!
Plug away I played with that project already and its one of my favorites on this forum. I am looking forward to the Soopa Loopa release and the Parameter functions sound really interesting. I would be honored if you joined me here when you get the time, it sounds as if we have some ideas in common. 😉
AngularMomentum ·
sunsynth wrote: btw: Here in our region we also have Frog-months 😈
Even walls and/or tunnels are build 4 them so they can cross the streets!
It's really dangerous to drive across this 'frog-mud' 😝
We have the "paddentrek", toads instead of frogs, but just as slippery when driving.
Wrote something here? To have a post removed or your name hidden, write to [email protected].
SynthMaker