No, seriously--what is it?
What exactly is granular synthesis?
-
- KVRist
- 367 posts since 29 Dec, 2004 from east coast, usa
In layman's terms....I'm looking for an explanation of granular synthesis. Sounds interesting. Does it build up a lot of sand inside your computer?
No, seriously--what is it?
No, seriously--what is it?
-
- KVRAF
- 9217 posts since 23 Jul, 2002 from Pequot Lakes, MN
Granular synthesis works by breaking down a sample into equal length sections called grains. Each grain can have a specified pitch and/or duration. If you keep the duration intact while changing the pitch, you get pitch shifting without any "chipmunk" artifacts. If you keep the pitch intact while changing the duration, you get different formants emphasized. Do both and you have more fun than you would with Pamela Anderson (well...almost 
)
ew
)ew
A spectral heretic...
-
- KVRAF
- 2460 posts since 3 Oct, 2002 from SF CA USA NA Earth
It's a method of building up a sound out of "grains", usually derived from a sampled sound. Basically you take a sampled sound, apply a very short volume envelope to it, and play it to get a "grain"; you play dozens of overlapping grains to get the final sound. Maybe the grains come from different places in the sample. Maybe you vary the length of the grain or the sharpness of the ends of its envelope. Maybe you process the grains further by putting them through a filter. Maybe you pulse the density of the grains. Maybe you take some of the output sound and mix it into the source sample. Maybe you stretch or squish the grain in time to transpose it in pitch.trappist wrote:In layman's terms....I'm looking for an explanation of granular synthesis. Sounds interesting. Does it build up a lot of sand inside your computer?
No, seriously--what is it?
KTgranulator and Audiomulch both showcase a lot of the potential of granular synthesis.
Don't do it my way.
-
- KVRAF
- 3364 posts since 16 Feb, 2004 from atop a katamari
normally you treat sound as one big long ribbon. you start at one end and play thru it, til you decide to stop or it runs out.
granular techniques become most relevant when you try to consider alternative ways of travelling along this ribbon.
sound is the result of pressure variations. in our electronic version, sound is a changing number, representing a wave. as we travel along this 'ribbon', the wave appears to wiggle past us. the wiggles determine the sound etc..
if we were to slow down, the wiggles would be seen as 'slower wiggles', which to our ears makes a lower pitched sound. if we were to stop, these wiggles wouldn't happen; we'd be sat upon a single unmoving point.
how would it be possible to stay still on the ribbon, whilst still hearing the sound at that point? well- to have sound, we have to have movement. we have to expand our 'point' to cover a short piece of the sound. if we continually 'scrub' forwards thru that single little bit of the ribbon, we might be able to have some kind of idea what the sound is like at that point, without moving along the ribbon any real distance.
what this is doing is breaking the sound into 'grains'. at its most basic, that would be this one repeating grain. you can get more complex tho, and say "at this point, we'll play 40 little pieces every second, and each of those little pieces will fade in and fade out, and they'll overlap and it will sound lovely". now there is a stream of these grains, and they are all showing roughly the frequency content of the wave at that point, without having to traverse the whole wave.
whilst this is happening, the point along the ribbon/wave can be changed; if we were to move the point forwards gradually, the frequency content constructed by the rapidly played back grains would resemble that of the changing wave, but it would change at a rate determined by the movement speed; this is often used a crude way of timestretching audio.
basically, granular systems use tiny little 'grains' of sound; very short (enveloped) bursts of a wave, played back at a rate so that they blend together, and allowing for the playback of spectral content without necessarily traversing the wave proper. [sorry for the silly late-night playschool first section of the wave, i've been in work and it's bed time!]
granular techniques become most relevant when you try to consider alternative ways of travelling along this ribbon.
sound is the result of pressure variations. in our electronic version, sound is a changing number, representing a wave. as we travel along this 'ribbon', the wave appears to wiggle past us. the wiggles determine the sound etc..
if we were to slow down, the wiggles would be seen as 'slower wiggles', which to our ears makes a lower pitched sound. if we were to stop, these wiggles wouldn't happen; we'd be sat upon a single unmoving point.
how would it be possible to stay still on the ribbon, whilst still hearing the sound at that point? well- to have sound, we have to have movement. we have to expand our 'point' to cover a short piece of the sound. if we continually 'scrub' forwards thru that single little bit of the ribbon, we might be able to have some kind of idea what the sound is like at that point, without moving along the ribbon any real distance.
what this is doing is breaking the sound into 'grains'. at its most basic, that would be this one repeating grain. you can get more complex tho, and say "at this point, we'll play 40 little pieces every second, and each of those little pieces will fade in and fade out, and they'll overlap and it will sound lovely". now there is a stream of these grains, and they are all showing roughly the frequency content of the wave at that point, without having to traverse the whole wave.
whilst this is happening, the point along the ribbon/wave can be changed; if we were to move the point forwards gradually, the frequency content constructed by the rapidly played back grains would resemble that of the changing wave, but it would change at a rate determined by the movement speed; this is often used a crude way of timestretching audio.
basically, granular systems use tiny little 'grains' of sound; very short (enveloped) bursts of a wave, played back at a rate so that they blend together, and allowing for the playback of spectral content without necessarily traversing the wave proper. [sorry for the silly late-night playschool first section of the wave, i've been in work and it's bed time!]
Kick, punch, it's all in the mind.
-
- KVRAF
- 2321 posts since 23 Mar, 2004 from Two lower than LS6
as haydxn said 
Basically, granular synthesis is one of the firsr purely computer based forms of synthesis.
The reason being , is that to do it otherwise would take many hours of work for theoretical results (lets chop up 30 mins of tape into 5ms chunks!!)
granular synthesis is a very old form of synthesis (see papers by Xenaxis and Stockhausen) but only with computers has it become a possibility.
IMHO it is the most exciting form of sythesis as it allows you to manipulate the sound any way you like -very powerful
Basically, granular synthesis is one of the firsr purely computer based forms of synthesis.
The reason being , is that to do it otherwise would take many hours of work for theoretical results (lets chop up 30 mins of tape into 5ms chunks!!)
granular synthesis is a very old form of synthesis (see papers by Xenaxis and Stockhausen) but only with computers has it become a possibility.
IMHO it is the most exciting form of sythesis as it allows you to manipulate the sound any way you like -very powerful
Phil
"The fool who persists in his folly will become wise" - William Blake
*No more band for me* | **My Host**
"The fool who persists in his folly will become wise" - William Blake
*No more band for me* | **My Host**
-
- KVRAF
- 1927 posts since 30 Oct, 2003 from Frolicking in Dirac's Ocean
-
- Skunk Mod
- 21249 posts since 10 Jun, 2004 from Pony Pasture
I'll just add that granular resynthesis is possible without computers... just awfully cumbersome.
When I took electronic music classes in the early 1970s, one exercise consisted of recording some sound -- <GrandpaSimpson> we used "magnetic tape" for this back in ye olden dayes </GrandpaSimpson> :-D -- then chopping the tape into bits of random (bot short) lengths, mixing them thoroughly, and splicing them alllll back together into a new sequence. The sound was definitely changed! Since we used mono tape and flipped some pieces end for end, some parts of the sound were even reversed.
A tedious exercise using mylar and razor blades. Using Reaktor and Grainstates, or Ableton Live and its granular delay unit, gives far more power... and lots more options than the old splicing block ever did.
Ghawds above and below, things are so much better now, with computer audio and soft synths and such! Love it. :-)
When I took electronic music classes in the early 1970s, one exercise consisted of recording some sound -- <GrandpaSimpson> we used "magnetic tape" for this back in ye olden dayes </GrandpaSimpson> :-D -- then chopping the tape into bits of random (bot short) lengths, mixing them thoroughly, and splicing them alllll back together into a new sequence. The sound was definitely changed! Since we used mono tape and flipped some pieces end for end, some parts of the sound were even reversed.
A tedious exercise using mylar and razor blades. Using Reaktor and Grainstates, or Ableton Live and its granular delay unit, gives far more power... and lots more options than the old splicing block ever did.
Ghawds above and below, things are so much better now, with computer audio and soft synths and such! Love it. :-)
-
- KVRAF
- 2321 posts since 23 Mar, 2004 from Two lower than LS6
No kidding?? I always thought the possibility was there but the means wasnt. Way to go Meffy!!
I suppose its what delia derbyshire did for the bass line to the dr who theme, after a fashion
I suppose its what delia derbyshire did for the bass line to the dr who theme, after a fashion
Phil
"The fool who persists in his folly will become wise" - William Blake
*No more band for me* | **My Host**
"The fool who persists in his folly will become wise" - William Blake
*No more band for me* | **My Host**
-
- KVRAF
- 1927 posts since 30 Oct, 2003 from Frolicking in Dirac's Ocean
I think Xenakis was one of the early pioneers and he used to cut analog tape into tiny little quanta. There are many things that suck about living in today's world but I am so grateful to be living in a time where such immense audio infinitude is made available to our desktops.
-
- KVRAF
- 3588 posts since 13 May, 2004 from montreal
True... thousands of tiny splices... I went down that road myself in the 80's - sometimes it took weeks to edit just a few minutes of sound. Much easier now...mayan wrote:I think Xenakis was one of the early pioneers and he used to cut analog tape into tiny little quanta.
-
- addled muppet weed
- 111304 posts since 26 Jan, 2003 from through the looking glass
f**k,i thought it was the sand 
