Alright, Granular synthesis is much more modern (and complex!) than the other two synthesis types I mentioned...
Addative synthesis has been used in pipe organs for centuries, and Subtractive synthesis was what the first synthesizers (moogs in particular!) utilized (and it is also how we produce many sounds... Our lips act as primitive "filters" on the sound produced by our "oscillator", our vibrating vocal chords! Neat, huh?) :D
Granular Synthesis, however has only been around for 60 years, max. From what I understand, it was invented by a very talented composer named Iannis Xenakis.
He had this crazy idea that he could cut sounds into tiny little pieces called "grains", and then do things with them, to manipulate sound in ways that never had been before. Since this was early on, he had to record something (a violin, for example) with a tape recorder, cut the tape into teeny tiny little pieces, and reassemble them to form all new sounds. This, obviously, took a very very long time, and the results were not perfect, but very innovative.Here is a very early piece using granular synthesis, called Analogique A + B.
Anyways, back into the idea of what Granular synthesis is. Now-a-days, we have computers, much faster and larger than any back when granular synthesis was being developed, so we dont need to fumble around with tape recorders any longer, and we can even use this synthesis in real time, if we wish to do so. Unlike addative or subtractive synthesis, which are very consistent in concept, granular synthesis can be used in an almost infinite variety of ways, the main constant being that the sound is being methodically chopped into pieces called "grains". Past this, these grains can be manipulated, combined, edited, sliced into more grains, looped, effected, shifted in speed or pitch, or any combination of these or other methods... You can see why this is such an interesting field!
Because granular synthesis is so computation-heavy, most hardware synthesizers, even today, do not have it as a function (though some rather expensive ones do). However, granular synthesis can be easily done on a computer via software programs like Audiomulch, Kluppe, Reason, and so forth.
Ok, so that is neat and all, but what are some of the applications of granular synthesis? Well, before granulation existed, the only way to speed up a song would be to increase its "frequency", how fast the speaker plays the song through... The problem with this method is, it makes everything sound high pitched, like Alvin and The Chipmunks! The same goes for slowing down a sound, it could not be done without dropping the pitch to the point of sounding ridiculous. This would certainly not do for most serious modern day music producers, but luckily, granulation allows a simple method to speed up (or slow down) a song WITHOUT changing its pitch! How?
Well, lets say we wanted to make a 2 minute long song take up the space of one minute. The old way to do this would be to simply squish together the entire 2 minutes into the space of one minute, doubling the frequency of all the sounds in the song, and thus, doubling its pitch. However, what if we cut the 2 minute song into millions and millions of equal pieces, cut each of THOSE pieces in half. Half of the data of the song has been lost, resulting in a one minute long song becoming two minutes, but luckily, the pitch has stayed the same! Slowing down a song is the same method, in reverse: cut the song into millions of equally sized pieces, and then space them apart evenly over the time fame you want the song to fill. This allows the frequencies to stay intact throughout the stretch.
If you have ever used the "change tempo" function in the popular music editing program, Audacity, you are performing the exact type of granular synthesis that I just described!
The "change pitch" function works using granular synthesis as well, to change the pitch of a song/sound without changing its length... The millions of pieces the song are cut into are then sped up, the old fashion way, to create a higher pitch without effecting the song's length itself!
There are many other applications for Granulation as well, from small fixes on a damaged sound file, to creating beautiful sound textures and pad sounds out of completely different files, to just fucking with a song for the fun of it! There are nearly infinite applications of granular synthesis, many of which don't quite cover the scope of this tutorial, but if you want to learn more, I suggest checking out http://www.granularsynthesis.com/guide.php, which allows an in depth explanation of many concepts of granular synthesis, as well as links to software, publications, etc.
Let me know if I you have any questions, as it can be an very odd topic at times, but I hope I sufficiently explained the basic idea of it, and maybe stayed interesting throughout :)
Thanks again for reading! :D
