Grain size..?

[Sequence2]

I am working with an audio sample in Ableton live 8 and I am trying to figure out what grain size is exactly..
The manual says:
Grain Size provides rough control over the average grain size used. The actual grain size is determined in a signal-dependent manner. for signals with a clear sense of pitch contour, a small grain size works best. Larger grain sizes help to avoid artifacts that can occur when the pitch contour is unclear, but the tradeoff can be audible repetitions.

Nothing really explains grain size here so I was wondering if someone could give me a a simple explanation of what I am actually modulating when I +/- grain size

 

[bandcoach]

Sounds like you are using the granular synth.

Grains are small snippets of audio that are used to create a larger waveform that is not one of the classic subtractive waveforms - sine/triangle/square/pulse/ramp/saw - instead the grain is used to create both the small and large features of the sound.

see this for more insight

 

[krushing]

Grains are small snippets of audio that are used to create a larger waveform that is not one of the classic subtractive waveforms - sine/triangle/square/pulse/ramp/saw - instead the grain is used to create both the small and large features of the sound.

That's pretty much it, although the application here is Live's warping/timestretching system - depending of the warping mode selected, Live chops up the audio in very small pieces in order to stretch it elastically & the grain size controls how small these pieces are. With the tones mode, the grain size is, as stated, dynamical so the setting is an average; the actual size will depend on the characteristics of the audio being treated. With the texture mode, Live uses the grain size as-is instead of changing (around the average size) it in regards to the audio. But basically it just comes down to trying it out and finding what sounds best.

 

[bandcoach]

so it's actually more like wavelets than grains? makes sense sort of.....

 

[hollandturbine]

A grain is an extremely short section of audio and as you apply time stretching each grain gets multiplied, for example if each letter represents a grain, ABCDEFG would become longer as you multiply each grain so you get AABBCCDDEEFFGG or AAAABBBBCCCCDDDDEEEEFFFFGGGG depending on how much stretching you apply. The grain size just determines how long each grain that gets multiplied is.

 

[KoenTanghe]

Well, the above explanations say it all.

I sometimes tend to think of grains as "sound dots" that are combined together to form a new "audio picture", much like the pointillistic painters did at the end of the 19th century by combining a lot of paint dots of various colors and sizes to create an overall image ;-)