I've been wondering about this for some time. Finally found an XLN post on the subject.
2012-08-01 14:10
Why are the Addictive Drums and Addictive Keys libraries so small compared to some other libraries?
Part of the answer is that we use our own format to store the audio data.
All our sounds are recorded, edited and mixed at 24 bit and 88.2 or 96 kHz, followed by conversion to 44.1 kHz using the best algorithms out there.
Finally we store the audio in our own format which uses variable bit depth, which is very efficient on percussive/decaying sounds like drums and pianos.
In the audio world, different compression techniques are often referred to as lossy or lossless.
Lossless - how can you reduce data size and not lose anything? Well, this can be done in many ways. A simple analogy would be the image formats BMP and PNG. In reality they are a bit more complicated than this, but to give you a rough idea:
Consider an all black image with a white dot in the middle. Lets say this image is 50 pixels.
BMP would store the colour value of each pixel, like so:black, black, black, black, black, black, black, black, black, black
black, black, black, black, black, black, black, black, black, black
black, black, black, black,
white, black, black, black, black, black
black, black, black, black, black, black, black, black, black, black
black, black, black, black, black, black, black, black, black, black
PNG would store it as: 24 black. 1 white. 25 black.
Same data, but stored with a lot more efficiency using PNG. The 'Zip' format works in a very similar way.
WAV and AIF files work like BMP - even if the file contains nothing but silence, it is just as big as if it contains a whole mix. So using WAV/AIF to store samples is really wasteful, and we don't like that :)
Virtually LosslessWe call our compression algorithm 'virtually lossless' - the signal to noise ratio is around 140db, equivalent to 24-bit resolution.
It works nothing like mp3 or other similar lossy formats, which introduce artifacts that are often audible.
Rest assured that when we developed this format we did extensive testing using sound engineers and producers with really golden ears to make sure we got it right!