Actually better than just the sin/cos test, if someone wants to run test signals, then try this one: http://mystran.googlepages.com/stereotest.wavmichu wrote: scoobz, can i ask you to make a simple test with uad?
settings: attack / release as fast (as long as it doesn't cause distortion: it's a static test but I didn't feel like making the file three minutes long), ratio to 10 (the more the merrier), and calibrate threshold such that the first beep is about 10dB or so (15 shouldn't hurt) into compression... and don't bother with makeup (don't wanna clip stuff and levels really don't matter as long as they are above noise floor).
Contents of the file? There's some sinewaves.. mid channel moved to left channel (equal power to the mid channel version, tests stereo summing), then mixed with another frequency on purely side channel (sanity check for M/S processing) and finally a few sines of different amplitudes to do a rough map of the compression curve just in case it's not a pure 10:1 compression.
edit (1.4.2009): had adl run the test on UAD, and run the test on the renewed demo period from 1.0.10 for the Glue.. turns out in static test it looks like UAD runs on L/R sum (or average, same thing) and Glue runs on L/R max. No idea if they sum/max/average/whatever before or after side-chains, and frankly I don't personally care... but one would expect UAD to push things into more compression with middle-panned signal that equal power side-panned signal (which could kinda widen things), and glue to do the opposite (which could kinda narrow things); in both cases you have exactly preserved stereo image IFF the amount of the compression remains static, but obviously that won't happen in realistic music, and with UAD reacting less, and the Glue more, to sounds panned to left/right (at least as long as they aren't overpanned to the "side" channel).. you got balance changes depending on how things are panned.
