naulen last edited by
In the IR loader there is a knobs called "Length" what's the function with that? In zero position it is a weak signal and on full there is more power.
Hopefully someone with more grasp on DSP can answer, but in the mean time, it seems like the shorter the filter length, the less low end accuracy you're going to get.
At the other end, on the capture side, if you are doing the capture in the analog world, the longer the test signal you use, the better the signal to noise ratio you get. With the types of captures I used to do, this meant that the impulse file would be longer as well, and if you shorten it past a certain point, you lose the low end accuracy as above.
And I say accuracy here, because it isn't necessarily going to mean less low end, it can be more, its just "wrong" low end
From Matt Hill at Liquidsonics
"I don’t have a link for you, but it’s reasonable to assume that if you truncate an IR to the point that it can’t contain a full cycle of a given frequency because it’s not long enough, you’ll severely compromise the response of an IR below that frequency.
So for example at 44.1kHz a 1 Hz wave needs 44100 samples, 10 Hz needs 4410 samples, 20 Hz 2205 samples. And so on the maths is quite simple."
Tannhauser last edited by Tannhauser
This got a bit of attention at the Fractal forum a few years ago, and one reason FA doubled the size of their IR files. Then there's the wav crowd, who seem to advocate the length should be....infinite.