Oh, you are talking,about the number of samples in an IR?
The most intuitive answer is trim it at the point where the response file has returned to zero, leaving maybe a bit more, so the convolving can finish. I'm not positive; but it might be prudent to make sure the impulse lasts at least 1 period at the lowest frequency you intend to use (20 mS for 50 Hz, 40 mS for 25 Hz).
The quickest way to find out is to get an IR of, say, a bass speaker cab, then make shortened copies of it of varying lengths and see what happens. If you do, let me know what you find, I'm curious.
Last edited by Philbo King; 01-31-2018 at 03:10 AM.
|