# HG changeset patch # User Chris Cannam # Date 1443714961 -3600 # Node ID 1c027151f7ecea6a05ba6f293f0b90ec7589a450 # Parent d7c216b6a84ffa4c703f157754faba236ff68b5b Beginnings of a test script diff -r d7c216b6a84f -r 1c027151f7ec .hgignore --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/.hgignore Thu Oct 01 16:56:01 2015 +0100 @@ -0,0 +1,2 @@ +syntax: glob +*~ diff -r d7c216b6a84f -r 1c027151f7ec fft/test.html --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/fft/test.html Thu Oct 01 16:56:01 2015 +0100 @@ -0,0 +1,95 @@ + + +

If 2150 iterations of real-to-complex FFT of size 2048 takes less than 10 seconds, then we may be able to make a high quality real-time phase vocoder (just).

+ +

A phase-vocoder of course must use overlapped +windowed FFT (although you can choose the size, within limits), IFFT, +and cartesian-polar conversion to calculate the phase for the +instantaneous frequency.

+ +

A reasonable estimate of CPU cost for the whole thing is somewhere +around 10x the cost of simple non-overlapping short-time forward +Fourier transforms across the signal.

+ +

2150 iterations corresponds to 100 seconds of audio non-overlapped at +44.1kHz, so if that takes less than 10 second, then in theory we might +be OK.

+ 
+
+
+
+
+
+ +