........................................................................
Sockpuppet........................................................................
SamplesBoiCan you prove that was a fast transform and not a slow one? Why bother with fast anymore with so much computing power anyway? Fast was for the 80's and 16-bit world. Slow for the win these days. Deep thoughts.
........................................................................
SockpuppetSamplesBoi wrote:
Can you prove that was a fast transform and not a slow one? Why bother with fast anymore with so much computing power anyway? Fast was for the 80's and 16-bit world. Slow for the win these days. Deep thoughts.
The app says FFT. The app would never lie.
........................................................................
SamplesBoiSockpuppet wrote:
The app says FFT. The app would never lie.
Best to reverse compile it and study its algorithms and verify. Trust but verify, the man said.
........................................................................
SockpuppetSamplesBoi wrote:
Best to reverse compile it and study its algorithms and verify. Trust but verify, the man said.
I actually don’t know shyt about it.
It looks pretty.
........................................................................
Assclown Mini TacosSamplesBoi wrote:
Can you prove that was a fast transform and not a slow one? Why bother with fast anymore with so much computing power anyway? Fast was for the 80's and 16-bit world. Slow for the win these days. Deep thoughts.
If you are doing visualization then there is no real downside to using an FFT - sure, you need to use a sample blocksize that's a power of 2, but for a real time spectrum display the blocksize is pretty much irrelevant anyway. I know I would use an FFT - not only is it faster (O(n log n) rather than O(n^2)), but it's really easy to implement in-place for simpler memory management.
........................................................................
SockpuppetAssclown Mini Tacos wrote:
If you are doing visualization then there is no real downside to using an FFT - sure, you need to use a sample blocksize that's a power of 2, but for a real time spectrum display the blocksize is pretty much irrelevant anyway. I know I would use an FFT - not only is it faster (O(n log n) rather than O(n^2)), but it's really easy to implement in-place for simpler memory management.
Trevs ruining my pretty music thread.
........................................................................
Previous | First | 1 | Last | Next