I’m running the flucoma plotter-5-completed.scd on my 9 minute wav file. I’ve increased the memory s.options.memSize=8192*32;
When it is doing the fa.doAdjaceentPairs, the s.sync seems to take a fairly long time without resolving. I can repeatedly evaluate the code at that point and eventually the ~analyses gets printed. The 2D plot of this seems to consist of shortened clips of the track. On the console I get this QabstractSocket::connectToHost() called when already looking up or connecting/connected to "127.0.0.1"
The question is, is the s.sync supposed to take that long? What can i do to get it to finish faster? Is my observation that calling this analysis part of the code multiple times causes at least something to terminate and the analysis to finish?
I can run audiostellar or borderlands on this without much issue.
I’ve changed the period for s.sync from i%100 to i%5 and it completes, but it still seems as if the slices are not long enough to cover 9 minutes. Is it supposed to be this way?
The analysis does take a bit of time. Can you be more specific about how long it is taking?
That is right. The code is slicing up the original audio file using a FluCoMa slicer, so each dot on the plot is one of those slices.
Can you post the code you’re using and perhaps also the wav file? The more info I have the better!
The code is exactly the same as plotter-5-complete.scd. It doesn’t finish really. But I’ve shortened the number of slices before calling s.sync to i%15 during the analysis, and that seems to work. I’ve replaced FluidBufOnsetslice with FluidBufNoveltyslice though. What i meant about the slices is, shouldn’t the slice start from slice point n to slice point n+1. So a 68 slice point output for a 9 minute track should give 8 seconds per point on average. But it doesn’t seem to be this way. A slice seems to last like 6 seconds or less. It could just be my impression of it.
There could be a few different version of that floating around. Did it come with when you downloaded FluCoMa or did you get it from the FluCoMa website or somewhere else?
Yes, this is what it does.
Some slices might be 6 seconds or less. It depends on what’s in the file. Both FluidBufOnsetSlide and FluidBufNoveltySlice are analyzing the file for slice points, so the duration between slice points can’t be known without knowing what’s in the file.
I got the file from the youtube link. So the entire track is represented on screen as points? I probably should check by running the play slices sequentially.
Yes, the entire track will be represented, except for potentially some of the buffer on the outer limits. Consider a buffer that was analyzed and found to have 4 slice points:
|------|--------------|----------|------------|-----------|
beginning 1 2 3 4 end
The FluCoMa slicer is giong to return the slice points just at 1, 2, 3, and 4, which will make for 3 slices to analyze:
- from slice point 1 to 2
- from slice point 2 to 3
- from slice point 3 to 4
Therefore whatever audio is from the beginning to slice point 1 and from slice point 4 to the end won’t be represented on the plot (because they’re not being analyzed).
If you do want these sections of the buffer to be analyzed, make sure to add a 0 to the beginning and the total number of samples in the buffer to the end of your array fa before running doAdjacentPairs on line 47.