New release for all: 1.0.5



I hope Cycling will update the Package Manager version soon, but if you can’t wait, see on the GitHub pages.

1.0.5 release notes

New Objects

  • (buf)SineFeature provides interpolated sinusoidal peak tracking, the first step of the algorithm in the (buf)Sines object

New Features

  • DataSet now can do kNearest brute force search, convenient for one-off queries on moderately-sized datasets where the cost of building the KDTree is not worth it.
  • LabelSet print is now on par with DataSet’s
  • KDTree’s kNearestDist now has an optional (k) parameter

Interface Changes

  • (buf)Sines trackingMethod is now trackMethod for interface consistency
  • Max/PD: parameters with non-fft Maxima, for instance MFCC numCoeffs, will now take the maximum needed at instantiation, in case of conflicting parameter settings.

Fixes and improvements

  • Major memory allocation overhaul, should behave much better for SuperCollider real-time objects (although you may need to increase the server pool size more often)
  • (buf)Loudness true peak is now more efficient and accurate (see below for a caveat)
  • Fixes to all examples in help files and example folders
  • Many typos and small edge-case bugs

Known Bugs

  • (buf)Loudness has some cases where the True Peak and Loudness overshoot by a small amount
  • (buf)Transient(Slice) will emit different values to 1.0.0 and is very slightly different in behaviour between different OSes
  • a few typos and help changes have been found in the release process - these are corrected on tip of dev should you want them

I’m really new to using PD externals, so I’m not sure if I’m overlooking something here, but there appear to be a bunch of externals (.pd_darwin) files not in the download for this.

I can’t get fluid.knnclassifier to instantiate, and checking the download folder there’s a ton that have .pd files, without corresponding .pd_darwin files.

Am I overlooking something here?

Did you follow this guide:


I had not!

That does the trick.

1 Like