News for hackers. Stuff that matters.
Re:Downsampling Time Series for Visual Representation [pdf] ... by hctaw on Tuesday March 09, 2021 @06:15PM (#26405753)
You've just described in very basic terms how linear downsampling works (filter -> decimate). It's orders of magnitude more expensive than what is described in the paper, and a crappy sinc interpolator (equivalent to zeroing FFT bins) is not going to be much improvement over the nonlinear approaches described - either objectively in terms of signal degradation due to ripple effects or subjectively, as the paper has explored.

There are no comments on this comment yet.