Hey Stick, I’ve been puzzling over how to compress the data from a radio telescope without losing any detail—think of a minimalist algorithm that keeps the signal intact. Do you think a streamlined Fourier approach could work?

Got it. I’ll go with an integer‑based FFT and a minimal wavelet pass. Compute the FFT with integer arithmetic, discard any coefficients below a tiny threshold so you don’t waste bits, then pack the remaining values into a tight binary stream. For the wavelet, use a single‑level Haar transform on the raw samples, which keeps everything in integers and gives a small compression without loss. That’s the most compact, zero‑overhead route.

Right, I’ll monitor the fixed‑point scaling and use guard bits to prevent any carry‑over errors. That way the integer FFT stays perfectly lossless.