Day 30

Hours worked/Total: 8/96.5

Day Recap

Started Fourier Transform experiments for spectral analysis. Got some low-frequency baselines, but results feel off. Struggled with filtering and visualization, and realized rectangles look ugly as hell compared to the smooth sine waves in pro DJ softwares.

What I Did

• Extracted low-frequency data—seems wrong, initial frequencies overpower everything.
  
  • Rendered low-frequency waveform to visually debug, still doesn’t map right.
• Bars look blocky and unrefined—real software uses smooth sine-like interpolations.
• Tried to split into 3 bands with filters (low, mid, high).
  • Not working. Everything registers as high-pass frequencies, meaning filtering logic is broken.
• Realized I Need a Proper Fourier Approach:
  • Need to correctly map frequency bins back to time domain.
  • Might have to rethink rendering method—probably need smoother interpolation instead of basic bars.

Remarks

Fourier transforms are pure pain if you don’t know what you’re doing. This might take a while…

Next Goals

I am burnt our and will need few days to recoup.