compare
track a is the reference — something properly mastered that you like the sound of. track b is the one you're checking against it. it measures both the same way — loudness, dynamics, spectral balance, tempo, stereo width — then reports how track b differs from the reference and what to consider changing. nothing is uploaded, it all runs in your browser.
tracks
headline
spectrum share of total energy, 20 Hz–20 kHz, relative to each track's own mix
loudness over time short-term (3s) loudness relative to each track's own integrated level, timeline stretched 0–100%
what to consider changing
loudness, true peak and spectral balance use the same measurement code as thelufs normalizer on this site, so numbers agree between the two tools. thresholds for what gets flagged are heuristics tuned for club techno, stated in the text so you can disagree with them — none of this listens to the track.