Nullpane Audio Difference

What changed between these two bounces? Nullpane subtracts one audio file from another so you can see, hear, and export the residual itself. Bring two versions of the same recording from Files. Check the shared time view, adjust B by integer samples, match a constant gain if appropriate, and optionally invert polarity. The offset suggestion searches channel 1 in the first two seconds within ±4096 samples. Listen and inspect the result before applying it. Switch between A, adjusted B, and the difference. See residual RMS relative to A, overlapping duration, actual waveforms, and measured one-second windows. Select a window and inspect a five-second range. The difference waveform can be displayed at ×8; this is a visual scale, not an export gain. Export three real files: an unscaled 32-bit float PCM WAV containing A minus adjusted B, a CSV of second-wise residual RMS/peaks and source positions, and a printable PDF with settings, sources, waveform and your note. Keep local comparisons with their original source copies, names, adjustments, notes and generated files. Delete a comparison to remove its local sources and exports. iPhone keeps the comparison and listening controls together. iPad places the sources beside a larger difference view. A clearly labeled synthesized audio example lets you learn the workflow; your library starts empty. One upfront purchase includes the complete tool. No account, subscription, ads, analytics, tracking, cloud processing or developer-server upload. Supports audio that iOS can decode, up to 100 MB per file, 0.2–120 seconds, mono or stereo, 8–96 kHz. Both files must have the same sample rate and channel count. Only the overlapping range is compared. Nullpane does not resample, correct clock drift or fractional delay, provide spectral repair, perform blind ABX tests, or decide sound quality or perceptual equivalence. Constant gain matching is least squares, not LUFS. A near-zero result is not an audibility verdict. Float exports preserve the mathematical residual and may exceed full scale. Use an editor that supports float WAV. Playback is attenuated when needed to avoid exceeding full scale; it is not an amplified listening test. Export and keep files separately before deleting local comparisons or uninstalling.