Most stereo wideners fake it. They shift phase, add delays, or manipulate frequency bands in ways that fall apart on headphones, in mono, or at the mastering stage. Velvet Noise takes a different approach, grounded in acoustic research into how the human brain actually perceives space.
At its core is a technique called velvet-noise decorrelation. Rather than treating the left and right channels as a single image to stretch, it makes the two channels genuinely less similar to each other in a way that mirrors how real acoustic spaces naturally scatter sound. The result is a wider stereo image that holds together in mono, survives the mix, and doesn’t pull the center apart.
A single knob controls how much widening is applied. Full counterclockwise is your original signal, untouched. Clockwise adds width progressively, blending the processed signal in with the dry. Maximum settings apply subtle modulation for extreme separation.
No phase weirdness, no pumping, and minimal latency.