Hey, I'm Hans

Christmas break 2025. I was mixing vocals that sounded harsh no matter what I threw at them. Tried every de-esser in my plugin folder—same story every time: either it missed the harshness, or it killed the life in the voice.

Ended up zooming in to 5ms resolution and drawing gain automation by hand around every sibilant. Twenty minutes per track. I still had eleven more songs to mix.

So I closed the DAW and started building. Three weeks later I had something that actually worked. Now it's the first plugin I load on every vocal session.

Norwegians value precision. I jump into icy water to sharpen my focus—because a clear mind leads to control, and control leads to precision.

Traditional de-essers take that control away. They rely on rule-based algorithms that decide for you, and they rarely behave exactly the way you want. The result? You fight the processor—or turn it off entirely.

That's why most of us end up manually de-essing vocals with clip gain or volume automation. It's slower, but it's the only way to shape sibilance with absolute intent.

SoftVox is built for that reality. It uses ML-trained detection to find exactly where the harshness lives—then gives you precise control over how much to reduce it. The detection is smart. The decisions are yours.