Every de-esser I tried made my vocals sound worse. So I stopped using them.

Started doing it manually instead—clip gain on every harsh S, every plosive, every breath. Volume automation drawn by hand, zooming in until I could see individual samples. Twenty minutes per track. But it worked.

Then I automated it. Same precision, same control. No more drawing curves at 2AM.

When you manually clean vocals with clip gain, you're targeting individual events—not EQ'ing the whole vocal. SoftVox does the same thing: detects sibilants, plosives, and breaths individually, applies precise gain reduction with smooth crossfades, leaves everything else untouched. The vocal stays open and alive—just like when you do it by hand. You just get your evenings back.