AXELL SIGNATURE · NEURAL DE-ESSER + WARMTH
Axell Velvet
Silk on your esses. Warmth everywhere else.
WINDOWS · LINUX · VST3 · NEURAL DE-ESSER · TAPE / CONSOLE / TUBE · UP TO 16x
A de-esser that recognises an S instead of listening for loud treble. Trained on 111 voices, it cuts the harsh esses and leaves everything else alone - and three analog warmth models give the voice its glow, with S-GUARD keeping the warmth off the esses.
01 WHY IT'S DIFFERENT
Five reasons, all measured.
30-80x gentler on your vowels
than a respected sibilance-aware de-esser, at equal harsh-ess removal, on 15 voices it never trained on.
Trained on 111 voices
and it hears accents it has never met: F1 0.82 on 12 unseen accents.
Zero latency when you record
the LIVE de-esser is causal - with HEAT off, Velvet adds no latency at all.
Warmth that stays clean
aliasing at equal saturation: tape -108, console -129, tube -113 dBc - certified UADx plugins measured -70 to -96.
3-9 % lighter in 3.0.4
byte-for-byte the same sound.
Every figure comes from the plugin's own test suites and published benchmarks — the full measurements, the method and every version's history are on GitHub.
02 A CLOSER LOOK
Every screen, every mode.
HEAR IT
Before and after.
The same excerpt, the original and through Axell Velvet with a factory preset. Press play, then switch — both play in sync, so you hear only the plugin. Headphones recommended.
Drums
Tape warmth on the kit - thicker, rounder, no harsh edge.
Keys
Tube bloom on the chords.
Sung vocal
The esses smoothed by the neural de-esser, air and warmth everywhere else.
03 GET IT
Windows first. Linux too.
Free and open source. Copy one folder into your VST3 folder — or install the whole suite in one go with the Axell Signature Suite installer, and let the Dojo keep everything up to date.
Runs in any VST3 host — FL Studio, Ableton Live, Cubase, Studio One, Bitwig, REAPER
THE SUITE