CLEAN S
Take out what gets in the way.
Removes mud, harshness and ringing, then warms what is left. Boomy rooms, spitty esses and muddy low mids come down without the mix losing its body.
IN THE DRAWER · RESONANCE · HARSHNESS · MUD · WARMTH
Auto mixing for creators · Beta soon
PENROSE ULTRA
Your mix, finished.
PENROSE ULTRA is an auto mixer for creators with no time to mix, or no wish to learn. Put it on your master, pick YOUTUBE, PODCAST, SHORTS or STREAM, play your edit through once and press AUTO. Your mix comes out cleaner, fuller and on target for loudness, with the voice lifted above the music when you want it.
AUVST3AAXmacOS · Apple Silicon
IN ONE SIDE / OUT THE OTHER
Three steps
Pick. Play. Press AUTO.
No mixing knowledge needed. Every step below is one click, and every result is measured on your audio.
- 1PICK A PRESET
YOUTUBE, PODCAST, SHORTS or STREAM. Each one sets how much the voice is lifted, how it sits against the music, and the loudness target for that platform.
- 2PLAY IT THROUGH
Play your timeline or bounce it once. PENROSE ULTRA measures the loudness of the whole piece and learns where it gets quiet and where it gets loud.
- 3PRESS AUTO
LOUDNESS turns to land your mix on the target, and LEVELLING, set by your preset, evens out the quiet and loud sections. Play it again to check. Done.
REAL READINGS FROM THE PLUG-IN · YOUTUBE PRESET · TARGET −14 LUFS
Every platform has a loudness it expects. Upload too quiet and you sound weak next to the video before yours. Upload too loud and the platform turns you down and squashes what is left. Set TARGET anywhere from −24 to −9 LUFS (−14 is the default) and AUTO gets you there.
The true peak never passes −1 dBTP, however far AUTO has to go.
The controls
Four knobs, named for what they do.
When you want to go further than AUTO, the knobs say exactly what they change. Each sound knob has an S button that plays only what that knob is doing, so you can hear the change on its own.
ENHANCE S
Give back what makes it finished.
Adds weight, air and presence, then glues the mix together. The step that turns a rough edit into something that sounds produced.
IN THE DRAWER · WEIGHT · AIR · GLUE
VOICE S
Bring the voice to the front.
Lifts the range where speech is understood and eases the music back underneath it, on the finished stereo mix you already have.
FINE CONTROLS · MUSIC ↔ VOICE · BODY ↔ CLARITY · SENSITIVITY
LOUDNESS
Set the level you deliver at.
Raises the mix and holds the ceiling so peaks never clip. AUTO turns this one for you to land on your target.
WITH · TARGET · AUTO · LEVELLING
The sound
Better before you touch a thing.
PENROSE ULTRA's sound was rebuilt from the ground up for this release. It works the way a good mix engineer listens: it treats quiet moments and loud moments differently, instead of putting one fixed EQ over everything.
Turn CLEAN and ENHANCE up and you get all of it at once. No extra controls to learn.
- TIGHT BASSHeld bass notes stay firm and controlled, and the attack still comes through.
- CLEAR LOW MIDSThe muddy range just above the bass is pulled out of the way, so the low end sounds clean, not thick.
- AIR ON THE DETAILThe top end opens up on quiet detail: breaths, room, the tail of a word.
- PRESENCEA presence lift around 4.5 kHz puts voices and lead instruments forward on small speakers.
- SMOOTH LOUD PARTSWhen things get loud, they stay smooth instead of turning harsh.
VOICE
You control the voice to music ratio.
You have the final mix. The music bed is a touch hot, the host is a touch buried, and the separate tracks are long gone. VOICE works on the stereo file you already have.
Turn it up and four bands where speech carries its meaning come forward: body at 500 Hz, clarity at 1 and 2 kHz, presence at 4 kHz. Their depths follow the band-importance weights of the speech intelligibility standard (ANSI S3.5), so the help goes where listeners actually catch words. At the same time two bands where music keeps its energy, 90 Hz and 250 Hz, ease back, clearing the low end that buries a voice.
It follows the talking. When someone speaks, VOICE leans in. When they stop, it lets the music be music.
It changes balance, not loudness. VOICE measures exactly what it added and takes that much back, so turning it moves the voice forward without a jump in level fooling your ears.
VOICE · ILLUSTRATIONLISTENING
How much help
Ease the music, or lift the voice
Which part of the voice
How readily it hears talking
OUTPUT LEVEL CHANGE 0.0 dB · balance moves, loudness stays
MUSIC ↔ VOICEWhich way VOICE works. Toward MUSIC it only eases the music back. Toward VOICE it only brings the voice forward. In the middle it does both.
BODY ↔ CLARITYWhich part of the voice gets the help. Lean toward body for thin voices, or toward clarity for muffled ones.
SENSITIVITYHow readily VOICE decides someone is talking. Raise it for a quiet guest over busy music. Lower it when the music itself keeps waking it up.
S for solo
Hear exactly what each knob is doing.
Press S next to any of the three sound knobs and PENROSE ULTRA plays only what that knob changes.
S on CLEAN plays what CLEAN is changing, which is mostly what it takes out. If you hear mud and hiss, good. If you hear the voice itself, back CLEAN off.
S on ENHANCE plays the weight, air and presence being added, on their own.
S on VOICE plays the difference VOICE makes, and the bands it is working on light up red on the SPECTRUM view. If it sounds like speech, you are in the right place. If it starts to sound like the drums, ease VOICE or SENSITIVITY down.
Only one S lights at a time, and they never save with your session, so you can never bounce a solo by mistake.
BETA BUILD · VOID SKIN · INTERFACE MAY CHANGE BEFORE RELEASE
Levelling
Quiet bits up. Loud bits down.
Real recordings wander. A guest leans back from the mic, the intro music comes in hot, someone laughs. LEVELLING learns your timeline as you play it and evens out the quiet and loud sections, the way an engineer would ride the fader.
Choose GENTLE, MEDIUM or FIRM, or switch it OFF. It moves between sections, not between words, so speech keeps its natural shape. Silence stays silent. The plan is saved with your session and gets better each time you play, and ANALYZE starts a fresh one whenever your edit changes.
BEFOREAFTER LEVELLING
Presets
Start where you publish.
Four creator presets set VOICE and its fine controls, LEVELLING and the right loudness target in one click. Pick one, play, press AUTO.
A presenter over music, recorded in an ordinary room.
Speech on its own, or over very little.
Vertical video heard on a phone speaker. Extra clarity so words cut through.
A live voice over game audio or music.
Making music too? Nine more starting points cover DEFAULT · HOUSE · HIP-HOP · ROCK · METAL · POP · JAZZ · ACOUSTIC · CINEMATIC. DEFAULT starts with nothing engaged, so a fresh insert never changes your sound on its own.
The idea
In one side. Out the other.
The PENROSE ULTRA interface is drawn as a bridge through spacetime. Your signal falls in on the left, is stripped down to what holds up, and comes out on the right.
The picture is the signal path. The left half is CLEAN: it only ever takes away. The right half is ENHANCE: it only ever gives back. You always know which side is doing what.
CLEAN · TAKES AWAYTHE THROATENHANCE · GIVES BACK
The gold
Four lights. One reward.
When your mix comes out the other side intact, the whole bridge blooms gold. It is the only reward the interface gives, and it is earned.
Every light is measured on the audio, never read from where the knobs sit. Four lights rather than a score: they tell you which one is dark and which control fixes it. The verdict reads EXITED INTACT, PASSING THROUGH or OVERDRIVEN.
PASSING THROUGH
FOUND SOMETHINGCLEAN is actually removing something from this mix.
NOT SCOUREDAnd it is taking the faults, not the mix.
GLUED, NOT SQUASHEDGLUE sits inside its musical window.
CEILING HOLDINGLOUDNESS catches peaks instead of crushing the mix.
Three views
See what happened, three ways.
One click switches the plate between three views of the same measurements. Every cell on screen is drawn from the audio, so the views never disagree.
SPECTRUM
Input on the left, output on the right, frequency running round the ring from 20 Hz at the bottom to 20 kHz at the top. What CLEAN removed and what ENHANCE and VOICE added, band by band.
HISTORY
Loudness over the timeline. It follows your host's playhead, holds still when you stop, and clears only when you press RESET.
FLOW
The signal in motion. Input streams into the core on the left and is emitted on the right. What CLEAN cut rides in as white specks, limiting piles up against the ring, and VOICE's lift leaves in gold.
Meters
Loudness you can deliver to.
Input and output loudness for the last 3 seconds and for the whole piece, loudness range, true peak, how much CLEAN removed, how hard LOUDNESS is limiting, and how much of the time someone is talking. Integrated loudness follows your host's timeline: it keeps measuring while you play, holds its reading when you stop, and starts again only when you ask it to.
OUTPUT · LAST 3 s−21.1LUFS
INTEGRATED−20.9LUFS
TRUE PEAK−6.2dBTP
CLEAN REMOVED1.0dB
VOICE1%
TALKING19% OF THE TIME
RANGE0.5LU
LIMITING0.0dB
Skins and the drawer
Four looks. One plate.
Switch the whole interface between VOID, PAPER, PRESS and HEAT to suit your room and your eyes. Light emits on the dark skins, and ink is laid down on the light ones.
Open the drawer.
ADVANCED reveals the controls behind CLEAN and ENHANCE, for when you want to fix one thing precisely.
| CLEAN | RESONANCE · HARSHNESS · MUD · WARMTH |
|---|---|
| ENHANCE | WEIGHT · AIR · GLUE |
| BUS | MIX · OUTPUT |
Coming soon
IN DEVELOPMENT
Vocal mode, for singers.
VOICE is tuned for the spoken word. On a song it will lift the lead singer too, which is exactly what some mixes need and exactly wrong for others. A dedicated Vocal mode is on the way, retuned for sung voices. You choose the mode yourself, so it never changes behind your back.
SPEECHVOCAL
Specifications
| FORMATS | AU · VST3 · AAX |
|---|---|
| SYSTEM | macOS 11 or later, Apple Silicon |
| CONTROLS | CLEAN · ENHANCE · VOICE · LOUDNESS, with MUSIC ↔ VOICE, BODY ↔ CLARITY, SENSITIVITY, three S buttons and nine advanced controls |
| AUTO MIX | TARGET (−24 to −9 LUFS), AUTO, LEVELLING (OFF, GENTLE, MEDIUM, FIRM), ANALYZE |
| PRESETS | YOUTUBE · PODCAST · SHORTS · STREAM, plus nine music voicings |
| CEILING | True peak held at −1 dBTP |
| METERING | Short-term and integrated LUFS, loudness range, true peak, voice activity |
| VIEWS | SPECTRUM · HISTORY · FLOW |
| SKINS | VOID · PAPER · PRESS · HEAT, two interface sizes |
Built by THETA.
PENROSE ULTRA shares a name and a bridge with PENROSE, THETA's free two-knob mastering processor, but it is its own plug-in with its own sound, rebuilt for creators.
The two install side by side and never touch each other's sessions.
Availability
PENROSE ULTRA
The beta opens soon. Leave your email and we will send the installer when it is ready, along with the release notes.
Join the betaAU / VST3 / AAX INCLUDED