Mix Tools

V-PRESS

An instant vocal compressor. Put it on a vocal, move one control, get a finished vocal.

£29 £35Launch price

See the measurements

Formats: VST3 · AU · Standalone
Operating system: macOS · Universal
Version: v1.0.0

What it does

One control. A finished vocal.

No attack, no release, no ratio, no knee — not because they were too hard to expose, but because a vocal compressor that needs them is not doing its job.

One control moves a two-stage compressor, a noise-floor guard and the auto make-up together. The output stays level-matched as you turn it, so you are judging how controlled the vocal is rather than how loud it is. Measured across the whole knob range, output loudness moves 1.96 dB.

Inside it is the chain every engineer builds by hand: a slow RMS leveller doing the density and the bring-quiet-words-forward work, into a fast peak catcher that only ever sees consonants and stray peaks.

Key features

The detector is the whole trick

A 24 dB/oct high-pass so plosives do not duck the whole word on every “p”. A +4.5 dB bell at 2.6 kHz weighting detection toward the band that carries intelligibility. A −5 dB shelf above 7.5 kHz so sibilants do not duck the word on every “s”. Detector-only shaping — it never touches the audio.

Two stages, because one cannot do both jobs

Slow and gentle levels beautifully but lets consonants through; fast and hard controls peaks but flattens the performance. V-Press builds the leveller-into-peak-catcher chain in and moves both ends from one control.

Release lengthens as the knob comes up

The opposite of a general-purpose compressor. With 16 dB of make-up in play, a short release lets the gain recover between words so the next syllable arrives while the stage is still open — measured, that pushes the 99th-percentile short-term peak up by 2.4 dB and lifts room tone by an extra 3 dB.

Look-ahead chosen at the knee of the curve

The detector reads 2 ms ahead. Without it, a syllable after a gap takes the full make-up on its transient — measured at maximum compression, the short-term peak came out 3.4 dB louder than it went in. 1 ms recovers half the benefit; 3 ms adds almost nothing.

A limiter with a proof, not a margin

The applied gain can never exceed what the sample needed, and because the smoothing kernel is triangular the gain is continuously differentiable and cannot click. Run inside the oversampled path, the ceiling holds for the reconstructed waveform, not just for the samples.

Deterministic auto make-up

Computed analytically from the static curves rather than measured from the signal, so the same automation always produces the same output.

How it works

The signal path, as the plugin’s own documentation states it.

input ──► input meter
      ──► oversample (1× / 2× / 4×)
            ├─► sidechain analyser ──► noise floor guard
            │      (undelayed)     ──► leveller  (RMS,  slow)
            │                      ──► peak      (peak, fast)
            ├─► look-ahead delay (2 ms)
            ├─► compressor gain + auto make-up
            ├─► peak saturator
            ├─► presence
            ├─► mix with dry
            ├─► output gain
            └─► true-peak limiter
      ──► downsample ──► output meter

Audio

Audio demos for V-PRESS are not published yet.

We would rather show you nothing than show you a demo that was not recorded through this plugin.

Measured

VPressTests — 102 checks · synthetic vocal · 48 kHz

Short-term dynamic range (p90−p10): 18.5 dB in → 13.1 dB at 100 %
Quiet syllables (p10): −28.1 → −24.0 dBFS (+4.1 dB forward)
Output loudness spread across the whole knob range: 1.96 dB
THD at nominal vocal level (−15 dBFS): −86 dBc (0.005 %)
Limiter overshoot, at every ceiling tested: zero
Oversampler image rejection: −108 dB
CPU, stereo, one core: 0.46 % (1×) · 1.10 % (2×) · 1.98 % (4×)
auval: AU VALIDATION SUCCEEDED

Figures are produced by this plugin’s own test suite and benchmark, on the hardware and sample rate named above. They are not estimates.

Controls

Everything there is

Compression: 0–100 %. Threshold, ratio, ballistics and make-up, all at once.
Presence: Post-compression clarity, articulation and air.
Low Cut: 20–300 Hz. Detector high-pass — plosive sensitivity.
Mix: Parallel compression.
Output: ±24 dB level trim.
Ceiling: −12–0 dBFS safety limiter threshold.
Limiter: True-peak safety limiter, on by default.
Quality: 1× / 2× / 4× internal oversampling. Deliberately not automatable, because it changes reported latency.

Technical details

Formats: VST3 · AU · Standalone
Operating system: macOS 10.13 or later
Architecture: Universal — Apple Silicon and Intel
Plugin code: Vpr1
Version: 1.0.0
Latency: 154–240 samples (3.5–5.0 ms) depending on quality. Reported exactly, and constant within a setting.
CPU: 0.46–1.98 % of one core
Factory presets: 8

Clean Pop Vocal, Aggressive Modern, Intimate Vocal, Rap Vocal, Podcast Voice, Voice Over, Background Vocal, Natural Ride. Exposed as host programs, so they appear in the DAW's own preset browser.

Worth knowing

Low-frequency plosives pass through the compressor largely uncontrolled. That is a direct consequence of the detector design, not an oversight — the sidechain high-pass is what stops the compressor ducking the whole word on every “p”. The safety limiter catches them, and it is on by default.
Thresholds are absolute, not program-adaptive. A vocal tracked much quieter than about −40 dBFS will not reach the leveller. An adaptive threshold would make the response depend on material history, which breaks automation reproducibility.
The MODE control in the artwork has no parameter behind it and is rendered disabled. COLOR affects the interface accent only and touches no audio.