Overview
The Jeesonic Metering Cockpit is a loudness metering plug-in with a flexible frequency analyzer and monitoring aids. It shows what is going on in you mix: loudness to BS.1770-4, peak level, dynamics, spectrum, stereo correlation, waveform, and a loudness history.
Everything is visible at the same time. Each meter is answering a different question. While you mix, you want to see everything at a glance without the need to change tools.
The plug-in adds no latency. You can leave it anywhere in the chain. It leaves your audio alone unless you ask it not to. You can compare your mix to a reference file, you can listen only to the low-, mid-, or high frequencies or to a selected frequency band on its own.
Installation
macOS
Run the macOS installer (a notarized .pkg). It adds two formats:
- VST3 to
/Library/Audio/Plug-Ins/VST3 - Audio Unit (AU) to
/Library/Audio/Plug-Ins/Components
Windows
On Windows, the Metering Cockpit ships as a VST3 inside a zip (Audio Unit is macOS-only). Unzip it and copy the folder Jeesonic Metering Cockpit.vst3 into your system VST3 folder:
C:\Program Files\Common Files\VST3\
Create that VST3 folder if it does not exist. Writing there needs administrator rights. Windows may ask you to confirm.
After installing on either platform, rescan plug-ins or restart your host.
The plug-in is free to use, privately and commercially, with no registration and no time limit. Please download it from jeesonic.de rather than from a copy elsewhere. That way you always get the current build. See the terms of use.
The interface
The window has four areas: the meter column on the left over the full height, the loudness card at the top right with the numbers and the history, the analyzer below it, and the waveform along the bottom.
The loudness numbers
- Integrated. The gated average over everything measured, in LUFS. This is the number a delivery spec means. It turns green when you are within 1 LU of your target.
- Short-Term. The last three seconds. This is the value that moves with the music.
- Max True Peak. The highest peak of the session, in dBTP, measured with four-times oversampling. It turns red at and above -0.1 dBTP. Switch the level column to sample peak (see below) and this cell follows: it then reads Max Sample Peak in dBFS.
- LRA. Loudness range in LU: how far the loud and quiet parts of the programme are apart, measured over the whole session to EBU R128.
The loudness history
The curve beside the numbers scrolls from oldest on the left to newest on the right. The box on the right of the switch row sets how much time it spans: 15 s, 30 s, 1 min, 2 min or 5 min. The time axis always spans exactly what the box says. The curve then scrolls at an even pace.
The row of switches at the top left picks which curve that is, one at a time:
- MOMENTARY (violet) loudness over a 400 ms window
- SHORT-TERM (green) loudness over a 3 s window
- INTEGRATED (amber) Integrated Loudness according to ITU-R BS.1770
The curves appear one at a time. All three measure the same quantity over different time frames.
Whatever you pick is saved with your project.
The dashed line is your target. Grab it and drag it up or down; the value chip moves with the line. Your host sees this as one automation gesture. You can record it like any other parameter.
The meter column
- TRUE PEAK / SMPL PEAK. Peak level per channel from 0 to -48 dB, with a held peak tick. The gradient goes from blue through yellow to red. Red is only the very top.
- CREST. This one is not a level meter. The bar floats from the short-term RMS up to the held peak. Its length is the crest factor and visualizes how much dynamics is left. The colour follows that length: red when the material is squashed close to its peak, through orange and yellow, to green when there is plenty of headroom.
The crest colour has these anchors: green from about 10 dB, yellow around 8 dB, orange around 6 dB, red at 4 dB and below. Take them as orientation, not as a spec.
Both columns use the same dB mapping. This means they line up, and the comparison does not mislead you.
Click the label above the level value to switch between the two ways of measuring a peak. True peak (the default) reconstructs what happens between the samples with four-times oversampling. That is the number that tells you whether a converter or an encoder will clip. It is also what delivery specs mean. Sample peak reads the sample values themselves, which is what a waveform editor shows you. True peak is never lower than sample peak. Both are measured the whole time. Switching only changes which one you see: the maximum still covers the whole measurement, and nothing else is reset.
The crest factor is calculated based on sample peak in both modes.
Below each column stands one number: for PEAK the held peak of the louder channel, for CREST the crest of the quieter-sounding one. Both report the worse channel, because that is the one you need to know about. The columns still show each side separately.
At the bottom, CORR is the stereo correlation from -1 to +1. Green above +0.5, yellow above 0, orange above -0.5, red below. On a mono track it reads "mono", and the needle disappears. A signal correlated with itself tells you nothing.
The analyzer
BANDS sets the number of bands from 1/1 to 1/24 octave, or Lin for the raw FFT bins.
The analysis window follows that choice: 8192 points up to 1/3 octave (fast and lively), 16384 points from 1/6 on (finer in frequency, slower to react).
AVERAGING sets how the bars rise: Inst follows the signal, RMS averages over about a second. How fast they fall back is the Decay setting in the settings menu.
The thin caps above the bars hold the recent maximum and then sink slowly. That way you see the peak of a passage without staring at it.
The waveform
Two lanes, left on top and right below, drawn like an oscilloscope: each column stays where it is and the bright head writes over the oldest parts of the waveform. The trace fades toward the older material.
The box at the top right sets how much time is on screen, the same way the box over the loudness curve sets its time axis: 2 s, 5 s, 10 s, 30 s, 1 min, 2 min or 5 min. The measurement is not affected.
The waveform is kept, not just drawn: five minutes of it, for your mix and for the reference at the same time. A longer window shows what was there before right away, instead of filling up from now on.
RESET clears the waveform too. The trace starts again at the left.
The row under the numbers holds three things, all of them plain words: RESET on the left, the measurement clock in the middle, SYNC on the right.
The clock between the two buttons counts how long the current measurement has been running. RESET puts it back to zero.
SYNC, next to RESET, ties the new measurement to your DAW. Whenever the playhead jumps back, a new measurement starts. Each pass then stands on its own, and you do not have to remember to press RESET. Pausing and playing on from the same place never resets. It is on by default and is saved with your project.
Listening to a reference
You can load a reference track that you know well and switch between it and your own mix: MIX and REF sit at the top, next to the file name. Click the name to load a file, to load another one, or to remove it.
The reference follows the playhead. The same loop always plays the same part of it. When you stop the transport, it goes silent with everything else. Past the end of the file there is silence; it never starts over on its own, because that would compare two different places with each other.
Level it before you judge it. A reference that islouder always sounds better, and that has nothing to do with the mix. Drag the dB field next to the file name to set the level. Double-click it to go back to zero. The file's own loudness sits in the menu behind the file name, measured while it loaded. That gives you a number to set the level against, instead of guessing.
The level is a listening aid. It does not change the numbers the reference shows.
The numbers follow what you hear. Both your mix and the reference are measured the whole time, whichever one is playing. Switching only changes which set of numbers you see: LUFS, peak, crest, spectrum, waveform and the history curve all belong to the source you are listening to. Both have been measured from the start. Their integrated values therefore cover exactly the same stretch of the timeline, which makes comparing them fair. The level trim does not count towards the reference's numbers: it is there for your ears, while the numbers should tell you how loud the reference really is.
RESET starts both measurements over. They then stay comparable.
While REF is selected, a warning sits in the header. The reference is never part of an offline render: a bounce, an export or a freeze always contains your own mix, whatever the switch says.
The file is converted to your session rate once, when it loads, and then lives in memory. That is why loading takes a moment, and why files are limited to ten minutes.
Listening to one band
Under the analyzer sits a strip split into LOW, MID and HIGH. Click a field and you hear that band alone. Click the same field again and the whole signal is back.
The strip sits under the frequency axis of the analyzer. Each field is exactly as wide as its range in the bars above it, so you see what you are about to hear. Drag the two splits to move them. They cannot pass each other, and the middle band stays at least an octave wide. The filters are 24 dB per octave and add no latency. Nothing in your project shifts in time.
The measurement does not follow. LUFS, peak, crest, the analyzer and the waveform always show the full signal, whichever band you are listening to.
While a band is soloed, a warning sits in the middle of the header. The split never applies while your host renders offline: a bounce, an export or a freeze always contains the full signal, even if you forgot to switch the strip back. A real-time recording out of your interface is the one case the plug-in cannot protect you from; that is what the warning is for.
The question mark
Every display in this plug-in explains itself: hover over a bar, a curve or a number and a short text says what it shows. The question mark in the top right turns those texts off once you know your way around, and brings them back when you need them. It is lit while they are on. The setting is saved with your project.
Settings
The gear icon opens the settings.
Analyzer, Slope. Tilts the display by 0, 3, 4.5 or 6 dB per octave, pivoting at 1 kHz. Purely cosmetic: it lifts the top octaves and lowers the bottom ones. Pink noise or a typical mix then reads flatter. It changes nothing about the measurement.
Analyzer, Decay. How fast the bars fall back: Fast, Medium, Slow or Very slow. The rise is always fast, because a display that misses a transient does not measure it.
Make Default. Saves the current settings as the starting point for new instances. Once a default exists the entry reads "Update Default" and a "Factory Default" entry appears to remove it. A project that you reload always restores its own saved settings. The default therefore only applies to a genuinely fresh instance.
Version. The last entry shows the version you are running; opening it also names the loaded format (VST3 or AudioUnit). Please include both when you report a problem.
Requirements
- macOS 11 or newer, Apple silicon or Intel (VST3, AU)
- Windows 10 or newer, 64-bit (VST3)
- Any host that loads VST3 or AU