Robotic Audio · User manual

Loudness Pilot

A loudness meter that also does something about it. It measures your podcast, audiobook or mix the way the streaming services and broadcasters do, sets the gain that lands it on their target in one click, can ride the level of a voice like an engineer on a fader, and keeps the peaks under the ceiling with a true-peak limiter.

  • EBU R128 / ITU-R BS.1770 metering
  • Integrated, short-term, momentary, range, true peak
  • Trim to target in one click
  • Auto leveller for speech
  • True-peak limiter
  • 11 delivery presets
  • AU · VST3 · Standalone
Loudness Pilot's window
Loudness Pilot: the six readouts along the top, the last minute of loudness with the leveller's gain and the limiter under it, and the Target, Auto Level and Output cards below.

Quick start

From an unknown level to a file that meets the spec.

01

Put it last

Insert Loudness Pilot as the last plugin on your master or voice bus, after any EQ and compression, so it measures and limits what really goes out.

02

Pick the standard

Choose a preset for where it's going: Podcast Stereo -16 LUFS for Apple Podcasts, Spotify / YouTube -14 LUFS, Broadcast EBU R128 -23 LUFS and so on. That sets the Target and the Ceiling.

03

Play it through

Play the whole programme, or a typical stretch of it. Integrated fills in and tells you how far off the target you are.

04

Trim to target

Press TRIM TO TARGET. Gain is set so the integrated loudness lands on the target, and the measurement starts over.

05

Check again

Play it through once more. Integrated should read on target (in the accent colour), and True peak under the ceiling.

06

Bounce

With Reset on play on, the measurement starts over whenever the host starts playing, so a bounce is measured from its first second.

Voices that come and go? Turn Auto Level on: it rides the gain so quiet and loud speakers end up at the same loudness, before the trim and the limiter. Hover over anything and the strip along the bottom says what it does.

The window

Numbers on top, the last minute in the middle, the controls below.

Header
The preset bar (previous, next, the list, SAVE), the power button (bypass) and the menu (three lines) with this manual. Click the logo for the version and credits.
Readouts
Integrated (wide), Short-term, Momentary, Range, True peak and Input. See The readouts.
Graph
The last 60 seconds of loudness round the target, with the leveller's gain and the limiter's reduction in a strip under it. See The graph.
Target
The loudness to aim for and the true-peak ceiling.
Auto Level
The leveller, with its switch in the card's title: Speed, Range and Gate.
Output
Gain, the limiter switch, TRIM TO TARGET, RESET and Reset on play.
Tooltip strip
Along the bottom: what the control under the mouse does.

Signal flow

Measure, level, trim, limit, measure again.

In Inputmeter LevellerAuto Level Gaintrim Limitertrue peak Outputmeter Out momentary loudness

The input is measured first (the Input readout). The leveller takes that measurement every 100 ms and sets its gain from it, then Gain is added, then the limiter keeps the true peaks under the Ceiling. Everything else on screen measures what comes out at the end: what your listeners will hear.

The limiter looks 2 ms ahead, so the audio is delayed by that much (plus a few samples), on or off. The host is told and compensates. The measurements follow ITU-R BS.1770 and EBU R128, as the streaming services and broadcasters use them.

The readouts

Six numbers, coloured by how they sit against the target.

Integrated
The loudness of everything since the last reset, gated as EBU R128 does it (silence and pauses don't pull it down). This is the number delivery specs talk about. Under it: on target (within 1 LU), or how many LU too loud or too quiet. Its colour says the same: the accent on target, red too loud, blue too quiet.
Short-term
The loudness of the last 3 seconds, and the highest it has been since the reset.
Momentary
The loudness of the last 400 ms, and its highest.
Range
How much the loudness varies (LRA, EBU Tech 3342), once there are 3 seconds or more to judge. Speech usually sits between 3 and 8 LU; under it, Even, Dynamic (over 6 LU) or Very dynamic (over 10 LU).
True peak
The highest peak since the reset, between samples too, as a decoder will see it. It turns red when it went over the ceiling.
Input
The integrated loudness of what comes in, before the leveller, the gain and the limiter, with its short-term loudness under it. Compare it with Integrated to see how much Loudness Pilot moved.

RESET (in the Output card) starts Integrated, Range, True peak and the maximums over.

The graph

The last minute, round the target.

Time runs from right (now) to left (a minute ago). The vertical scale is centred on the target: 12 LU above it and 24 below, with a line every 6 LU labelled in LUFS and the target drawn as a band across the graph.

Momentary
The filled area: the loudness of each 400 ms, jumping with every word or beat.
Short-term
The bright line: the loudness of the last 3 seconds, the one to keep near the target band.
Gain strip
Under the graph, ±12 dB round 0. The green line is the leveller's gain (LEVELLER, with its current value); red bars down from the top are the limiter's reduction (LIMITER). Frequent red bars mean the limiter is working hard: lower Gain or the Target.

Target

Where it has to land.

Target
The integrated loudness to aim for: Apple Podcasts −16, Spotify and YouTube −14, mono podcasts −19, EBU R128 −23 LUFS. The readouts, the graph, TRIM TO TARGET and the leveller all work towards it. −30 … −10 LUFS
Ceiling
The highest true peak allowed out; the limiter keeps peaks under it. −1 dBTP suits most services; audiobooks often ask for −3. −9 … 0 dBTP

Auto Level

An engineer on the fader, for speech.

Switch it on with the On pill in the card's title. The leveller listens to the input's momentary loudness, keeps a slow average of it and sets the gain that brings that average to the target. A soft speaker comes up, a loud one goes down, and the integrated loudness stays close to the target before any trim. Its gain glides between steps, so it never clicks. With it off, its gain glides back to 0 dB and the three knobs grey out.

Speed
How many seconds the leveller listens back before it moves the gain. Short follows every phrase; long only rides big changes, like a new speaker. It turns down four times faster than it turns up, as a compressor would. 0.5 … 10 s
Range
The most it may turn up or down. 1 … 24 dB
Gate
Under this loudness (pauses, breaths, room noise) it holds its gain instead of turning silence up. Raise it in a noisy room. −70 … −30 LUFS
Gaps between words don't count. A dip more than 10 LU under the average is taken for a gap, not a quieter voice, and only starts to count once it has lasted 1.5 seconds. So the leveller doesn't pump up the noise between words.

Output

Trim, limit and measure.

Gain
Added after the leveller, before the limiter. TRIM TO TARGET sets it for you. −24 … +24 dB
True-peak limiter
Keeps peaks, between samples too, under the Ceiling. It looks 2 ms ahead, so the gain is already down when a peak arrives, and recovers smoothly (120 ms). Off, the audio is still delayed by the same amount, so switching it never moves the timing.
TRIM TO TARGET
Sets Gain so the integrated loudness lands on the target, then starts the measurement over. It needs a measurement first: play the whole programme, or a typical stretch of it.
RESET
Starts the integrated, range and peak measurements over.
Reset on play
Starts the measurement over whenever the host starts playing, so a bounce or a take is measured from its start.
Limiting is a safety net. If you trim a dynamic programme up a lot, the limiter has to take the peaks down and the integrated loudness ends up a little under the target. Trim, play it through again and trim once more, or level it first with Auto Level or a compressor.

Presets

The delivery standards, and the same with the leveller riding.

PresetTargetCeilingLeveller
Default−16 LUFS−1 dBTPOff
Podcast Stereo -16 LUFS−16 LUFS−1 dBTPOff
Podcast Mono -19 LUFS−19 LUFS−1 dBTPOff
Spotify / YouTube -14 LUFS−14 LUFS−1 dBTPOff
Audiobook -20 LUFS−20 LUFS−3 dBTPOff
Broadcast EBU R128 -23 LUFS−23 LUFS−1 dBTPOff
Broadcast ATSC A/85 -24 LKFS−24 LKFS−2 dBTPOff
Auto Level Podcast−16 LUFS−1 dBTP2 s, ±12 dB
Auto Level Podcast Mono−19 LUFS−1 dBTP2 s, ±12 dB
Auto Level Gentle−16 LUFS−1 dBTP6 s, ±6 dB
Auto Level Interview−16 LUFS−1 dBTP1 s, ±18 dB

Press SAVE in the header to keep your own. The preset menu has Save As, Delete and Open Folder. An asterisk after the preset's name means you've changed something since loading it. Every control is a host parameter and can be automated.

Recipes

Starting points for common jobs.

Finished mix

Choose the service's preset, play the whole track, TRIM TO TARGET, play it again to check. Leave Auto Level off: it would flatten the music's dynamics.

Two-person interview

Auto Level Interview on the voice bus. If the quieter guest still drops, shorten Speed; if breaths come up, raise the Gate.

Solo narration

Auto Level Gentle: it only evens out the slow drift of a long read, not the expression in it.

Just measuring

Limiter off, Auto Level off, Gain 0: Loudness Pilot only delays the audio by its lookahead and tells you where you are.

Mono podcast

Podcast Mono -19 LUFS. A mono file plays from both speakers, so it sounds as loud as stereo at −16.

Check the range

For speech, a Range over 8 LU means listeners in a car will keep reaching for the volume. Level it with Auto Level or a compressor before trimming.

Specifications

The numbers.

PropertyValue
MeteringITU-R BS.1770 K-weighting; momentary 400 ms, short-term 3 s; integrated gated at −70 LUFS absolute and −10 LU relative (EBU R128); loudness range per EBU Tech 3342
True peak4× oversampled estimate (BS.1770)
LevellerSteps every 100 ms; Speed 0.5–10 s (four times faster down than up); Range ±1–24 dB; Gate −70 to −30 LUFS
LimiterTrue-peak, 2 ms lookahead, 120 ms release, stereo-linked
Latency2 ms plus 16 samples, on or off, reported to the host
GraphThe last 60 seconds, 12 LU above to 24 LU below the target
CPUThe meters rest on digital silence; the window redraws only what changes
ChannelsStereo (mono tracks work too)
FormatsAU, VST3, Standalone