LRA, PLR, PSR: Dynamic Range

LRA measures loudness range, PLR peak-to-loudness, PSR peak-to-short-term. Learn targets for music, podcast, film.

LRA — Loudness Range

EBU R128 defines LRA as difference between 10th and 95th percentile of short-term loudness (3s window, > -70 LUFS). Music: 4-9 LU typical, Podcast: 4-7 LU, Film: 15-20 LU.

PLR — Peak to Loudness Ratio

PLR = dBTP - Integrated. Indicates how much headroom you left. Mastered pop: PLR 6-9, Dynamic jazz: PLR 12-16. Lower PLR = louder, more compressed.

PSR — Peak to Short-term Ratio

PSR = Short-term Max - Integrated. Shows punch of loudest 3s vs whole track. EDM drop PSR 6-9, Podcast PSR 4-6.

LRA in depth: how it's actually calculated

LRA isn't "max minus min." EBU R128 defines it precisely: take the short-term loudness (3-second sliding window) across the whole program, discard anything below -70 LUFS absolute, then find the level exceeded by 95% of the remaining distribution (the 10th percentile, LRA low) and the level exceeded by 5% (the 95th percentile, LRA high). LRA = high − low, in LU.

Why percentiles instead of min/max? Because a single gunshot or a single silent gap would otherwise define your entire dynamic range. The 10/95 spread describes the body of your program — where the loudness actually lives. Two tracks can have identical integrated LUFS and wildly different LRA, and they'll feel completely different.

LRA targets by content type

Music — 4 to 9 LU typical. Modern pop/EDM/hip-hop: 4–6 LU (dense, consistent). Rock/indie: 6–9 LU. Jazz/classical/acoustic: 9–15 LU. If your pop master shows LRA 12, it's probably under-compressed for the genre — or it's beautifully dynamic and you should leave it alone. Context decides.

Podcasts & spoken word — 4 to 7 LU. Speech needs consistency: listeners in cars can't ride the volume. If LRA exceeds 8, quiet guests get lost and loud laughs startle. Gentle compression and manual gain riding beat heavy limiting here.

Broadcast & film — up to 20 LU. EBU R128 allows LRA up to 20 LU for most broadcast content (5 LU for short-form commercials, which must stay consistent). Film uses the full range deliberately — dialogue at -27, explosions at -7.

Audiobooks — 4 to 6 LU. ACX/Audible-adjacent work wants tight, consistent narration. Wild dynamics annoy listeners falling asleep.

PLR: reading your headroom

PLR = true peak (dBTP) − integrated LUFS. It tells you how much peak headroom exists above your average loudness — essentially, how hard you hit the ceiling to achieve your level.

Typical values: heavily limited pop master at -8 LUFS with -1 dBTP ceiling → PLR 7. Dynamic acoustic track at -14 LUFS, -1 dBTP → PLR 13. Orchestral → PLR 16+.

PLR is most useful comparatively: if your reference tracks sit at PLR 9 and yours is at 5, you're pushing 4 dB harder into the limiter than the genre norm. Maybe that's the sound you want. But now it's a decision, not an accident.

Warning sign: PLR below 5 on anything but intentional lo-fi means the limiter is doing structural damage — pumping, flattened transients, inter-sample distortion risk even with true-peak limiting.

PSR: the punch meter

PSR = short-term maximum − integrated LUFS. While PLR measures peak headroom, PSR measures how much harder your loudest 3 seconds hit compared to the whole-track average. It's the "drop factor."

EDM drop: PSR 6–9. Pop chorus: PSR 4–7. Podcast: PSR 3–5 (a laugh or emphatic moment). If PSR is under 3 on a dance track, the drop isn't dropping — the whole track is one flat plateau and listeners feel no arrival.

PSR pairs with short-term metering: watch the short-term readout during your loudest section, subtract integrated, and you have it. JSM Loudness computes it automatically.

How the three work together: reading a full picture

Integrated LUFS tells you how loud. LRA tells you how much it moves. PLR tells you how hard you pushed. PSR tells you how hard the peaks hit. Together they describe a master more completely than any single number:

• -8 LUFS, LRA 4, PLR 6, PSR 4 → modern loud pop/EDM. Dense, competitive, intentional.
• -14 LUFS, LRA 8, PLR 12, PSR 7 → dynamic indie/acoustic. Breathing room, transients intact.
• -14 LUFS, LRA 3, PLR 6, PSR 2 → problem: quiet and crushed. Turned-down loud master — the worst of both worlds.
• -23 LUFS, LRA 18, PLR 18, PSR 12 → broadcast/film mix. Huge range, totally appropriate.

Learn to read these four as a set and you'll diagnose masters in seconds.

Fixing dynamics problems

LRA too high for the format (podcast at 11 LU): use serial gentle compression (2–3 dB GR per stage) rather than one hard limiter. Automate dialogue levels first — fader rides beat compression for natural speech.

LRA too low (lifeless acoustic at 3 LU): back off the mix bus compressor, raise the limiter ceiling, or accept that the arrangement wants density. Don't "fix" intentional aesthetics.

PLR too low (everything at PLR 5): you're limiter-bound. Either embrace it (it's a valid sound) or rebuild gain staging so peaks have somewhere to go — often the mix needs 2–3 dB more crest factor before mastering.

PSR too low (drop at PSR 2): the issue is usually arrangement/mix, not mastering. Automate the build 2–4 dB hotter into the drop, or sidechain the drop's low end so the transient punches through.

Frequently asked questions

What's the difference between LRA and dynamic range?

Classic "dynamic range" (peak minus noise floor, or crest factor) is an electrical measurement. LRA is perceptual — it's the range of K-weighted, gated loudness the listener actually experiences. A track can have 20 dB of electrical dynamic range but only 5 LU of LRA if the loud parts dominate.

Can LRA be too low?

Artistically, no — some genres live at LRA 3–4. Technically, very low LRA (under 3) with high integrated loudness means heavy limiting; check for pumping, distortion, and listener fatigue. If it's intentional and sounds good, it's fine.

Do streaming platforms use LRA?

Not for normalization — they use integrated LUFS. But LRA predicts translation: high-LRA content varies more between quiet and loud playback systems, and very low-LRA content can sound flat after platform transcoding. It's a quality metric, not a delivery spec.