A 20-show network running a mix of interview, narrative, and branded content will land episodes at wildly different integrated loudness levels if you're not normalizing at the network layer. Spotify reads -14 LUFS integrated and turns down anything louder — which sounds fine until you realize your show measured at -10 LUFS is now being attenuated 4 dB, and the compression applied in mastering sounds pumped and unnatural at reduced gain. Meanwhile, Apple Podcasts targets -16 LUFS and doesn't loudness-normalize at all for RSS-delivered content. The result: the same master sounds different depending on where a listener finds you.
LUFS (Loudness Units Full Scale) is the unit used to measure perceptual loudness over time. It's the metric that replaced RMS as the broadcast and streaming standard because it models human hearing more accurately — specifically, it weights frequencies the way ears do, giving midrange content more influence over the final number than subsonic rumble or extreme high frequencies. For podcast networks, understanding what the number means and how platforms interpret it is not optional. It's the difference between consistent-sounding output and a portfolio where each show feels like it came from a different studio.
The Platform Landscape Is Not Uniform
Before going into mastering targets, you need to understand what each major distribution platform actually does with your audio:
- Spotify: Normalizes playback to -14 LUFS integrated (with a -1 dBTP true peak ceiling). Content louder than -14 LUFS is turned down. Content quieter than -14 LUFS is turned up. This means there's no benefit to mastering hotter than -14; you'll just trigger attenuation that can make compression artifacts more audible.
- Apple Podcasts: Does not apply loudness normalization to podcast RSS feeds at the app level (unlike Apple Music, which normalizes to -16 LUFS). Your master is delivered as-is. This makes your choice of LUFS target a direct listener experience decision, not just a platform-compliance issue.
- YouTube: Normalizes to -14 LUFS for video content. For podcast-style audio-only uploads, the same normalization applies.
- Pocket Casts / Overcast / Castro: Some apps offer their own volume leveling as a user-controlled feature, applied on the client side. This is independent of your master.
- Amazon Music / Audible: Applies loudness normalization to -14 LUFS. Behavior on non-normalized content varies.
The practical takeaway: if you're mastering for broad distribution, -16 LUFS integrated with a -1 dBTP true peak ceiling is a defensible target that sounds good on Apple (where you land as mixed), plays well on Spotify (slight gain applied), and complies with YouTube normalization (slight gain applied). Mastering hotter than -14 primarily helps situations where your episode will be played from a file download with no normalization — and that's a shrinking use case.
True Peak vs. Integrated Loudness: They're Not the Same Problem
Networks conflate these two measurements constantly, which leads to mastering mistakes that cause real issues at delivery.
Integrated LUFS measures average loudness across the full duration of the program. It's gated — silence periods are excluded from the calculation. A 45-minute interview episode measured integrated will give you a sense of the show's average energy level. This is the number you target for platform compliance.
True peak is different. It measures the actual peak level of the reconstructed audio waveform, including intersample peaks that occur during digital-to-analog conversion. An episode can measure -16 LUFS integrated while having a true peak of -0.3 dBTP — and that intersample peak can cause clipping on some playback systems and DAC chains even though the waveform looks clean in your DAW.
The standard ceiling for podcast distribution is -1 dBTP. Some engineers target -2 dBTP for extra headroom, particularly for content destined for platforms that apply additional gain. If your integrated target is -16 LUFS and your limiter is set to -1 dBTP, you have 15 LU of headroom between your target and your ceiling — which is plenty for speech content but can be tight for music-heavy intros or sound-designed narrative episodes with wide dynamic range.
Loudness range (LRA) is the third variable that matters for network consistency. LRA measures the variation between the quietest and loudest sections of the program (excluding extremes). A conversational interview might have an LRA of 4–7 LU. A serialized narrative with dramatic pauses and music swells might run 12–15 LU. Both can target -16 LUFS integrated, but they'll sound very different on a platform that applies normalization — the high-LRA content will have its peaks pushed louder relative to the quiet sections than the low-LRA content will.
What Network-Level Mastering Actually Requires
Single-show mastering is straightforward: you have one editorial voice, one recording environment, one style of guest. Network mastering means you're managing variance across shows that record in different rooms, with different microphones, with hosts who have very different speaking volumes and dynamics.
Consider a 15-show network producing a mix of business interview content (typically measured -14 to -16 LUFS before processing, relatively consistent levels), true crime narrative (often mixed with music beds at -18 to -22 LUFS before processing, high LRA), and weekly news rundowns (fast-talking hosts, high dynamic range, often clipping if not carefully gain-staged). A single mastering chain with fixed settings will not produce consistent output across all three formats.
What actually works at network scale:
- Format-specific processing chains: Each show format gets its own set of processing parameters — compression ratio, attack and release times, limiter ceiling, de-essing aggressiveness. These are templated so they're consistently applied across episodes of the same show, not hand-tuned each time.
- Pre-mastering gain staging review: Before mastering, the integrated LUFS of the raw mix is measured. Episodes landing more than 6 LU away from the expected range for that format flag for review before automated processing begins. This catches recording session problems — a host who was unusually far from the mic, a guest call recorded at a lower level than normal.
- Post-mastering QC measurement: After the mastering chain runs, integrated LUFS, true peak, and LRA are measured and logged. Episodes outside tolerance are flagged, not auto-approved.
- Archival masters vs. delivery files: Network archives should retain the processed master at 48kHz/24-bit WAV before lossy encoding. The distribution MP3 at 128kbps or AAC at 192kbps is derived from this master. Re-encoding from a previously encoded MP3 introduces generation loss; encoding from the WAV master doesn't.
The -14 vs. -16 Decision Isn't About One Number Being Better
You'll see arguments for -14 LUFS on the basis that it sounds louder on Apple Podcasts (where there's no normalization) and that podcasts compete for attention against other media that's often mastered hotter. You'll see arguments for -16 LUFS on the basis that it's more conservative, preserves more dynamic range, and plays cleanly on every normalization scheme without any gain attenuation.
We're not saying -14 LUFS is the wrong target. We're saying the choice should be deliberate and documented, not ad hoc. For a network that distributes to Spotify, Apple, and Google Podcasts equally, -16 LUFS integrated at -1 dBTP true peak is a safe, defensible default. For a network whose primary audience listens on Apple devices without normalization, mastering to -14 LUFS may genuinely produce a better perceived experience — provided your source material can sustain that level without artifacts.
The pathological case to avoid: mastering each episode to a different target, or processing the same show differently across episodes, because your engineer eyeballs the level and applies processing by feel. At two episodes a week across 15 shows, that's 30 unique mastering decisions every week. The variance adds up fast, and listeners notice — not necessarily as "this episode was too loud" but as "something feels off about this show lately."
Encoding for Distribution: Where LUFS Work Gets Undone
Mastering to -16 LUFS and then encoding to MP3 at 128kbps joint stereo introduces a variable that many producers don't account for: codec psychoacoustic processing can shift the perceived loudness of an episode by 0.5–1.5 LU depending on the content. Speech-heavy content typically encodes cleanly. Episodes with significant low-frequency content — music-heavy intros, narrative sound design — can shift more.
The correct workflow is to measure the final encoded delivery file, not just the WAV master, and confirm it falls within tolerance. A master that measures -16.2 LUFS integrated as a WAV might measure -15.8 LUFS after MP3 encoding. That's usually within acceptable variance. A shift to -14.1 LUFS on an episode with heavy music should trigger a review of the mastering chain, not just acceptance of the delivered file.
For networks that use dynamic ad insertion (DAI), there's an additional consideration: your episode content and the inserted ads may have different loudness levels. If your episode content is mastered to -16 LUFS and the ad spot that's inserted is at -14 LUFS (which is common for programmatic audio ads), listeners will experience a loudness jump at the ad break. The best DAI platforms apply loudness normalization at insertion time, but not all do — and even those that do may use different target levels than your mastering target.
Practical Starting Points for Network Audio Standardization
If you're bringing consistency to a network that's currently inconsistent, a few practical steps in rough priority order:
- Audit your existing back catalog: measure integrated LUFS, true peak, and LRA for the last 20 episodes of each show. Map the variance. Understand where the problem is worst.
- Document a mastering spec per show format — not per show, but per format class. Interview format, narrative format, branded content format. Three specs cover most networks.
- Build measurement into your publishing workflow, not as an optional step. LUFS measurements should be logged and stored for every episode before it goes live.
- Apply normalization to your RSS feed distribution consistently. Whether you use -14 or -16 as your target matters less than whether that target is applied consistently across every episode of every show.
The goal isn't perfection — it's predictability. A listener who subscribes to three shows on your network should feel like they're listening to a coherent editorial product, even if the shows have very different formats. Loudness consistency is one of the cheapest ways to achieve that, and it's entirely within your control.
