Guide
AI Mastering for Podcasts: Consistent Loudness
Marc Aaron Jacobs
Founder, DropCue · Composer
Every podcast episode should land at the same loudness, hit the platform targets, and sound clean from intro to outro. DropCue's AI mastering and loudness tooling, built for music creators, handles spoken word too.
Who this is for
This page is for podcasters, voiceover artists, and audio-drama producers who are tired of episodes that play back at wildly different volumes. You recorded episode 12 a little hot, episode 13 came out whisper-quiet, and now a listener has to ride the volume knob every time they switch shows. That inconsistency is the single most common reason people bail on an episode in the first thirty seconds.
You probably already know that loudness is measured in LUFS (Loudness Units Full Scale, defined by the ITU-R BS.1770 standard), and that the major platforms each have an expectation. Apple Podcasts and Spotify normalize toward roughly minus 16 LUFS for stereo spoken-word programming, and a lot of the podcast-specific guidance lands near minus 16 to minus 14 LUFS integrated with a true peak ceiling around minus 1 dBTP. If you have ever exported an episode, uploaded it, and then heard it get quietly turned down by the platform's normalization, you have met loudness compliance the hard way.
If you are an audio-drama or narrative-fiction producer, you have an even harder version of this problem: dialogue, music beds, and sound design all fighting for the same loudness budget, and you need the master to feel cohesive without crushing the quiet scenes. And if you are a solo voiceover artist delivering reads to clients, you often get a spec sheet that names an exact loudness and peak target, and "sounds about right" is not an acceptable deliverable.
The audience-specific reality
Here is the honest framing, because the memory of every composer who reads this will appreciate it: DropCue is built for music creators. It is a catalog, pitch-delivery, and analytics platform for people who write and place music. The AI mastering and AI mix analysis tools were built to make composers' tracks louder, cleaner, and competitive against commercial releases.
But loudness physics do not care whether the source is a string quartet or a two-host interview show. LUFS is LUFS. True peak is true peak. The ITU-R BS.1770 measurement that tells a composer their cue is sitting at minus 14 LUFS integrated is the exact same measurement a podcast platform uses to decide how much to turn your episode down. So the tooling that serves music masters also serves spoken-word masters, and the analytics that diagnose a mix translate directly to diagnosing a podcast export.
What is genuinely different about podcast and voiceover audio is the content shape. Music is usually dense and continuous; speech has gaps, breaths, plosives, and long dynamic swings between a calm intro and an animated debate. A mastering pass that is great for a four-on-the-floor track is not automatically great for a sparse interview. That is why the workflow here leans on the diagnostic read first (what is my integrated loudness, where are my true peaks, is my stereo image collapsing in mono on phone speakers) and then a mastering style choice second. Podcasts are also frequently consumed in mono or near-mono on a single phone speaker or a smart speaker, so mono compatibility and phase are not academic for you, they are the difference between a clear read and a hollow one.
Why DropCue fits this workflow
Most podcasters end up bouncing between four tools to ship one episode: the DAW or editor to assemble the cut, a standalone loudness meter to check LUFS, a spectrum analyzer to spot a muddy region, and then a separate mastering service or plugin to finalize. That is four context switches, four places to make a mistake, and usually four subscriptions.
DropCue collapses that. The AI mastering engine, powered by Chosen Masters, lets you pick a style (Modern, Open, or Powerful), master in seconds, and then A/B the before and after loudness-matched, with a moving waveform and a full studio analytics readout. Loudness-matched A/B matters more than it sounds: it stops you from being fooled into thinking "louder equals better," because both versions are leveled to the same perceived loudness so you are judging tonal balance and clarity, not just volume. For a spoken-word read, that means you can hear whether the master genuinely opened up the voice or just turned everything up.
The AI mix analysis engine, powered by ROEX, gives you the diagnostic side: integrated LUFS to ITU-R BS.1770, true peak (including inter-sample peaks that can clip on a listener's device even when your meter says you are under zero), clipping detection, stereo width, mono compatibility, phase, and tonal balance, plus plain-language recommendations on what to fix and how. For a podcaster who is not a mastering engineer, that plain-language layer is the point. It tells you "your true peak is over the ceiling, pull it down" instead of leaving you to interpret a wall of numbers.
One honest caveat, because it matters for spoken word specifically: if you analyze a lossy, already-compressed source (a re-exported MP3 instead of your original WAV or FLAC), the loudness and peak readings can be slightly skewed versus the original and are flagged as approximate. So run the diagnostic on your full-quality export, not on the compressed file you are about to upload. Get the master right at full resolution, then encode for the platform.
And because your audio already lives in DropCue, there is no upload-to-a-separate-service step. The episode you stored is the episode you analyze and master.
The features that matter most
✓ AI Mix Analysis (ROEX): integrated LUFS + true peak
See exactly where your episode sits against the roughly minus 16 to minus 14 LUFS spoken-word target and whether your true peak is under the minus 1 dBTP ceiling before a platform normalizes it for you.
✓ Mono compatibility and phase check
Most podcast listening happens on a single phone speaker or a smart speaker. The analysis flags if your stereo image collapses or your voice goes hollow in mono, which is invisible on studio headphones.
✓ AI Mastering (Chosen Masters) with style choice
Pick Open for a natural, breathy interview read or Powerful for a punchy ad-style narration, master in seconds, and get every episode landing in the same loudness zone instead of guessing.
✓ Loudness-matched A/B with moving waveform
Judge whether the master actually improved your voice's clarity and tonal balance, not just made it louder, because both versions are leveled to the same perceived loudness.
✓ Plain-language recommendations
You are a host or a voice actor, not a mastering engineer. The tool tells you what to fix and how in normal words, instead of leaving you to decode raw meter readings.
✓ Built into your catalog, no tool-hopping
Your episode audio already lives in DropCue, so you analyze and master it in place instead of bouncing between an editor, a separate loudness meter, a spectrum analyzer, and a standalone mastering service.
Names you may know in this space
ITU-R BS.1770
The international loudness measurement standard (integrated LUFS, true peak) that both music platforms and podcast platforms use to normalize playback, and the standard DropCue's analytics report against.
Apple Podcasts
Normalizes playback toward a spoken-word loudness target near minus 16 LUFS, which is why a hot or quiet master gets quietly adjusted on listeners' devices.
Spotify
Applies its own loudness normalization to uploaded audio, so mastering to a consistent integrated LUFS keeps your episodes from being turned up or down unpredictably.
Chosen Masters
The AI mastering engine powering DropCue's master step, offering Modern, Open, and Powerful style choices applied in seconds.
Pricing for this audience
DropCue pricing is transparent with no add-ons, no per-track fees, and no revenue share. Pro plans start at $15/mo billed annually (1,000 tracks, scaling by catalog size) and get 10 free masters and 10 free mix analyses to start. After that, mastering packs are 10 for $9.99, 25 for $21.99, or 50 for $39.99 and never expire, or 15 per month for $12; mix analysis packs are 20 for $4.99 or 50 for $9.99 and also never expire. Starter, from $5/mo annually with 500 tracks, can buy the same packs. For a podcaster shipping a handful of episodes a month, a pack you buy once and use until it runs out fits the workflow better than a recurring per-minute mastering bill, with professional studio analytics and every feature included at one transparent price.
Frequently asked questions
What LUFS should I master a podcast to?
Most spoken-word guidance lands around minus 16 LUFS integrated for stereo (Apple Podcasts and Spotify normalize toward roughly that range), with mono delivery sometimes targeted closer to minus 19 LUFS and a true peak ceiling near minus 1 dBTP. DropCue's mix analysis reports your integrated LUFS to the ITU-R BS.1770 standard and your true peak so you can confirm you are in range before uploading.
DropCue is built for music. Will it actually work for spoken word?
Yes, with the honest framing that it was designed for music creators. Loudness, true peak, mono compatibility, and phase are measured the same way regardless of whether the source is music or speech, so the analytics and mastering serve voiceover and podcast audio too. The main thing to watch is choosing a mastering style that suits a natural read, which is usually the Open style for interview-type content.
Why do my episodes sound like they have different volumes?
Almost always because each episode was exported at a different integrated loudness. Without a loudness check, one episode lands at minus 12 and the next at minus 20, and platforms normalize them inconsistently. Running every export through the same analysis and mastering target gives you a consistent loudness across the whole feed.
Should I analyze my MP3 or my original recording?
Analyze your full-quality WAV or FLAC export. For lossy, already-compressed sources like an MP3, the loudness and true peak readings can be slightly skewed versus the original and are flagged as approximate. Master at full resolution, confirm your numbers, then encode the MP3 for the platform.
Does the master fix true peak clipping that my meter misses?
The mix analysis detects true peak, including inter-sample peaks that exceed your ceiling even when a standard sample meter reads under zero. That matters because those inter-sample peaks can clip on a listener's phone or smart speaker. The analysis flags it in plain language, and the mastering pass brings the level under control before export.
Is there a per-minute or per-episode mastering fee?
No per-minute and no per-episode billing. Pro plans get 10 free masters to start, then you buy packs (10 for $9.99, 25 for $21.99, 50 for $39.99) that never expire, or a 15-per-month option for $12. You only spend a credit when you master, so a pack stretches across a lot of episodes.
