Loudness and Delivery

You finished the mix. It sounds great on your headphones. Now it has to sound right on a podcast app, a streaming service or a TV station, and each of those has its own idea of "right." The last step of every audio job is setting the overall loudness for the destination and checking the peaks. Skip it and your podcast is either a whisper next to everyone else's or so loud the app turns it down and it sounds squashed.

Peaks are not loudness

Two files can both peak at -1 dBFS and sound completely different in volume. The peak is one instant. Loudness is how loud the whole thing feels to a human, which depends on the average level and which frequencies carry the energy. The theory page explains it with pictures at Loudness: peak, RMS and LUFS.

The measure the whole industry uses comes from a standard called ITU-R BS.1770. It filters the sound roughly the way ears hear, averages over the whole piece and ignores silences. Europeans write the unit as LUFS; the US broadcast world writes LKFS. Same measurement, same numbers. The figure you hit for a whole file is its integrated loudness.

The other number is true peak, in dBTP. Digital audio is a series of samples, and the real wave that gets rebuilt on playback can swing higher between two samples than either sample is. MP3 and AAC encoders add their own overshoot. So a file whose highest sample is -0.2 dBFS can still clip on a phone. True peak measures the rebuilt wave, which is why delivery specs ask for it.

Targets by destination

These are the numbers as of October 2026. Platforms change their specs; check the linked pages before a big delivery.

Notes on the numbers

Podcasts. Apple's spec asks for about -16 LKFS with a 1 dB tolerance and true peak not above -1 dB FS, measured per BS.1770. Most of the podcast world has settled on -16 for stereo. For a mono file, Podnews and common practice say -19 LUFS, because a mono file plays through both speakers and sounds about as loud as a -16 stereo file. Sources do not fully agree here: an Audio Engineering Society document, TD1008, recommends -18 LUFS for speech streams. -16 stereo is the safe choice for a podcast feed today.

Streaming music. Spotify's help page says it adjusts tracks to -14 LUFS, with Loud (-11) and Quiet (-19) settings for Premium listeners, and asks for true peak below -1 dBTP. A master much louder than -14 just gets turned down, and loses its punch for nothing. YouTube does not publish its number; independent measurements put it around -14.

US broadcast. The CALM Act made the FCC require TV stations, cable and satellite providers to follow ATSC's A/85 recommended practice for commercials. The FCC rule (47 CFR 73.682(e)) incorporates the 2013 edition. ATSC released a revised A/85 on July 15, 2026, which adds guidance for streaming; its target for delivered content stays at -24 LKFS, with true peak kept below -2 dBTP. Stations often send their own spec sheet. When they do, it wins.

Measuring and normalizing in SoundStudio

All of this is in the Enhance tab. Like everything in SoundStudio, nothing happens until you press a button.

1. Delivery Check. Pick a target: TV US (ATSC A/85), TV and radio Europe (EBU R128), Podcast, Music streaming, Social video or Custom. Click Check. It measures integrated loudness, true peak, sample peak, loudness range and the loudest 3 seconds, says pass or fail, and tells you how many dB to move. It changes nothing.
2. Press Use this target in LUFS Normalization to copy that target into the normalization slider.
3. Normalize to Target LUFS. With Use limiter unticked (the default), it stops short of the target if going further would push peaks past -1 dB, and tells you where it stopped. Tick Use limiter to reach the target with the limiter holding peaks at -1 dB.
4. Run Delivery Check again to confirm.

If the target is -24 with -2 dBTP, normalizing down is easy and peaks drop with it. If the target is -16 and your voice is very peaky, normalizing up may hit the peak limit first. That is a sign to revisit compression from lesson 4, not to crush it with a limiter. A Limiter in the Dynamics tab is a safety net for the last dB or two, nothing more.

Tracks matter here too. Delivery Check's As it would export setting mixes your tracks the way Export does and only counts Volume Pins you have applied. Measure the mix, not just the voice track.

Formats

Work in WAV the whole way, and make a lossy file once, at the end. Every MP3 encode throws away a little more; editing an MP3 and saving it as MP3 again stacks the damage. The theory page has the sizes at File formats.

• Your master: WAV, 24-bit, at your project rate (48 kHz). SoundStudio's Export offers WAV 16-bit, WAV 24-bit or MP3 at 128 to 320 kbps.
• Podcast feed: Apple, as of October 2026, recommends MP3 stereo at 128 to 256 kbps (mono 96 to 128) at 44.1 or 48 kHz, and says it strongly prefers AAC for RSS feeds. Most hosting services accept MP3 and many convert for you.
• Broadcast: whatever the station asks for, usually WAV at 48 kHz, 16 or 24-bit.
• Video: 48 kHz WAV into your video editor. Let the video export make the AAC.

Batches with StudioConverter

When you have a dozen episodes or a stack of spots to deliver, do the loudness and format step in StudioConverter. Its Convert audio tab has a Make it this loud setting with targets of -14, -16, -19, -23 and -24 LUFS (or Off), and a -1 dBTP peak limit by default. It measures each file first, then sets the level in a second pass. It outputs MP3, AAC (.m4a), WAV, FLAC and others, and it runs the whole queue with the same settings. It never overwrites the original.

Mind the peak limit: StudioConverter's default is -1 dBTP, while A/85 wants below -2. For a US broadcast spot, check the result in SoundStudio's Delivery Check with the TV US target before you send it.