LUFS for YouTube and Podcasts: Loudness Targets

Mix podcasts to about -16 LUFS integrated with true peaks no higher than -1 dBTP: that's what Apple Podcasts recommends. YouTube publishes no loudness target; creators widely observe that it turns videos louder than about -14 LUFS down and leaves quieter ones alone, so -14 is a sensible upper aim for video. Spotify publishes -14 LUFS for music normalisation, and broadcast uses -23 (EBU R 128, Europe) or -24 (ATSC A/85, US).

Below: what LUFS and true peak mean, a table of targets with the source for each, and step-by-step normalising in FFmpeg and Audacity. Checked on 1 November 2026 on Apple Podcasts for Creators, Spotify, YouTube Help, the EBU, ATSC and ITU documents, the FFmpeg docs and the Audacity manual. More in our video formats, codecs and numbers guides.

The short answer

Podcast loudness standard: -16 LUFS integrated, plus or minus 1 dB, true peak at most -1 dBTP (Apple Podcasts). YouTube has no official LUFS target; the widely observed reference is about -14 LUFS, and YouTube turns louder uploads down rather than quiet ones up. Spotify normalises music to -14 LUFS but its podcast spec sets no loudness target. Broadcast: -23 LUFS (EBU R 128) or -24 LKFS (ATSC A/85). Measure and fix with FFmpeg's loudnorm filter or Audacity's Loudness Normalization.

Key takeaways

• LUFS measures perceived loudness over the whole programme; LKFS is the same unit under its ITU name.
• Apple Podcasts: -16 LKFS plus or minus 1 dB, true peak no higher than -1 dBFS, measured per ITU-R BS.1770-5.
• YouTube publishes no target; about -14 LUFS is the observed reference, and quieter videos aren't turned up.
• Spotify: -14 LUFS for music normalisation (Loud -11, Quiet -19); its podcast delivery spec has no loudness figure.
• Normalise to loudness, not peak: peak normalising only guarantees nothing clips, not that two shows sound equally loud.

What is LUFS?

LUFS (Loudness Units relative to Full Scale) measures how loud audio sounds to people, not just how high its peaks are. It's defined in ITU-R BS.1770, now at version 5 (November 2023), which calls the same unit LKFS; the EBU notes the two are equivalent. A change of 1 LU is a change of 1 dB.

True peak matters because lossy encoders can push a peak above the sample values you saw in your editor. That's why Spotify advises keeping masters below -1 dBTP, and lower (-2 dBTP) for masters louder than -14 LUFS.

Loudness targets by platform

YouTube

YouTube's Help Center doesn't publish a loudness target. Creators who read the "content loudness" line in Stats for nerds report a reference of about -14 LUFS, with louder videos turned down and quieter ones left as they are; treat that as an observation, not a rule. Separately, YouTube's stable volume feature, on by default and documented for Android phones, tablets and TVs, evens out quiet and loud parts during playback; YouTube says it's off for YouTube Music and official music videos.

Podcasts

Apple asks for -16 LKFS with a 1 dB tolerance and true peaks no higher than -1 dBFS, measured per BS.1770-5, and says to do this before encoding, since encoding can clip a signal that breaks the true-peak limit. Spotify's podcast delivery spec, which covers formats and metadata, has no loudness figure, so -16 works there too. AES TD1008, a 2021 recommendation for streaming and on-demand audio that became AES77-2023, puts speech at -18 LUFS and mixed speech and music at -17 according to published summaries; we couldn't open the AES document itself.

One file for YouTube and podcast apps?

A -16 LUFS, -1 dBTP master meets Apple's spec and plays on YouTube about 2 dB quieter than -14 LUFS videos, which most viewers won't notice. If YouTube is your main outlet, make a separate -14 version. Setting up a show that goes to both is covered in how to start a video podcast.

How to measure loudness

You measure integrated loudness with a BS.1770 meter. Most editors include one; FFmpeg has one built in:

ffmpeg -i episode.wav -af ebur128=peak=true -f null -

At the end, FFmpeg prints a summary with Integrated loudness (I, in LUFS), Loudness range (LRA) and the true peak. FFmpeg's docs describe the filter's short-term (3 s) and momentary (400 ms) readings and note that its on-screen target band defaults to -23 LUFS, the EBU figure.

How to normalise to -16 or -14 LUFS

With FFmpeg (two passes)

1. Measure: ffmpeg -i in.wav -af loudnorm=I=-16:TP=-1.5:LRA=11:print_format=json -f null - and note input_i, input_tp, input_lra and input_thresh.
2. Apply with those values: ffmpeg -i in.wav -af loudnorm=I=-16:TP=-1.5:LRA=11:measured_I=-23.4:measured_TP=-4.2:measured_LRA=6.1:measured_thresh=-33.8:linear=true -ar 48000 out.wav (replace the four measured numbers with yours).
3. Measure the result again with the ebur128 command above.

FFmpeg's docs explain the details: loudnorm's defaults are -24 integrated and -2 true peak, so set I and TP yourself; linear mode applies one gain change when the measured values allow it and falls back to dynamic mode otherwise; and in dynamic mode it upsamples to 192 kHz, which is why -ar 48000 sets the output rate. A TP of -1.5 leaves margin under Apple's -1. For a video, keep the picture and re-encode only the sound: -c:v copy -c:a aac -b:a 192k.

With Audacity (free)

1. Select the whole track.
2. Choose Effect > Volume and Compression > Loudness Normalization.
3. Pick perceived loudness and change the target from the default -23 LUFS to -16 (podcasts) or -14 (YouTube).
4. Decide on "Treat mono as dual-mono": it's on by default and makes a mono track measure 3 LU louder than many other meters would, so a mono episode normalised this way plays quieter in apps that measure it as plain mono.
5. Apply, then check true peak in another meter and add a limiter if peaks exceed -1 dBTP; the Audacity manual doesn't say this effect limits peaks.

Common loudness mistakes

• Peak normalising and calling it done. Raising a file until its highest peak hits -1 dB doesn't make two episodes equally loud. QuickVeed's free audio volume booster works this way: it boosts or normalises the peak to -1 dB with a soft limiter, which is handy for a quiet recording but doesn't measure LUFS, so check the result with a meter.
• Normalising after encoding. Apple says to do it before encoding to MP3 or AAC; normalise the WAV, then encode.
• Mixing music beds too loud. Integrated loudness averages everything, so a loud music bed lets the speech end up quieter than the target.
• Forgetting the clips. Short clips cut from an episode with Podcast to Clips or any editor keep the episode's level; they don't need a separate pass unless you change the mix.

If the source is simply too quiet, how to make audio louder covers gain, compression and limiting, and the free audio extractor pulls a WAV from a video so you can work on the sound alone. Sample rates are in 44.1 kHz vs 48 kHz.

Frequently asked questions

What LUFS should a podcast be?

-16 LUFS integrated, plus or minus 1 dB, with true peaks at or below -1 dBTP, per Apple Podcasts. It works on Spotify too, which publishes no podcast target.

What LUFS does YouTube use?

YouTube doesn't publish one. Creators observe about -14 LUFS, with louder videos turned down; aim for -14 to -16 with peaks under -1 dBTP.

What is the Apple Podcasts loudness standard?

-16 LKFS with a 1 dB tolerance and a true-peak maximum of -1 dBFS, measured per ITU-R BS.1770-5.

Is LUFS the same as LKFS?

Yes. LKFS is the ITU's name and LUFS the EBU's; the EBU's R 128 notes they're equivalent.

Does Spotify normalise podcasts?

Spotify's loudness article covers music (-14 LUFS by default) and its podcast spec has no loudness figure, so Spotify doesn't publish a podcast target.

What is -23 LUFS?

The EBU R 128 broadcast target for European TV and radio, and the default in Audacity's and FFmpeg's EBU-based tools. It's too quiet for podcasts and YouTube.

Try it with your own video

Recording a video podcast? Turn each episode into captioned vertical clips. Upload an episode and let QuickVeed find clips, frame the speaker and burn in captions. QuickVeed plans start at $18/month, and new accounts get a 14-day trial with 250 credits, no card needed. Sign up or see pricing.

Sources

Competitor details in this post were checked against these public pages in November 2026. Products and prices change, so confirm with each vendor before you decide.

• Apple Podcasts for Creators: audio requirements (loudness, formats)
• Spotify for Artists: loudness normalization
• Spotify for Creators: podcast specification doc (links the delivery spec v1.10)
• YouTube Help: control video volume (stable volume)
• ITU-R BS.1770-5 (11/2023): algorithms to measure loudness and true-peak level
• EBU R 128-2023: loudness normalisation and permitted maximum level
• ATSC A/85:2013: techniques for establishing and maintaining audio loudness
• Telos Alliance: understanding loudness for streaming audio (summary of AES TD1008)
• FFmpeg filters documentation (loudnorm, ebur128, tonemap, zscale)
• Audacity manual: Loudness Normalization