French podcast loudness: 30 episodes measured
Public radio, studios, independent shows: how loud the podcasts you listen to really are, and what these numbers say about your own delivery.
By Hanna Eng·Audio engineer, Abbey Road Institute Paris
I measured 30 French podcasts, the latest full episode in each feed, to the EBU R128 standard. Median: -16.2 LUFS integrated, within Apple Podcasts' recommendation. 22 of 30 are within 1 LU of -16 LUFS, every file has two channels, and 18 go above -1 dBTP true peak after decoding. Public radio is the most consistent, independent shows the most variable.
Everyone knows the loudness recommendations: -16 LUFS for Apple Podcasts, a true peak of -1 dBTP. But how loud do French podcasts actually come out? Rather than repeat a target, I measured 30 well-known shows, the latest full episode of each, on the file listeners actually receive. Here are the results, the method and the limits, with the full data to download.
The study in numbers
| Measure | Result |
|---|---|
| Podcasts measured | 30 (10 public radio, 10 studios and private media, 10 independent) |
| Median integrated loudness | -16.2 LUFS |
| Full range | -19.7 to -13.0 LUFS |
| Within 1 LU of -16 LUFS | 22 of 30 |
| Louder than -14 LUFS | 1 of 30 |
| True peak above -1 dBTP | 18 of 30 (6 above 0 dBTP) |
| Mono files | 0 of 30 (all two-channel) |
| Measured on | 21 November 2026 |
Measurements by Hanna Eng, ffmpeg 8.0 (ebur128 filter). Full data as CSV: /data/podcast-loudness-fr-2026.csv
How were the episodes measured?
I picked 30 well-known, varied podcasts: 10 from public radio (France Inter, France Culture), 10 from studios and private media (Binge Audio, Louie Media, Nouvelles Écoutes, Slate, Europe 1, RTL, Le Monde) and 10 independent shows. For each one, I took the latest full episode from the official RSS feed, skipping trailers, previews of other series and episodes under 10 minutes. 28 episodes date from October or November 2026; FloodCast and 2 Heures de perdues have not published since 2025 and 2024.
The measurement follows the EBU R128 standard (ITU-R BS.1770 method), using ffmpeg 8.0's ebur128 filter: integrated loudness, loudness range (LRA) and true peak. It runs on the MP3 or AAC file as served by the feed, before any normalization by Spotify or Apple. The files were deleted after measuring; only the numbers are published.
How loud are French podcasts?
The median lands at -16.2 LUFS integrated, almost exactly the target Apple Podcasts recommends. 22 of 30 episodes are within 1 LU of -16 LUFS. The full range runs from -19.7 to -13.0 LUFS, and only one episode is louder than -14 LUFS, Spotify's normalization level for music.
In other words, the -16 LUFS target isn't theory: it is already what most shows do. The outliers are mostly independent productions.
- Public radio10 of 10
- Studios8 of 10
- Independent4 of 10
Public radio, studios, independents: who is most consistent?
Public radio is remarkably consistent: its 10 episodes sit between -16.7 and -15.8 LUFS, all delivered as AAC at 192 kbps and 48 kHz. Studios and private media are on target too, with two exceptions (-19.7 and -17.1 LUFS), delivering MP3 at 128 kbps and 44.1 kHz.
Independent shows vary the most: from -17.6 to -13.0 LUFS. The three loudest episodes in the study are all independent. I don't know their production chain: these numbers show the level, not the intent.
- Public radio-16.7 to -15.8 LUFS
median -16.2 LUFS
- Studios-19.7 to -15 LUFS
median -16.4 LUFS
- Independent-17.6 to -13 LUFS
median -15.7 LUFS
- Apple Podcasts target (-16 LUFS)
- Spotify, music (-14 LUFS)
| Producer | Median (LUFS) | Range (LUFS) | Within 1 LU of -16 | True peak above -1 dBTP |
|---|---|---|---|---|
| Public radio | -16.2 | -16.7 to -15.8 | 10 of 10 | 2 of 10 |
| Studios and private media | -16.4 | -19.7 to -15.0 | 8 of 10 | 6 of 10 |
| Independent | -15.7 | -17.6 to -13.0 | 4 of 10 | 10 of 10 |
Medians computed over the 10 episodes in each group.
Do true peaks go above -1 dBTP?
Often: 18 of 30 episodes go above -1 dBTP true peak, 6 of them above 0 dBTP. Every independent episode goes above -1 dBTP.
An important detail: the peak is measured after decoding the MP3 or AAC. Encoding creates overshoots the original master may not have had. That is exactly why you keep headroom before encoding: a master that touches 0 dBFS can come out above it once encoded.
Mono or two channels, and which file format?
None of the 30 files is mono: every file has two channels, even for a single voice. Public radio delivers AAC at 192 kbps and 48 kHz; studios and independent shows deliver MP3, almost all at 128 kbps and 44.1 kHz. One episode is delivered at 32 kHz, an unusual sample rate for a podcast.
All the results, episode by episode
The full table, loudest to quietest. The loudness range (LRA) shows how much the level moves over the episode: low for a steady interview, higher for a documentary edit with ambience and music.
| Podcast | Producer | Integrated loudness (LUFS) | Range (LU) | True peak (dBTP) | File |
|---|---|---|---|---|---|
| Le Podcast de Pauline Laigneau | Independent | -13.0 | 4.9 | +4.0 | MP3 128 kbps, 44.1 kHz |
| Change ma vie (Clotilde Dusoulier) | Independent | -14.1 | 4.3 | -0.2 | MP3 128 kbps, 44.1 kHz |
| FloodCast | Independent | -14.4 | 5.7 | +0.9 | MP3 128 kbps, 44.1 kHz |
| L'Heure du Monde (Le Monde) | Studio or private media | -15.0 | 6.0 | 0.0 | MP3 128 kbps, 44.1 kHz |
| 2 Heures de perdues | Independent | -15.3 | 9.4 | +0.4 | MP3 128 kbps, 44.1 kHz |
| Thinkerview | Independent | -15.3 | 10.5 | -0.9 | MP3 128 kbps, 44.1 kHz |
| Le Code a changé (France Inter) | Public radio | -15.8 | 4.1 | -1.7 | AAC 192 kbps, 48 kHz |
| Hondelatte raconte (Europe 1) | Studio or private media | -15.9 | 2.6 | +0.8 | MP3 128 kbps, 44.1 kHz |
| Avec philosophie (France Culture) | Public radio | -16.1 | 6.1 | -0.6 | AAC 192 kbps, 48 kHz |
| Le Masque et la Plume (France Inter) | Public radio | -16.1 | 4.5 | -1.2 | AAC 192 kbps, 48 kHz |
| Génération Do It Yourself (Matthieu Stefani) | Independent | -16.1 | 3.9 | 0.0 | MP3 128 kbps, 44.1 kHz |
| Affaires sensibles (France Inter) | Public radio | -16.2 | 4.3 | -1.3 | AAC 192 kbps, 48 kHz |
| Les Pieds sur terre (France Culture) | Public radio | -16.2 | 6.4 | -1.9 | AAC 192 kbps, 48 kHz |
| LSD, la série documentaire (France Culture) | Public radio | -16.2 | 7.6 | -1.2 | AAC 192 kbps, 48 kHz |
| La Science, CQFD (France Culture) | Public radio | -16.2 | 8.5 | -0.1 | AAC 192 kbps, 48 kHz |
| Sismique (Julien Devaureix) | Independent | -16.2 | 3.7 | 0.0 | MP3 128 kbps, 44.1 kHz |
| Les Couilles sur la table (Binge Audio) | Studio or private media | -16.3 | 3.7 | +0.6 | MP3 128 kbps, 44.1 kHz |
| La Terre au carré (France Inter) | Public radio | -16.4 | 2.7 | -1.4 | AAC 192 kbps, 48 kHz |
| Le Cœur sur la table (Binge Audio) | Studio or private media | -16.4 | 4.8 | -0.6 | MP3 128 kbps, 44.1 kHz |
| La Poudre (Nouvelles Écoutes / Lauren Bastide) | Studio or private media | -16.4 | 9.2 | -1.1 | MP3 128 kbps, 44.1 kHz |
| L'Heure du crime (RTL) | Studio or private media | -16.4 | 2.8 | -1.8 | MP3 128 kbps, 44.1 kHz |
| Émotions (Louie Media) | Studio or private media | -16.5 | 5.5 | -0.7 | MP3 128 kbps, 44.1 kHz |
| Zoom Zoom Zen (France Inter) | Public radio | -16.5 | 3.6 | -1.8 | AAC 192 kbps, 48 kHz |
| Passages (Louie Media) | Studio or private media | -16.7 | 6.0 | -1.3 | MP3 128 kbps, 44.1 kHz |
| Le Cours de l'histoire (France Culture) | Public radio | -16.7 | 5.5 | -2.4 | AAC 192 kbps, 48 kHz |
| Au cœur de l'histoire (Europe 1) | Studio or private media | -17.1 | 2.3 | -0.9 | MP3 128 kbps, 44.1 kHz |
| Vlan! (Grégory Pouy) | Independent | -17.3 | 4.3 | -0.4 | MP3 128 kbps, 44.1 kHz |
| Passion Médiévistes (Fanny Cohen Moreau) | Independent | -17.4 | 6.5 | +3.2 | MP3 192 kbps, 32 kHz |
| Les Baladeurs (Les Others) | Independent | -17.6 | 3.7 | -0.6 | MP3 128 kbps, 44.1 kHz |
| Transfert (Slate Audio) | Studio or private media | -19.7 | 5.2 | -3.3 | MP3 128 kbps, 44.1 kHz |
Measured with ffmpeg 8.0, ebur128 filter (EBU R128 / ITU-R BS.1770), on the file served by the official RSS feed, latest full episode in each feed, on 21 November 2026. Full data: /data/podcast-loudness-fr-2026.csv
What these numbers mean for your podcast
Aim for -16 LUFS integrated in stereo: it is Apple Podcasts' recommendation, and it is what most of the shows measured already do. Keep the true peak under -1 dBTP on the encoded file, not only on the master. And measure every episode: consistency like public radio's suggests systematic measurement, not a magic setting.
If your podcast comes out louder or quieter than its neighbors, listeners notice at the first switch between shows. I mix and master podcasts to these targets, measured before every delivery: send me an episode and I'll tell you where it sits in this table.
Frequently asked questions
How many LUFS should a podcast be?
-16 LUFS integrated in stereo, with a true peak under -1 dBTP: that is Apple Podcasts' recommendation. Across 30 French podcasts measured, the median is -16.2 LUFS.
Do French podcasts hit -16 LUFS?
Mostly yes: 22 of 30 are within 1 LU of -16 LUFS. Public radio scores 10 of 10, independent shows 4 of 10.
Why does the true peak go above 0 dBTP on an MP3?
MP3 or AAC encoding creates peak overshoots that may not exist in the master. That is why you keep headroom: a -1 dBTP ceiling before encoding, and a check on the encoded file.
How do I measure my podcast's loudness?
With a meter that follows ITU-R BS.1770, or with ffmpeg and its ebur128 filter, which gives integrated loudness, range and true peak. Measure the final file, the one you publish.
Can I reuse this data?
Yes. The full CSV file is available to download; cite the source and the measurement date, 21 November 2026.