You exported your podcast file. You listened on your headphones and it sounded fine. Then you played it back through your laptop speaker during a commute and it sounded quiet and thin. Then a listener wrote in to say the volume feels low. The same file, three different playback experiences.
This is a very common problem and it has a specific cause. Understanding LUFS, loudness normalization, and how different platforms handle your audio will help you produce files that hold up consistently.
Why the same file sounds different on different playback systems
Different listening environments have different reference levels. Headphones can deliver full bandwidth audio at reasonable volume from almost any source level. Laptop speakers are small drivers with limited low-end capability; they rely heavily on mid-range frequencies to produce perceived volume. Phone speakers are similar. Car stereos run through amplified systems with different EQ curves applied.
Your file sounds fine on headphones because headphones give the audio nowhere to hide -- you hear what is in the file. Your file sounds quiet on a laptop speaker because the laptop speaker rolls off the low and low-mid frequencies that provide much of the perceived body and volume of a voice. What is left is the high-mid and high content, which sounds thin and quiet by comparison.
This is a physical limitation of small speaker drivers, not something fixable in the file per se. What is fixable is the voice's presence in the frequency range that small speakers can actually reproduce.
What LUFS means and why it matters
LUFS stands for Loudness Units relative to Full Scale. It is an international standard for measuring perceptual loudness (defined in ITU-R BS.1770), developed specifically because peak level (dBFS) is a poor proxy for how loud audio actually sounds to humans.
Two audio files can have the same peak level but sound dramatically different in perceived loudness. A spoken word recording with lots of quiet passages has a lower average loudness than a heavily compressed music track at the same peak level. LUFS measures the average loudness over time, weighted to approximate how human hearing perceives loudness across frequencies.
Podcast platforms normalize episode loudness to a target: Apple Podcasts uses -16 LUFS, Spotify uses approximately -14 LUFS for podcasts, and most other apps follow similar targets. When you upload an episode, the platform measures its integrated LUFS and adjusts playback gain to hit the target. If your episode measures at -18 LUFS integrated, the platform turns it up 2 dB. If it measures at -14 LUFS, Apple Podcasts turns it down 2 dB.
The quiet podcast problem explained
If your podcast sounds quiet on most playback systems despite the platform normalization, the usual cause is one of three things.
High dynamic range. If your episode has large swings between quiet and loud passages, the integrated LUFS may hit the platform's target while individual quiet sections still feel low. The normalization targets the average, not the minimum. Sections where your voice drops off significantly stay quiet even after normalization.
Missing presence frequencies. If your voice recording has limited energy in the 1-4 kHz range (the "presence" range that cuts through small speakers), the audio will sound thin and low-volume on small playback systems even at correct LUFS. This is a combination of room acoustics, microphone characteristics, and EQ applied during processing.
Over-processed audio. Aggressive noise reduction can attenuate frequencies in the presence range along with the noise. The result is technically at the right LUFS but subjectively sounds quieter because the frequencies that make speech sound present and intelligible have been reduced.
How to check your file before upload
The tools for measuring LUFS are widely available. Reaper, Audacity, Adobe Audition, and almost every professional DAW have integrated loudness meters or plugins. For a quick check outside a DAW, LUFSMeter (free) or the integrated loudness function in ffmpeg will measure your file's integrated LUFS and true peak.
The numbers to check: integrated LUFS should be at your target (-16 LUFS is the safe standard across most platforms). True peak should not exceed -1 dBTP (decibels relative to full scale, true peak) -- this prevents clipping on decoding.
If your file measures at -20 LUFS and you have not applied compression, the platform will boost it by 4 dB, but the dynamic range problem will remain. The boost applies uniformly; it does not make the quiet sections relatively louder. Apply compression to reduce dynamic range before normalization.
A practical processing chain for consistent loudness
Record at healthy levels without clipping (aim for peaks around -6 dBFS during the session). Apply compression to control dynamic range (ratio 3:1 to 4:1, threshold around -20 dBFS for typical voice). Apply make-up gain to bring average level back up after compression. Apply EQ if needed to address thin sound on small speakers (slight boost in the 2-3 kHz range can improve small-speaker intelligibility without adding harshness on full-range systems). Normalize to target LUFS (-16 LUFS integrated). Check true peak.
The order matters. Normalize after compression and EQ, not before. Normalizing before compression will create a different gain structure that makes the compression settings less predictable.
Why the number alone is not enough
Hitting -16 LUFS is necessary but not sufficient for sounding loud on small speakers. The perceptual loudness at which speech is heard on a laptop depends on the content of the audio, not just its integrated level. A voice with strong presence in the 2-4 kHz range sounds louder on a laptop speaker at the same LUFS as a voice with weak presence in that range.
This is why two podcasts at the same LUFS target can sound different in level to the same listener on the same device. The LUFS number is correct in both cases; the sonic result reflects the frequency content of the voice and how much it cuts through small speakers.
The fix for this -- if you are consistently getting "sounds quiet" feedback -- is usually a moderate presence boost in EQ and tighter compression, not just targeting a higher LUFS number.
