You are on the train, earbuds in, an interview podcast running, and every third sentence dissolves into wheel noise. You turn the volume up. The host’s voice gets louder, and so does the rumble underneath it — you still miss the joke. Play a song on the same earbuds in the same seat and it sounds fine. Dialogue is the one thing that keeps losing to the cabin.
Why the default mix loses to the train
A podcast is mixed for a quiet room, where a voice sits comfortably above the noise floor. A train or bus cabin adds a wall of broadband, low-frequency noise: wheels, engine, HVAC, road. That noise piles up exactly where a voice’s low end and body live, and it masks the consonants — the s, t, f, k sounds — that carry most of what makes speech intelligible, since consonants are already quieter than vowels even in a clean recording. Raising the volume raises the noise floor and the voice together. It doesn’t change which one wins.
The fix, step by step
- In ZonaReq, start from a flat curve or a plain voice preset, not a music one — most music presets add sub-bass, which is exactly the range working against you here.
- Cut everything below 120 Hz. Almost no speech information lives there; it’s rumble and headroom, and removing it clears space for what’s left.
- Lift the 2 to 4 kHz band by 3 to 4 dB. This is the presence range where consonants sit, and it’s the fastest way to make dialogue legible without adding overall loudness.
- If the voice still sounds thick or boxy, trim 300 to 500 Hz by a couple of dB. That range holds warmth and muddiness in roughly equal parts.
- Save the result as its own preset — “train” or “commute” — so it’s one click instead of five sliders every time you get on.
Is the free tier enough
Not quite, for the precise version. ZonaReq’s free tier gives you a 3-band system equalizer, which can approximate this — bass down, treble up — but it can’t isolate 120 Hz from 300 Hz, or hold a narrow presence lift at 2 to 4 kHz without also dragging up everything above it. If your podcast app already does dialogue enhancement — Apple Podcasts’ Enhance Dialogue and Overcast’s Voice Boost both do a version of this — that may be all you need, and it costs nothing. The system-wide EQ earns its place when you want the same fix everywhere: a browser tab, a downloaded episode, a call recording, without waiting for each app to have solved it on its own.
If it’s not quite right
If the voice starts to sound harsh or sibilant, the presence lift is too aggressive — pull the 4 kHz band back by 1 to 2 dB before touching anything else. If dialogue still gets lost under a particularly loud stretch of track, add a little more cut around 300 Hz rather than more volume; muddiness, not loudness, is usually what’s left standing in the way.