Vocal Mastering Online — Clean, Intelligible Voice
Two presets for voice: Dry Vocal for a recording without music, and Voice for podcasts and voice-over. We even out loudness, remove rumble and lift intelligibility — no reverb, no extra colouring.
Master a vocal →Two presets — and when to use which
- Dry Vocal — target −14 LUFS. For a vocal track heading into a mix or onto a song. Compression is not pushed, exciter and reverb are off: the voice stays exactly as you recorded it, only with corrected loudness and rumble removed.
- Voice — target −16 LUFS. For podcasts, voice-over and interviews. Denser dynamically so quiet phrases don't drop out and loud ones don't bite. No reverb is added — the voice stays close.
What the processing does
- Evens out loudness to the target LUFS with true-peak headroom — the file won't be scorched by platform transcoding.
- Removes low-frequency rumble — air-conditioning hum, footsteps, microphone thumps.
- Lifts intelligibility in the range of speech consonants without turning the voice harsh.
- Keeps dynamics in check — no pumping or breathing, which becomes obvious across a long recording.
What vocal mastering does not do
- It does not remove room echo. Reverberation captured together with the voice is part of the signal. For difficult cases there is restoration with de-reverberation.
- It does not fix pitch or timing. That's an editor's job in a DAW, not mastering.
- It does not extract a voice from a finished mix. For that you need stem separation.
- It does not rescue clipping from the recording. Clipped peaks are only partly recoverable — record with headroom.
What to do before mastering — at the recording stage
Mastering doesn't repair what was broken at the recording stage, but a good recording turns it into a formality. Four things matter more than any later processing:
- Record with −12 to −6 dB of peak headroom. Clipping on the way in is irreversible: there is nothing left to restore a flattened waveform from.
- Remove sources of constant noise. Air conditioning, a laptop fan and a fridge produce a steady hum that becomes more audible the denser the processing.
- Move away from walls. Early reflections from a nearby wall add boxiness, and no equaliser fixes that.
- Keep a constant distance to the microphone. Jumps in distance mean jumps in level and tone, which a compressor only partly smooths.
Target loudness: where you publish matters
Platforms normalise loudness differently, and «making it louder» wins nothing — the extra level is simply taken back, while the lost dynamics stay lost.
- Podcast platforms — the reference is −16 LUFS, and our Voice preset lands exactly there.
- Music streaming — around −14 LUFS; for a vocal inside a song, use a music preset rather than the speech one.
- Audiobooks — quieter still, around −19 LUFS: long listening demands less fatigue. There is a separate preset for them.
- Peak headroom — in every case keep true peak at or below −1 dBTP, otherwise the platform codec will add distortion when it re-encodes.
Questions and answers
Which preset is for a song and which is for a podcast?
For a song's vocal track use «Dry vocal» targeting −14 LUFS: it adds no extra compression and no colouration. For podcasts, voice-over and interviews use «Speech» targeting −16 LUFS: denser in dynamics so quiet phrases do not drop out.
Why is the target loudness for speech quieter than for music?
−16 LUFS is the Apple Podcasts reference. Voice has a narrow dynamic range, and pushing it up to the musical −14 makes it feel pressed and tiring over a long recording.
Will processing remove room echo?
No. Reverberation is recorded together with the voice and is part of the signal — mastering does not separate it. For those cases there is separate restoration with de-reverberation.
Can I process vocals pulled out of a finished song?
You need stem separation first — it extracts the vocal track as its own file. After that you can process it with a vocal preset.
See also: Podcast mastering · Stem separation · Audio restoration · Pricing