Some singers just sing slightly under the note, and no amount of EQ will fix it. I get that kind of take regularly, and I still don’t reach for pitch correction. Not out of principle. I just find that a tuned vocal often stops sounding like the person who walked into the live room.
What I do instead is make the mix work harder for the vocal, and there’s more room there than most people think. Here’s the process I run when the take is competent but flat, the performance is right, the pitch is just shy.
Slow compression first, pitch later (probably never)
Flat sang vocals usually have no dynamic life in them either. The singer was concentrating on the note, not the phrase, so everything comes out at one even level with no push and pull. That’s the real problem, not the cents.
I put a compressor on it with a low ratio and a slow attack, maybe 2:1 or 3:1, and let it take gentle peaks down by a dB or two. The idea is to make the quiet words louder, not squash the loud ones. Mastering engineers lean on this exact approach for keeping a mix polished without killing it, and a vocal benefits from the same restraint. Gain makeup afterward does more for a flat vocal than any pitch tool because it gives the words somewhere to push against.
FabFilter Pro-C 2 is what I use here. Its Clean style gets out of the way, and if the acoustic nature of the track is bringing out harshness, I’ve found backing off the compression on the direct signal and running a parallel copy gives a cleaner result. It’s the one compressor I keep on every vocal chain; the attack and release are the controls I actually touch, and everything else is set and forgotten.
Carve a pocket, don’t boost the vocal
Broad, gentle EQ moves. I take care of the obvious problems first with a narrow cut somewhere in the high mids if that range is stinging, then use a second gentler EQ for a small air lift up top, around 10kHz or so, if the vocal is dull. Small numbers. A flat vocal with 5dB of presence boost just sounds like a flat vocal with a cold.
The other half of this is subtraction on everything around it. Guitars, keys, whatever the arrangement has, get a scoop in the range where the vocal lives. The vocal then sits without needing to be turned up. That’s the move that actually rescues these takes.
Double tracking is the real cheat code
If the singer can come back in, this is what I’d do before touching a pitch correction plugin. Recording the same performance a second time and layering it in gives width, depth and presence that no amount of processing on a single take will reproduce. The small timing and pitch differences between the two takes are what make it sound human, which is exactly the opposite of what Autotune does to a vocal.
In the mix, I tuck the double just under the main vocal, panned slightly, volume adjusted so it thickens without competing. On choruses it’s the difference between a section that lands and one that doesn’t. For a flat-sounding lead, a well-blended double is worth more than any plugin on this list.
If the singer can’t come back, artificial doubling plugins get you most of the way for a demo. Not identical to the real thing, but usable.
Attack the sibilance, not the pitch
Flat vocals often come with brittle esses, because the singer is over-articulating to hold the note. That harsh top end reads as pitchiness even when the pitch is close, and fixing it changes how the whole vocal is perceived.
A de-esser set so it only catches the offending sounds is cleaner than a broad EQ cut, since you’re only acting on the moments that need it. If the sibilance is stubborn, a dynamic tool that reacts to the incoming signal and only takes down the problematic frequencies works better than a multiband compressor, which will pull down whatever is in the band whether or not it’s actually an “S”. Soothe 2 is the one I reach for; I keep the processing limited to the upper part of the spectrum and the attenuation small, and it does things to a harsh vocal that nothing else does. I put it before any saturation so I’m not making the harshness more audible.
Where saturation helps and where it lies
A little harmonic distortion on a flat vocal can add life, but only a little. Tape-style settings thicken the low end and pull the word forward slightly. The temptation is to hear that and add more, and the vocal stops sounding like the singer. I’ll add it in parallel, or split the processing into bands so I keep control of where the harmonics land, and keep the wet level well under the dry.
What I don’t do
I don’t reach for pitch correction on a good performance. I’ve heard tuned vocals that sound like they were carved from another singer’s face, and I’d rather have a slightly under-the-note take that still sounds like the person. A close-tuned vocal also sits worse in a dense mix, because the tuning fills the same pocket the rest of the arrangement is fighting for.
If the singer is more than a little under the note, that’s a different conversation. One elastic pass at the worst three or four notes, plus the compression and the double, will get you there without the artifact becoming the story.
One note of caution on the compression: I keep the attack slow enough that consonants survive. A flat vocal run through a fast compressor loses its consonants and sounds muffled, which is worse than flat. Check the top end after every move.
That’s the chain. Slow compressor, gentle EQ with a pocket carved around it, de-esser, maybe a parallel double, small saturation. It won’t make a bad take good. It will make a half-decent take sit properly in a mix, which is most of what I’m doing in a session anyway.