You searched for how to mix vocals in hip hop because your vocal never sounds like part of the record. It either floats on top of the beat like it was recorded in a different room, or it drowns the second the instrumental gets busy, and turning the fader up just swings you between the two. The vocals on records you reference do neither. They sit in the mix: locked to the beat, every word readable, no fader anxiety. That sound is an order of operations, not a magic plugin, and this guide walks through it with exact numbers. The beat side of the equation, 808s, kick, and the low end, is covered in the full hip-hop mixing guide; this one is the vocal deep-dive.
Clean the vocal first: high-pass, mud cut, de-ess
Every problem you fail to remove here gets amplified by every processor after it, so cleanup comes before compression, always. Start with a high-pass filter: 70 to 100 Hz for a male rap vocal, 100 to 140 Hz for a female vocal. Nothing useful lives below those points, only rumble, plosive thumps, and mic handling noise that will fight the 808 for headroom. Sweep the filter up until the voice starts thinning, then back off slightly.
Next, check 200 to 400 Hz. If the vocal sounds boxy or woolly, like the rapper is cupping the mic, cut 1 to 3 dB there with a wide Q. Skip the cut if the vocal already sounds clear. Then de-ess before you compress. Sibilance lives between 5 and 8 kHz, and a compressor placed first will squash the loud words while leaving the esses untouched, which makes them relatively louder. Set the de-esser so it only triggers on actual esses, pulling 3 to 6 dB on the worst offenders, and the rest of the chain inherits a vocal that is already polite.
How to compress a vocal: two stages, never one slam
A rap vocal swings 10 dB or more between a shouted line end and a tucked-in triplet run, and one compressor working hard enough to control that range pumps audibly. So the way to compress a vocal here is two compressors in series, each doing a light job, instead of one slamming the whole range. Stage one evens out the performance: ratio 3:1, attack 10 to 20 ms so the consonants punch through before the compressor grabs, release 40 to 80 ms, and a gentle 2 to 4 dB of gain reduction that moves almost constantly. This is the stage that glues the performance into one consistent level.
Stage two sets the level and catches what slips past: ratio 4:1, attack 1 to 5 ms, threshold set so you see 2 to 4 dB of gain reduction on the loudest words and almost none on the quiet ones. This stage exists only to stop the spikes that the first compressor let through. Two compressors each pulling a few dB will always sound more transparent than one slamming 8 dB, because neither of them ever has to work hard enough to be heard working. The same two-stage method is how you compress the vocal and mix voices in pop, R and B, or melodic genres too; only the de-essing and reverb amounts change with the style.
Carve the beat, not the vocal
Here is the move most tutorials skip: the space for the vocal does not come from the vocal. It comes from the instrumental. Intelligibility lives between 1 and 4 kHz, and a busy beat full of synth leads, hi-hat wash, and sample harmonics occupies exactly that range. When the vocal will not cut through, the instinct is to boost the vocal at 3 kHz, then boost again, until it cuts through by being harsh. Resist that. Instead, cut the instrumental 1 to 2 dB between 1 and 4 kHz with a wide bell, and the vocal steps forward without you touching it.
The better version of the same move is a dynamic EQ on the beat bus, sidechained to the vocal. Set a band at 1 to 4 kHz that dips 1 to 2 dB only while the rapper is rapping, then releases the moment the line ends. The beat stays full in every gap and steps back during every bar. That ducking is why pro hip-hop mixes sound dense and clear at the same time: the vocal and the beat never actually fight, they take turns.
Presence and air without the sibilance tax
With the beat carved, the vocal needs far less EQ than you think. Presence comes from a 1 to 3 dB shelf or wide bell between 3 and 6 kHz, just enough that consonants read at low volume. Air comes from a 1 to 2 dB high shelf at 10 to 14 kHz, which adds the expensive sheen without touching the harsh zone below it. Make both moves small and check them at a quiet listening level, because presence boosts that feel right loud are usually 2 dB too hot.
If the vocal starts spitting after the air shelf, the answer is the de-esser threshold, not less air. Lower the threshold 2 to 3 dB so it catches the esses the shelf lifted, and you keep the brightness without the pain. If the whole top end of your track hurts, not just the vocal, that is a different problem with its own fixes, covered in why your mix sounds harsh. The fastest way to know whether your vocal brightness is actually competitive is to upload the track to TrackSensei and compare your high-frequency balance against released hip-hop records instead of guessing on your own monitors.
Hip-hop vocals run dry: space that stays out of the way
Compared to pop or melodic genres, hip-hop vocals carry very little reverb, and what is there is engineered to vanish. Use a short plate, 0.6 to 1.2 seconds, mixed at 8 to 15 percent wet, with 40 to 80 ms of pre-delay. The pre-delay is the secret: it keeps the front of every word completely dry, so the diction stays sharp while the tail adds a sense of room behind it. Without pre-delay the reverb smears the consonants and the vocal slides backward in the mix.
A mono slap delay at 80 to 120 ms works as an alternative or an addition, giving thickness without any audible tail. And instead of a constant wash, automate delay throws on line ends: push the send up for the last word of a phrase, let it echo into the gap, pull it back down. The vocal stays dry where the words are and spacious where they are not.
Doubles tucked 6 to 10 dB under, ad-libs treated as percussion
Doubles thicken the lead only when you cannot hear them as separate voices. Tuck them 6 to 10 dB under the lead, pan them 30 to 70 percent left and right, and high-pass them harder than the lead, up around 150 to 200 Hz, so they add width without adding mud. Compress them more aggressively than the lead too, because their dynamics only need to support, never to perform.
Ad-libs are percussion, mix them like it. Pan them off-center, filter them aggressively at both ends, throw more delay or reverb on them than you would ever put on the lead, and place them in the pockets between lead phrases. An ad-lib that collides with a lead line is an arrangement problem, so mute or move it rather than fighting it with EQ.
Ride the fader last: the 1 to 2 dB moves no compressor makes
After all the processing, automate the vocal level phrase by phrase. Push the quiet conversational lines up 1 to 2 dB, pull the shouted hooks down 1 to 2 dB, lift the one word that carries the punchline. Compressors react to level; a fader ride responds to meaning, and that is why every professional rap vocal you have ever heard was ridden by hand on top of the compression. Thirty minutes of automation at the end of the mix does more for vocals sitting in the mix than any plugin purchase.
Run the chain in this order, cleanup, two-stage compression, carving the beat, presence, space, doubles, automation, and the floating or drowning problem resolves into a vocal that lives inside the record. The honest check at the end is whether your numbers actually land where released tracks land, and that is hard to hear on the same monitors that created the problem. If you want measurements instead of guesses, upload a bounce to TrackSensei and read the feedback against the hip-hop reference profile. It will tell you within two minutes whether your vocal presence range, low-mid buildup, and overall balance match the records you are aiming at.