How To High Pass A Vocal Without Thinning The Sound

The Fixed Filter Problem

A fixed high-pass filter is a compromise from the moment you use it. Set the cutoff conservatively at 60 Hz, and the rumble survives. Push it up to 120 Hz to clean things up properly, and you start thinning the lower notes of the performance. On a baritone delivering a quiet verse, that filter is doing useful work. On the same singer pushing into a chest-voice belt, that same filter is now removing some of the fundamentals. The voice loses weight exactly where it shouldn't.

Dialogue suffers the same way. A presenter moving closer to the microphone for an intimate read produces more low-end through the proximity effect. The same filter setting that sat invisibly during the wide shot now hollows out the close one. Of course, automation can help, but it's time-consuming and without close attention to detail is going to miss stuff.

The Compression Trap

There is a related problem that compounds this one. When the low end of a vocal shifts during a performance, the compressor responds. A passage with heavy fundamentals pulls more gain reduction than a passage at the same perceived loudness sitting higher in the register. The result is a vocal that ducks unpredictably, with the low passages sounding pushed back and the high passages sounding forward. The fix that often gets reached for is more EQ. Cut more low end before the compressor so the compressor stops reacting to it. Then add some back afterwards to compensate.

It works to a degree, but you have now built a corrective chain that thins the signal at one stage and tries to add it at another, and the version of the low end you put back isn't the same as the one you removed. You end up with a vocal that is consistent but lifeless. Clean but small. The performance is intact, and the levels are tidy, and somehow it does not sound like the take you nailed in the tracking room.

A Filter That Knows Where The Voice Is

The Lowend section of VoiceAssist takes a different approach. Its Pitch-Based low cut analyses the voice and sets the cutoff frequency relative to the lowest pitch the singer or speaker is producing at that moment. When the performance drops, the cutoff drops with it. When the voice climbs, the cutoff climbs too. The filter is always sitting just below the useful content of the performance, removing rumble and sub-bass material without ever crossing into the fundamentals.

This is the part that matters because the filter is not guessing at a safe number. It is tracking the actual pitch content of the audio and adjusting accordingly. A baritone delivering a verse and the same baritone belting a chorus get treated as the two different signals they are. The thinness that engineers spend so much time engineering around stops being an issue, because the filter never reaches up into the body of the voice in the first place.

For dialogue, the same principle applies to the proximity effect, register changes, and any actor with a wide vocal range. A character whispering in one scene and shouting in the next no longer needs two passes of EQ to manage the low end. The filter follows the performance through both.

Resonance That Does Not Stay In One Place

The other lower mid problem most engineers will recognise is resonance. Boxy, boomy build-ups around 200 to 400 Hz that turn up on certain syllables or certain notes and disappear on others. The traditional fix is a static EQ cut, often with a narrow Q, sometimes with a dynamic EQ if you have one to hand. The static cut helps with the problem notes, but can hollow out everything else. The dynamic EQ performs better, but it still relies on a single threshold and a single band.

Vocal resonances move; the frequency that rings out on an open vowel sitting at one pitch will not be the same frequency that rings out on a different vowel sitting at a different pitch. The room, the mic, the proximity, the phrasing, they all change where the build-up sits. A single band at a fixed frequency cannot keep up with that.

Resonance Reduction in VoiceAssist works dynamically across the lower frequency range. It identifies harsh resonances as they occur and reduces them at the frequencies where they are actually building up, on the syllables where they are actually a problem. The notes that do not need it stay untouched. This is the difference between cutting a frequency and cutting a resonance. The static EQ approach treats the whole performance as if every syllable has the same problem. The dynamic approach treats each syllable on its own terms.

What Stays In The Signal

The reason these two tools work well together comes down to what they leave alone. The Pitch-Based low cut removes the material below the performance. Resonance Reduction removes the harshness within the performance. Neither of them is affecting the body of the voice and pulling weight out of it to fix a problem that only exists in part of the take.

The signal that comes out the other side is one a compressor can deal with cleanly. The fundamentals are intact, so the compressor is not being driven by a moving low-end target. The resonances are tamed, so the EQ that follows can do tonal work rather than corrective work.

For engineers used to spending the first hour of a vocal mix building a corrective EQ chain by hand, the change in workflow is noticeable. The job stops being about removing the things that get in the way and starts being about shaping what is already there. A clean, consistent low end on a vocal track is something most mixers will tell you is harder to achieve than it sounds. The tools that try to do it with static settings rarely manage it without compromise.

Made For Your Sessions

VoiceAssist runs as an ARA plugin inside Pro Tools, Studio One, Reaper, Cubase, Nuendo, Logic Pro, and LUNA, and as an AudioSuite plugin in Pro Tools. The Lowend section sits alongside Clean, Gate, Dynamics, Breath, and Sibilance modules, all working on the same clip in the same pass. The analysis happens locally, with no cloud upload. Settings can be adjusted per region for material that needs different treatment in different sections.

For engineers working on vocals and dialogue, where the low end has always been the part that takes longer than it should to get right, the workflow change is real.

Less time spent fighting a fixed filter and more time spent on the parts of the mix that are creative and rewarding.