Multiband Processing for Guitar: Why Frequency Splits Fall Short
Multiband compressors, EQs, and stereo imagers split by frequency. Here's why that fundamentally conflicts with how guitar actually works.
Multiband processing is everywhere in modern mixing. Multiband compressors, multiband stereo imagers, multiband saturators — the idea of splitting a signal into frequency bands and processing each one independently is one of the most powerful tools in audio engineering.
For full mixes, drums, and mastering, it’s indispensable. For guitar? It has a fundamental problem that no amount of tweaking can fix.
How Multiband Processing Works
Every multiband processor follows the same basic architecture:
- Crossover filters split the input signal into 2–6 frequency bands
- Each band is processed independently (compressed, EQ’d, panned, saturated)
- The processed bands are summed back together
Common crossover points for a 3-band guitar setup might be: below 250Hz (lows), 250Hz–2kHz (mids), above 2kHz (highs). Each band gets its own set of controls.
This architecture works beautifully when frequency content maps neatly to what you want to control. On a drum bus, “lows” means kick, “mids” means snare body, “highs” means cymbals. You can compress each independently because the instruments occupy different frequency ranges.
Guitar doesn’t work this way.
The Harmonic Overlap Problem
Play a low E power chord. The open low E string vibrates at 82Hz. Its second harmonic is at 164Hz, third at 247Hz, fourth at 329Hz, fifth at 411Hz — and the harmonics extend well past 5kHz. Every one of these harmonics is part of that single note’s tone.
Now play the B string at the same time. Its fundamental is 247Hz — the exact same frequency as the low E’s third harmonic. The D string’s fundamental at 147Hz sits right next to the low E’s second harmonic at 164Hz.
A multiband crossover set at 250Hz doesn’t see “low E note” and “B string note.” It sees frequency content, all tangled together:
- Low band (below 250Hz): Low E fundamental + D string fundamental + parts of A string harmonics
- Mid band (250Hz–2kHz): B string fundamental + low E harmonics 3-8 + D string harmonics + G string fundamental + everything overlapping
- High band (above 2kHz): Upper harmonics from all six strings, impossible to separate
When you process the “mid band,” you’re simultaneously affecting the B string’s fundamental, the low E’s critical upper harmonics, and the G string’s body. There’s no way to target one note without affecting the others.
What This Means in Practice
Multiband Compression on Guitar
Compress the low band to control the bass notes, and you’re also squashing the fundamental of the D and A strings. Compress the mids to tame a resonant note, and you duck the harmonics of the bass notes — making them sound thin and hollow.
This is why multiband compression on guitar often sounds “processed” and unnatural. The compressor is applying different gain reduction to different parts of the same note’s harmonic series, warping the tonal balance of individual notes in ways that don’t happen in nature.
Multiband Stereo Imaging
Pan the highs wide and keep the lows centered. Now your low E string’s fundamental stays centered while its upper harmonics pan outward. The note sounds like it’s being torn apart — the bass is here, the brightness is over there. This isn’t natural stereo width; it’s harmonic smearing.
Meanwhile, the B string’s fundamental (at 247Hz, in the “low” band) stays centered while the B string’s harmonics pan wide. Every note in the chord is being fractured across the stereo field at its crossover boundaries.
Multiband Saturation
Saturate the low band to add warmth to the bass notes. But you’re also saturating the D string’s fundamental and generating intermodulation distortion between the low E and A string frequencies that share the band. The result is muddy distortion instead of warm harmonic richness.
The Crossover Problem
Even if you could perfectly choose your crossover frequencies (you can’t — they’d need to change for every chord), the crossover filters themselves introduce issues:
Phase shifts at crossover points. Linear-phase crossovers avoid this but add significant latency (30-50ms for steep filters — unusable in a live context). Minimum-phase crossovers are low-latency but rotate the phase at the crossover frequency, causing subtle cancellation when bands are recombined.
Steep vs. gentle slopes. Steep crossovers (24dB/oct) minimize overlap between bands but create more phase artifacts. Gentle crossovers (6dB/oct) have less phase impact but more frequency overlap between bands — reducing the effectiveness of independent processing.
Band count tradeoffs. More bands give you finer control but multiply the crossover problems and make setup more complex. Fewer bands are simpler but each band contains more overlapping note content.
Where Multiband Does Work for Guitar
Multiband processing isn’t useless on guitar — it just needs to be used for frequency-domain problems, not musical problems:
Low-cut per band: Cleaning up sub-bass rumble below the guitar’s range (high-pass the low band at 60-80Hz) is purely frequency work and multiband handles it fine.
De-essing fizzy high gain tones: If your amp sim has harsh fizz above 6kHz, a multiband compressor on just that band can tame it without affecting the mids. This works because the problem is frequency-specific, not note-specific.
Mastering context: On a full mix, multiband processing controls the overall spectral balance — kick vs. vocal vs. cymbals. Individual instrument harmonics are less relevant at the mix bus level.
Parallel processing: Blending a small amount of multiband-processed signal with the dry original can add subtle color without the artifacts of full multiband processing.
The Alternative: Note-Domain Processing
The core issue is that multiband processors divide the world by frequency, while music is organized by notes. A single guitar note spans the entire frequency spectrum — its fundamental might be 100Hz while its harmonics extend past 10kHz.
Note-domain processing flips the architecture:
- Detect individual notes — identify which frequencies belong to which musical note
- Group harmonics — keep each note’s complete harmonic series together as one unit
- Process per note — apply effects, panning, or dynamics to complete notes, not frequency slices
When you pan a note in the note domain, its fundamental and all its harmonics move together. When you add reverb to the “high notes,” you’re affecting the actual high-pitched notes — not the upper harmonics of bass notes that happen to share the same frequency range.
This is what TONIQ’s Zone FX system does. Low/Mid/High zones are defined by musical pitch, not frequency bands. The “Low Zone” contains the lowest notes in the chord — their fundamentals and all their harmonics, no matter what frequency range those harmonics occupy.
A Direct Comparison
Take a simple C major chord: C3 (131Hz), E3 (165Hz), G3 (196Hz), C4 (262Hz), E4 (330Hz).
Multiband stereo imager (3-band, crossovers at 200Hz and 2kHz):
- Low band (center): C3 fundamental, E3 fundamental, G3 fundamental — three different notes forced to the same position
- Mid band (slightly wide): C4 fundamental, E4 fundamental, plus harmonics 2-6 of C3, E3, G3 — a mess of overlapping content from all five notes
- High band (wide): Upper harmonics from all notes — harmonics of C3 panned wide while C3’s fundamental stays centered
Note-aware stereo imager:
- C3 (all harmonics, 131Hz through 5kHz+): panned slightly left
- E3 (all harmonics): panned right
- G3 (all harmonics): panned left
- C4 (all harmonics): panned center-right
- E4 (all harmonics): panned center
Each note is a complete, coherent musical voice positioned in the stereo field. No harmonics separated from fundamentals. No notes forced together because they share a frequency range.
The Bottom Line
Multiband processing solves frequency-domain problems. For guitar — a polyphonic instrument where every note’s harmonics overlap with every other note — the interesting problems are in the musical domain: which notes to pan, which notes to add reverb to, which notes to compress.
Using frequency bands to solve musical problems is like using a ruler to measure weight. The tool measures something, just not the thing you actually care about.
Written by
INSEKTIQ Team