KERN PUSH vs FabFilter Pro-C 3: different tools for different jobs

the problem broadband compression cannot solve

you know the move. drum bus into Pro-C 3, Bus algorithm, 3:1 ratio, auto-gain on. the body tightens, the snare sits, the mix sounds more finished. then you notice: the cymbals are pumping against the kick in the 2-5 kHz range, and the low-end thump is triggering gain reduction that ducks the entire spectrum. you reach for the sidechain EQ. it helps. but the detector is still reading one number for the whole signal.

broadband compression sees level. one detector, one gain-reduction value, applied everywhere from 20 Hz to 20 kHz. when a kick drum hits hard at 80 Hz, the compressor pulls down the cymbals at 8 kHz by the same amount. that is the design working as intended. it is how broadband dynamics behave, and it is why bus compression sounds like bus compression.

spectral compression is a different problem. KERN PUSH splits the signal into 40 bands spaced on the Equivalent Rectangular Bandwidth (ERB) scale[^2], then compresses each band with its own detector and its own gain computer. the kick drives gain reduction in the low bands, the cymbals are handled in theirs. they are not fully independent, and that is deliberate: PUSH blends a share of the energy-weighted broadband gain reduction back into every band (60% on SETTLE, 40% on TIGHT, 15% on DRIFT) so the result still behaves like a compressor rather than a fast dynamic EQ. what you avoid is the kick dictating the gain of the whole spectrum.

what Pro-C 3 does well

Pro-C 3 earned its position. 14 styles running from Clean and Mastering transparency through Opto, Op-El, Vari-Mu and Vocal to Bus, Pumping, Upward and the multiband TTM. the visual feedback is a real strength: real-time gain-reduction trace, interactive transfer curve, and a sidechain EQ of up to 6 bands using the Pro-Q 4 filter shapes. Dolby Atmos support up to 9.1.6. auto-gain helps you level-match an A/B, though FabFilter describe it as an educated guess rather than measured loudness matching, so check it by ear. character modes add colour without leaving the plugin.

if you need one compressor for everything, Pro-C 3 is the right answer. its spectral work is limited to the multiband TTM style; the rest is broadband, and it is very good at it. i am not going to hand you a number proving it is the best, because nobody publishes one. the Pro-C line has been a default on a lot of desks for a long time, and that is the honest form of the claim.[^1]

what PUSH does differently

PUSH is not 14 flavors of broadband. it is three characters of spectral compression, each built for a different job across 40 ERB bands.

SETTLE is bus glue. gentle per-band downward compression with slow release curves. it does what you want a bus compressor to do, but per-band: each frequency range settles into place independently. the low end does not fight the top. think of it like a well-run meeting where everyone gets heard, instead of one loud voice setting the volume for the room. (that is the one analogy. i will stop.)

TIGHT is transient grip, and it runs the opposite way round to what the name suggests. crest factor modulates the ratio per band: when a transient arrives the ratio drops by up to 20%, so the attack gets through, and it climbs back for the sustain behind it. the per-band transient gate pushes the same idea further, pulling gain reduction back by 85% at onset. so when a snare hits, the crack comes out intact and the tail is what gets held. that behaviour is borrowed from FET compressors like the 1176, where fast programme-dependent response lets the attack past. PUSH is not an emulation of one: an 1176 has switched ratios and a single broadband detector, where this is a continuously varying ratio computed separately in each of 40 bands.

DRIFT is per-band upward compression. quiet spectral content gets lifted toward the median, filling in the gaps between peaks without raising the peaks themselves. Pro-C 3 does upward compression too: it has an Upward style, and TTM combines upward and downward across multiple bands. what it does not have is upward compression running across 40 ERB bands with per-band release, so this is a difference of resolution rather than a thing only one of us can do.

the positioning is simple: PUSH is the step before your compressor. it pre-shapes the spectral dynamics so your broadband compressor only has to handle the broadband job. Pro-C 3 glues. PUSH shapes what goes into the glue.

the comparison table

(this is where most comparison guides get dishonest. the table says PUSH is cheaper, so PUSH wins. no. the table says they solve different problems at different price points. read the “when to use which” section.)

every figure in that table is read off the vendor’s own page, checked 3 Aug 2026. where a row says “not published”, that is because the vendor does not publish it, and i am not going to invent a number to fill a cell. FabFilter and sonible price in euros, so the dollar figure moves with the exchange rate.

when to use which

get PUSH if

• you want spectral dynamics on a bus or stem before your broadband compressor catches the rest. the per-band processing tames frequency-specific peaks that a single detector cannot see.
• you mix in Ableton or Logic and do not need AAX or Atmos.
• you want upward compression running per band across 40 ERB bands (DRIFT). Pro-C 3 has an Upward style and a multiband TTM, so this is about resolution rather than existence.
• budget matters. $29 is not $199.

get Pro-C 3 if

• you need a very versatile broadband compressor. 14 styles, character modes, Atmos, AAX, sidechain EQ, A/B snapshots. it covers a wide range of broadband work in one plugin.
• you work in Pro Tools and need AAX. PUSH does not ship AAX.
• you need Dolby Atmos up to 9.1.6. not even close.
• you want zero-latency tracking. Pro-C 3 runs at 0ms with oversampling and maximum lookahead both off. PUSH LIVE mode reports ~23ms.

get both if

• you mix drums, dense arrangements, or anything where spectral and broadband dynamics serve different purposes. PUSH first, Pro-C 3 second.
• you want to A/B the two approaches on hard material and pick the cleaner result per source.

the honest gap

what PUSH lacks: no lookahead (spectral processing uses a fixed FFT window instead). no AAX. no Atmos. no sidechain EQ. no external sidechain input. three characters against fourteen styles, and no per-style transfer-curve editing. the UI is four knobs, a character selector and a spectrum. if you need the Swiss-army compressor, PUSH is not it.

what Pro-C 3 lacks: no per-band spectral processing on a psychoacoustic scale (TTM is a handful of crossover bands, not 40 ERB bands). no upward compression at that resolution. FabFilter do not describe anything like PUSH’s cepstral gain smoothing[^3], though what a plugin does not advertise is not proof of what it does not do. and it costs about 7x more for a different job. that price gap is not a value judgment. Pro-C 3 earns its price with scope. PUSH earns its price with specificity.

frequently asked questions

frequently asked questions

is PUSH a real alternative to Pro-C 3?

they solve different problems. PUSH is spectral pre-shaping across 40 ERB bands. Pro-C 3 is mostly broadband dynamics, across 14 styles, one of which (TTM) is multiband. most sessions benefit from both. PUSH at $29 does not replace a $199 broadband compressor. it handles the spectral work the broadband compressor cannot.

what does Pro-C 3 do that PUSH cannot?

14 compression styles, a sidechain EQ of up to 6 bands using the Pro-Q 4 filter shapes, up to 32x oversampling, Dolby Atmos up to 9.1.6, AAX and AudioSuite for Pro Tools, CLAP, and adjustable lookahead up to 20ms. Pro-C 3 is the Swiss-army compressor. PUSH is a specialist.

what does PUSH do that Pro-C 3 cannot?

per-band compression across 40 psychoacoustic ERB bands. per-band upward compression via the DRIFT character. spectral-aware gain smoothing via cepstral DCT. program-dependent ratio on transients via the TIGHT character. and it costs $29 vs $199.

can i use PUSH and Pro-C 3 together?

yes. the recommended chain is PUSH first (spectral pre-shaping), then Pro-C 3 (broadband glue). PUSH tames per-band peaks so Pro-C 3 only has to handle the broadband dynamics. the result sounds more relaxed than either alone.

does PUSH have a low-latency mode for tracking?

yes. PUSH v1.1.0 added an opt-in LIVE mode at ~23ms reported latency. toggle it on for tracking, off for mixing. Pro-C 3 can run at zero latency with both oversampling and maximum lookahead switched off.

references

a note from the developer

i have had a FabFilter compressor on my mixes for years, and Pro-C 3 (which shipped in January) is the one there now. it sits on the drum bus, on the vocal chain, sometimes on the master. it is good at what it does, and what it does is broadband dynamics.

PUSH sits before it.

the idea was never to replace Pro-C 3. it was to handle the part of the dynamics problem that broadband compression is structurally blind to: frequency-dependent peaks that a single-number detector averages away. a kick drum at 80 Hz should not duck your cymbals at 8 kHz. that is just physics. the broadband detector sees one number. the spectral detector sees 40.

i built PUSH at $29 because the spectral compression step should not cost $199. it is a focused tool. four knobs, three characters, one job. the honest version of this comparison is that most sessions benefit from both: PUSH to pre-shape the spectral dynamics, then Pro-C 3 (or whatever broadband compressor you already own) to glue the result. they are complementary, not competing. if you only buy one compressor plugin this year, buy Pro-C 3. if you already own a broadband compressor and want to hear what spectral pre-shaping does to your buses, try the PUSH demo. your ears will tell you whether $29 is worth it.