BECOS FX Side-Chain Filter (LO-SCF)

The LO-SCF sits in the sidechain only. It changes what the compressor listens to, not the audio itself: it sets how easily each note, low or high, triggers compression.

The filter: a low shelf in the sidechain. Turning the LO-SCF away from the centre turns the low range up (BOOST) or down (CUT), by up to about ±12 dB at the bottom. The further you turn it, the higher the 3 dB corner moves: from about 50 Hz just outside the centre to about 400 Hz at the ends. Close to the centre the curve stays within ±3 dB, so there is no corner to show.

CUT (knob turned clockwise, above 50 %): the low range, below the corner, is turned down before the detector. Low notes need more level before triggering compression, so the compressor stops pumping on the low end, and the output keeps more of its lows. The green line is labelled SC CUT.

BOOST (knob turned anticlockwise, below 50 %): the low range is turned up before the detector. Low notes trigger compression sooner and harder, so the low end is held down more under compression, tighter and more even. The green line is labelled SC BOOST.

NEUTRAL (50 %): the sidechain hears the full range as it is. Double-click the slider to return there.

On the plot

Green line: the filter at the selected knob position; dashed grey lines: knob positions 0 / 25 / 50 / 75 / 100 % for comparison.

Corner (green dotted line, dot on the curve, label above the plot): where the curve is 3 dB away from its level at 2 kHz; − for cut, + for boost. Near NEUTRAL the curve stays within ±3 dB, so there is no corner.

50 Hz and 400 Hz lines (grey dotted, shaded between): the range the corner moves through. 90 Hz marker (yellow dashed): an arbitrary reference to read the amount of cut or boost, from 0 dB at the centre to about ±12 dB at the ends.

Audio response of the output of compressor: the mirror of the filter curve, taken from the treble level – blue where the output is freed up, red where it is clamped down, pink near 0; labelled AUDIO FREED UP / AUDIO CLAMPED DOWN. It is the upper limit: it holds only while the compressor is reducing gain, and the real amount is smaller, about (1 − 1/ratio) of it, roughly half at the ~2.2:1 slope. Below threshold the filter has no effect on the audio. The plot below shows the real effect on music.

Example – bass, CUT: with the lows cut by about 12 dB in the sidechain, the low E and A strings need about 12 dB more level before triggering compression, so their fundamentals stay big and open, and the mids (pick attack, growl) set how hard the compressor works.

Counter-example – guitar, BOOST: palm-muted low chugs start triggering compression early and hard, so the low end of a rhythm guitar is held tight and even, while lead lines higher up the neck pass with less compression. Boosting works like lowering the threshold, but only for the boosted frequencies, so even a thin-sounding guitar or weaker pickups start triggering compression in the lows.

What this plot shows – the cause (what triggers the compressor) and the effect (what changes at the output), for real music. We sampled two real recordings, a clean bass clip (about 21 s) and a clean guitar clip (about 16 s), and the plot uses their measured spectrum and dynamics, not a made-up example.

Bass / Guitar buttons (top right of the knob panel): choose which clip the plot uses. The bass puts most of its energy in the range the SCF acts on, so it reacts strongly to the knob; the guitar sits mostly higher (250 Hz – 1 kHz), so the same setting changes its compression less.

The boxes (NEUTRAL → with SCF action): average gain reduction over the clip; gain reduction on its loud moments (95th percentile); and how much of the time the compressor is working (gain reduction above 1 dB).

Spectrum profile (always shown). All charts share the frequency axis of the plot above (20 Hz – 20 kHz) and the same dB scale:

• Dark grey dashed line (over a light grey area): the clip's level in each third-octave band at its loud moments, at NEUTRAL, against the compressor's threshold at NEUTRAL (faint orange flat line, 0 dB). It never moves: it is the 'normal' state of the music.

• Thick orange curve – the cause: the effective threshold with the SCF action, following the filter's slope; its label (THRESHOLD: SCF CUTTING / BOOSTING) sits on the line at the right. CUT raises it in the lows, so low notes must be louder to trigger compression; BOOST lowers it there. Where the dashed line is above the orange curve, that band triggers compression; where it ends up below, it is freed up.

• Blue/red band and line – the effect: the solid line is the spectrum with the SCF action, blue where it comes out louder, red where quieter; the shaded band between it and the dashed line is the change.

• Bottom row: the octave ranges (31 Hz … 8 kHz+), separated by faint vertical lines, with their output change in dB – the same numbers as the bar charts.

• Green dotted line: the SCF's 3 dB corner, as in the plot above (its own tick box below the chart; none near NEUTRAL). Grey dotted lines and shading: 50 – 400 Hz, the corner's range; yellow: 90 Hz reference.

• Hover: a crosshair with the frequency, the clip level, the threshold and the output change at that point.

Magnify the output change ×2 (off by default): the changes are small next to the spectrum's range; tick it to draw the arrows and the blue/red band twice as long. Spectrum, bars and threshold stay at their true positions, and every value label and the hover readout stay in real dB.

Display octave bar frequency bands (extra chart, on by default): the same picture grouped into octaves, on the same axes. Each grey bar covers its octave's real frequency span; its height is the average level of the spectrum line over that span, so the dashed line (exactly the spectrum line from the chart above) runs across the bar tops. The orange dashed curve is the effective threshold. The arrow is the output change: it starts at the bar's top, and its tip marks the new level (blue up = louder, red down = quieter), with the value in dB next to the tip.

Separate chart: audio output change (extra chart, on by default): only the output change per octave range, as bars from the NEUTRAL 0 dB line, with the orange dashed line showing how far the SCF moves the threshold at each frequency, in dB from NEUTRAL – cause and effect on one scale. The axis runs from −16 to +16 dB to show the threshold's full range and every value.

Show the output change (on by default): switches the blue/red band and the arrows on or off.

Note by note (on by default): each frequency range (and each band of the profile) is measured on the moments when it is loud itself – its own top 10 % – instead of averaged over the whole clip. That is how the ear judges it: a low note is heard when it sounds, and the compressor ducks less exactly then with CUT (more with BOOST), so the lows move more than the highs, which are often heard between the low notes. Untick it for the plain average over the clip.

Frequency ranges used in this plot (octaves, as in the bar charts and the bottom row): 31 Hz = 22–45 Hz · 63 Hz = 45–90 Hz · 125 Hz = 90–180 Hz · 250 Hz = 180–355 Hz · 500 Hz = 355–710 Hz · 1 kHz = 710 Hz–1.4 kHz · 2 kHz = 1.4–2.8 kHz · 4 kHz = 2.8–5.6 kHz · 8 kHz+ = 5.6–20 kHz. The keep buttons use: Lows = 63 + 125 Hz (45–180 Hz) · Mids = 500 Hz + 1 kHz + 2 kHz (355 Hz–2.8 kHz) · Highs = 4 kHz + 8 kHz+ (2.8–20 kHz).

Make-up gain matching method: one fixed gain after the compressor, the same for every frequency, so on its own it doesn't tilt the sound – but where you set it decides where the 0 dB line sits. Changing the SCF changes how much the compressor ducks, so the whole output gets louder (CUT) or quieter (BOOST); in real life you re-set the make-up gain by ear. Unchanged: kept as at NEUTRAL (raw view). Fletcher-Munson: uses the ear's equal-loudness curves (ISO 226, the modern Fletcher–Munson curves) at the chosen listening level – Quiet 65, Normal 80 (default) or Loud 95 dB SPL for the loudest band; in the lows each dB of change sounds like more than a dB (up to about 1.6 phon per dB at normal and loud levels), while at quiet levels the deepest lows (around 30 Hz) fall toward the edge of hearing and count less: the make-up gain is set so the total perceived loudness (in sones) matches NEUTRAL, and all values are shown as perceived change (phon) instead of dB – in the deep lows a dB of change sounds bigger, so the freed lows show more. Zwicker: the full loudness model of ISO 532-1 (Zwicker, stationary, free field): the sound is split into critical bands, each band's level spreads upward and masks what sits above it (louder lows mask more), the ear's transmission and low-frequency corrections are applied, and the specific loudness of each band is computed from what remains audible. The make-up gain is set so the total loudness (sones) matches NEUTRAL, and the values are shown as loudness-level change (phon equivalent: +10 = twice as loud). Same listening level as Fletcher-Munson; checked against a reference ISO 532-1 implementation. LUFS: same loudness as at NEUTRAL by the broadcast LUFS measure (ear-weighted average, values in dB). keep: Lows / Mids / Highs (on top of the loudness correction): set the trim so that region reads 0 – lows = the 63 and 125 Hz octaves (about 45–180 Hz), mids = the 500 Hz, 1 kHz and 2 kHz octaves (about 350 Hz–2.8 kHz), highs = the 4 kHz and 8 kHz+ octaves (from about 2.8 kHz up – for bass and guitar, that is already the top end), as when you judge by that part of the sound; the trim slider moves to the value found. Click again to release it; moving the slider by hand also releases it. Fletcher-Munson is the default. Trim by ear (±6 dB, double-click to reset) adds a little more on top, e.g. to free the lows. The box 'Make-up gain change' shows the total. The model computes the shape of the change (the tilt between lows and highs); the make-up setting only moves it up or down.

What to notice: the orange line moves only in the lows, but the output changes in every range. The filter changes what triggers the compressor; the compressor then turns the whole signal up or down. The lows usually move a little more, because they play mostly in the moments that changed. A change is not a level: +3 dB on a band below the line doesn't mean it now triggers; it means it comes out 3 dB louder.

Simplifications: the detector reacts to all bands together, so 'a band above the line triggers' is a guide, not an exact rule. Model: the clip cut into 20 ms slices, measured in third-octave bands, passed through the SCF and a simple detector (fixed slope, instant attack, short release); the threshold is set so the loud moments get the chosen gain reduction at NEUTRAL. Not a circuit simulation.