k2k audio Sculpt the frequency spectrum like shaping clay on a potter’s wheel. Cut what you don’t want, boost what you do, and let resonances ring or vanish.
Multimode Filter
What it does — Runs your signal through one of five classic filter algorithms, each with its own character, from surgical Butterworth to screaming 303 acid.
When you’d reach for it — You need a filter and you want to pick the personality. Clean for transparent mix work, Moog for warm analog sweeps, 303 for acid bass lines, Diode for gritty saturation, ZDF for fast automated cutoff movements that stay stable.
Quick example
- Drop a Multimode Filter after your audio source.
- Set the algorithm to Moog and the mode to LP (lowpass).
- Bring the Cutoff down to around 800 Hz.
- Push Resonance to 0.7 for that self-oscillation bloom.
- Automate Cutoff with an envelope for a classic filter sweep.
Parameters
| Parameter | What it controls | Range | Sweet spot hint |
|---|---|---|---|
| Algorithm | Filter model: Clean, Moog, 303, Diode, ZDF | 5 choices | Moog for warmth, ZDF for modulation |
| Mode | Filter shape: LP, HP, BP, BR (Clean adds Notch, Peaking, Lo Shelf, Hi Shelf) | 4-8 choices | LP is home base |
| Cutoff | Corner frequency | 20 - 20 000 Hz | 800 Hz for bass warmth, 3 kHz for presence |
| Resonance | Peak emphasis at cutoff | 0 - 100% | 50% for color, 85%+ for self-oscillation |
| Drive | Pre-filter saturation | 0 - 100% | 20% for subtle grit |
| Steep | Cascade two stages for 24 dB/oct slope (Clean only) | On / Off | On for steep surgical cuts |
| Gain | Boost/cut at target frequency (Clean Peaking/Shelf modes) | dB | 0 dB neutral |
| Dry/Wet | Parallel blend with unfiltered signal | 0 - 100% | 100% for full effect |
Parametric EQ
What it does — Splits the spectrum into four adjustable bands (low shelf, two sweepable mids, high shelf) and lets you boost or cut each one independently.
When you’d reach for it — You want broad tonal shaping across the whole spectrum: tame a muddy low end, carve out a midrange pocket, add some air on top. The interactive frequency-response display makes it easy to see what you are doing.
Quick example
- Connect your source into the Parametric EQ.
- Pull the Low Gain shelf down by -3 dB to tighten the bass.
- Set Mid 1 Freq to 400 Hz and cut Mid 1 Gain by -4 dB to clear mud.
- Bump High Gain up +2 dB for some sparkle.
Parameters
| Parameter | What it controls | Range | Sweet spot hint |
|---|---|---|---|
| Low Freq | Shelf corner for bass band | Hz | Set once, leave around 80-120 Hz |
| Low Gain | Boost or cut below shelf corner | -18 to +18 dB | Subtle moves: +/- 3 dB |
| Low Q | Shelf slope steepness | Q value | Lower Q for gentle rolloff |
| Mid 1 Freq | Center frequency of first peaking band | 100 - 4 000 Hz | 300-800 Hz for body |
| Mid 1 Gain | Boost or cut at Mid 1 center | -18 to +18 dB | Cut mud at -3 to -5 dB |
| Mid 1 Q | Bandwidth of first band (scroll wheel in widget) | Q value | Narrow Q for surgical, wide for tonal |
| Mid 2 Freq | Center frequency of second peaking band | 500 - 16 000 Hz | 2-5 kHz for presence |
| Mid 2 Gain | Boost or cut at Mid 2 center | -18 to +18 dB | +2 dB presence lift |
| Mid 2 Q | Bandwidth of second band | Q value | Match width to the problem |
| High Freq | Shelf corner for treble band | Hz | 8-12 kHz for air |
| High Gain | Boost or cut above shelf corner | -18 to +18 dB | +2 dB for sheen |
| High Q | Shelf slope steepness | Q value | Lower Q for smooth shelf |
| Mix | Parallel blend with dry signal | 0 - 100% | 100% for standard EQ use |
Comb Resonator
What it does — Creates pitched resonances by feeding a delay line back into itself, producing metallic, robotic, or string-like tones from any input.
When you’d reach for it — You want to add a tonal, pitched ring to a sound. Use Comb mode for metallic robot vibes, String mode for plucked Karplus-Strong textures, and Waveguide mode for more natural resonance with damping that mimics a physical string decaying over time.
Quick example
- Feed a noise burst or percussion hit into the Comb Resonator.
- Set mode to String and Frequency to 220 Hz (A3).
- Push Feedback to 95% for a long, plucked decay.
- Set Mix to 50% to keep some of the original attack.
Parameters
| Parameter | What it controls | Range | Sweet spot hint |
|---|---|---|---|
| Mode | Resonator type: Comb, String, Waveguide | 3 choices | String for plucked tones |
| Frequency | Pitch of the resonance | 20 - 2 000 Hz | Tune to a musical note |
| Feedback | How long the resonance rings | 0 - 99.9% | 90% for a medium decay, 98%+ for infinite drones |
| Damping | Lowpass cutoff inside the feedback loop (mostly affects Waveguide) | 500 - 16 000 Hz | 5 000 Hz for natural string damping |
| Mix | Blend between dry input and resonated signal | 0 - 100% | 50% default keeps the attack intact |
Envelope Resonator
What it does — Places up to eight resonant peaks across the spectrum and scales their intensity in real time based on how loud the input signal is.
When you’d reach for it — You want resonances that breathe with the performance. A quiet passage stays clean; a loud hit rings out through the peaks. Set Env Depth to 0% and you get a static resonant EQ with no dynamic behavior, useful on its own.
Quick example
- Connect a vocal or drum loop to the Envelope Resonator.
- Click in the frequency-response widget to place two peaks: one around 800 Hz, another at 3 kHz.
- Set Env Depth to 75% so the peaks only fully open on loud transients.
- Shorten Attack to 5 ms and lengthen Release to 300 ms for a punchy, ringing tail.
- Adjust each peak’s Q by scrolling over its handle (tighter Q = sharper ring).
Parameters
| Parameter | What it controls | Range | Sweet spot hint |
|---|---|---|---|
| Attack | How fast peaks respond to rising level | 0.1 - 200 ms | 5-15 ms for percussive response |
| Release | How fast peaks fade after level drops | 10 - 1 000 ms | 100-300 ms for natural decay |
| Sensitivity | Shifts the detection level up or down | -24 to +24 dB | 0 dB default; raise for quiet sources |
| Env Depth | How much the envelope scales peak gain (0% = static resonance) | 0 - 100% | 75% for dynamic, 0% for fixed EQ |
| Detection Mode | Level measurement: Peak or RMS | 2 choices | RMS for smoother tracking |
| Min Freq | Low bound of detection range | 20 - 20 000 Hz | 20 Hz to detect full spectrum |
| Max Freq | High bound of detection range | 20 - 20 000 Hz | Narrow to focus on a specific band |
| Peaks (widget) | Frequency, gain, and Q of each resonance (up to 8) | Freq 20-20k Hz, Gain -24 to +24 dB, Q 0.5-20 | Place peaks at formant frequencies |
| Mix | Blend between dry and resonated signal | 0 - 100% | 100% for full effect |
| Output | Final output level trim | -12 to +12 dB | 0 dB unless compensating |
Adaptive Notch
What it does — Automatically detects up to four prominent spectral peaks and follows them with narrow notch filters, removing them even as they drift in frequency.
When you’d reach for it — A recording has hum, feedback whine, or a mechanical resonance that wanders in pitch. Instead of placing a static notch and hoping, this node tracks the offending frequency and stays locked on it frame by frame.
Quick example
- Feed the noisy recording into the Adaptive Notch.
- Set Notches to 2 (one for the fundamental hum, one for its harmonic).
- Lower the Threshold to -30 dB so quieter hum lines get detected.
- Narrow the Width to 15 Hz to avoid eating surrounding content.
- Listen and adjust Depth until the hum vanishes without artifacts.
Parameters
| Parameter | What it controls | Range | Sweet spot hint |
|---|---|---|---|
| Notches | Number of independent tracking filters | 1 - 4 | 1 for single hum, 2-3 for harmonics |
| Width | How wide each notch cuts | 1 - 200 Hz | 15-30 Hz for hum, wider for feedback |
| Depth | How deep the notch attenuates | -60 to 0 dB | -40 dB removes most problems cleanly |
| Speed | How fast the notch follows a moving frequency | 0.01 - 1.00 | 0.5 balances tracking and stability |
| Threshold | Detection sensitivity (relative to loudest peak) | -60 to -6 dB | -24 dB default; lower for faint hum |
| Persistence | Minimum frames a peak must appear before the notch activates | 1 - 20 frames | 5 prevents false triggers |
| Min Freq | Lower bound of the detection range | 20 - 20 000 Hz | 30 Hz to catch mains hum |
| Max Freq | Upper bound of the detection range | 100 - 20 000 Hz | 16 000 Hz default |
| Mix | Blend between original and notched signal | 0 - 100% | 100% for full removal |
| Gain | Output level trim | -24 to +24 dB | 0 dB unless compensating |
Monophonic Noise Filter
What it does — Detects the pitch of a monophonic source frame by frame, builds a harmonic grid, and separates the tonal content from everything inharmonic (breath, noise, bow scrape).
When you’d reach for it — You have a solo vocal, flute, cello, or any single-pitched source and you want to isolate just the clean harmonics or just the breathy, noisy texture. The node outputs both sides so you can process them independently and recombine later.
Quick example
- Connect a solo vocal recording to the Monophonic Noise Filter.
- Set the mode to Both Outputs to get harmonics and noise on separate outputs.
- Lower Tolerance to 30 cents for a tighter harmonic grid.
- Raise Softness to 0.7 for a gentle crossover between harmonic and noise.
- Route the noise output through a separate chain (e.g., gentle reverb on breath).
Parameters
| Parameter | What it controls | Range | Sweet spot hint |
|---|---|---|---|
| Mode | What comes out of the primary output: Noise Only, Harmonics Only, or Both Outputs (complement on second port) | 3 choices | Both Outputs for maximum flexibility |
| Tolerance | How close a frequency must be to a harmonic to count as tonal | 5 - 200 cents | 50 cents default; tighten for cleaner split |
| Softness | Mask crossfade between harmonic and noise regions | 0 - 1 | 0.5 for balanced; 0 for hard cut |
| Harmonics | Number of harmonic partials in the grid | 4 - 64 | 16-32 covers most instruments |
| Min F0 | Lowest expected pitch | 20 - 500 Hz | 50 Hz for bass voice, 200 Hz for flute |
| Max F0 | Highest expected pitch | 100 - 4 000 Hz | 2 000 Hz covers most solo sources |
| Confidence | Minimum pitch-detection confidence to accept a frame | 0.10 - 0.90 | 0.3 default; raise if tracking jumps |
| Median | Pitch-smoothing window size (odd numbers) | 1 - 15 frames | 7 for stable tracking |
| Smooth | Exponential smoothing on pitch contour | 0 - 0.95 | 0.7 for natural legato |
| Mix | Blend between dry and separated output | 0 - 100% | 100% for full separation |
| Gain | Output level trim | -24 to +24 dB | 0 dB unless compensating |