Built like a studio tool. Measured like one.
A build-up pushes everything at once: drive, width, pitch, reverb, level. That is exactly when cheap processing breaks. PREDROP 1.2 was rebuilt so that the top of your build is as clean as the start, and every claim on this page is a measurement of the real plugin.
worst inharmonic tone, 4× oversampled. A plain saturator: −19 dBc.
WIDTH at 100 %. A full-band widener: −7 dB.
even with +6 dBFS coming in and every effect at maximum.
of one core at 48 kHz (measured offline on a 2.1 GHz server core).
Saturation without the digital grit
When a saturator adds harmonics above half the sample rate, they cannot exist digitally: they fold back down as tones that are not in tune with your music. This is called aliasing, and it gets worse as the drive and the pitch climb, exactly what a riser does. PREDROP's HEAT runs at four times the sample rate, so those harmonics are filtered out before they can fold.
Spectrum of a 4.7 kHz tone through HEAT at full drive
Every grey peak that is not a multiple of 4.7 kHz is an alias: an out-of-tune tone added by the saturator. In PREDROP, they all sit below −74 dBc. Hover or tap the chart to read values.
| Measure | Plain saturator | PREDROP HEAT |
|---|---|---|
| Worst inharmonic tone | −19 dBc | below −74 dBc |
| DC offset (100 Hz tone) | yes (asymmetric drive) | 0, removed by a 7 Hz DC blocker |
| Oversampling | none | 4×, polyphase IIR half-band filters |
Hear the difference
A sine sweep from 600 Hz to 9 kHz through both saturators at the same drive, level-matched. With the plain saturator you will hear extra tones gliding the wrong way. With HEAT, only the sweep and its harmonics.
Start at a low volume: these are pure tones.
Wide highs, a solid centre
Club systems sum the bass to mono, and so do phones. A widener that also spreads the kick and the sub makes the low end weaker exactly when the drop should hit. PREDROP's WIDTH splits the signal with a steep 24 dB/oct crossover at 150 Hz and only widens what is above.
Stereo (side) signal created by WIDTH at 100 %, from a mono input
Lower is more mono. Above 1 kHz both widen the same; below 150 Hz PREDROP adds almost nothing: −39 dB at 60 Hz, −63 dB at 30 Hz.
| Frequency | Full-band widener | PREDROP WIDTH |
|---|---|---|
| 30 Hz | −6.9 dB | −62.9 dB |
| 60 Hz | −6.9 dB | −39.0 dB |
| 100 Hz | −6.9 dB | −22.6 dB |
| 150 Hz | −6.9 dB | −13.0 dB |
| 500 Hz | −6.9 dB | −7.0 dB |
| 4 kHz | −6.9 dB | −6.9 dB |
Then make the whole low end mono
In the ADVANCED panel, MONO BASS folds everything PREDROP outputs below 80, 120 or 160 Hz to mono, including stereo that was already in your track. Measured with a pure-stereo 60 Hz tone at 120 Hz: the side drops by 24.6 dB, the mids are untouched.
GROOVE handles the lows during the roll
NO KICK and KICK KEEP split the loop at 140 Hz so the sub never stutters into mud: cut it for the biggest drop contrast, or keep it playing live.
A pitch rise that stays in tune
Most simple pitch shifters crossfade between two delay taps. On sustained sounds this creates a warble: the note splits into neighbouring frequencies. PREDROP's SHIFT mode finds where the waveform matches before each splice, so the pitch stays clean. Measured at +12 semitones on a 440 Hz tone: artefacts 57 dB lower than a basic two-head shifter.
TAPE mode, the default for new instances, goes further: it plays the repeats faster, like a tape machine, with 4-point Hermite interpolation. There is no shifting at all, so there is nothing to smear.
A reverb that blooms, not rings
SPACE is an 8-line feedback delay network: four diffusion stages, slowly modulated lines and frequency-dependent damping. The result is dense and smooth, with far less of the metallic ringing heard in classic comb-filter reverbs. Measured on the decay: the tail's peaks rise 10.1 dB above its median, against 13.8 dB for a Freeverb-style design. The decay grows with the build, up to several seconds.
SPACE TAIL decides what happens on the drop: CUT closes the tail in 250 ms for a tight drop, RING lets it bloom into it.
Loud when you want it. Never clipped. Always on time.
A ceiling, not a squasher
An instant-attack safety limiter at −0.3 dBFS with a short hold and an 80 ms release. Fed +6 dBFS, the output peaks at −0.30 dBFS; below the ceiling it does nothing at all. Switch it off in ADVANCED if you prefer your own.
Parallel without phasing
The untouched signal in MIX is delayed by exactly the plugin's latency, so blending never causes comb filtering. At MIX 0 % you get your track back, sample for sample.
6 samples, compensated
The 4× oversampling costs 6 samples (0.13 ms at 48 kHz). PREDROP reports it to your DAW, which delays the other tracks to match: everything stays on the grid.
On the beat, even when you're late
PREDROP keeps a short history of the incoming audio, so an automation or a MIDI note that lands a few milliseconds late still captures from the beat.
No zipper noise
Every effect amount, the tape speed and the output stage are smoothed while you turn or automate them, so fast moves stay click-free.
Light enough for every bus
About 3.7 % of one CPU core with all seven effects at 100 %, measured offline at 48 kHz on a 2.1 GHz server core. Your own machine and DAW will differ, but not by much.
Tested before every release
A plugin you use on a live set or a deadline must never crash, never output garbage and never lose your settings.
- Passes pluginval, the industry's plugin validator, at its strictest level (10), at 44.1, 48 and 96 kHz and block sizes from 64 to 1024
- 83 automated tests render the real plugin and check timing, sound quality, presets, state recall and the interface
- Stress test: every effect at maximum for 30 seconds, output always finite and never above −0.3 dBFS
- Every setting is saved with your project and recalled exactly
- Projects made with PREDROP 1.1 open with the same sound
- Presets never change your performance settings, so switching mid-build is safe
Method: measurements come from PREDROP's offline test suite. The real plugin processes test signals at 48 kHz; spectra use a 32,768-point FFT with a Blackman-Harris window; "plain saturator" is the same tanh drive curve without oversampling; "full-band widener" adds a 13 ms delayed copy of the mid at 45 % to the side; "basic two-head shifter" crossfades two delay taps over a 35 ms window. CPU was measured single-threaded on one 2.1 GHz Xeon core.