> ## Documentation Index
> Fetch the complete documentation index at: https://docs.sampler.meiji.industries/llms.txt
> Use this file to discover all available pages before exploring further.

# Effects And Parameter Reference

> Reference ranges, defaults, and intended behavior for channel-level sound shaping.

# Effects and parameter reference

This page documents the main channel-level sound controls in Meiji Sampler.

## Channel defaults

| Parameter      | Default      | Typical meaning                                                |
| -------------- | ------------ | -------------------------------------------------------------- |
| Gain           | `1.0`        | neutral level                                                  |
| Saturation     | `0.0`        | no added drive                                                 |
| Pan            | `0.0`        | center                                                         |
| Width          | `0.0`        | neutral stereo width                                           |
| Pitch          | `0.0`        | no pitch shift                                                 |
| HPF            | `20.0 Hz`    | effectively off for most material                              |
| LPF            | `20000.0 Hz` | effectively open                                               |
| Reverb         | `0.0`        | dry                                                            |
| Offset         | `+0ms`       | no trigger delay                                               |
| ASR-10 Boost   | `Off`        | no boost processing                                            |
| Duck source    | `None`       | disabled                                                       |
| Duck amount    | `1.5`        | moderate ducking                                               |
| Duck ratio     | `6.0`        | medium-strong compression ratio                                |
| Duck threshold | `0.05`       | fairly sensitive trigger                                       |
| Duck attack    | `0.0025 s`   | fast response                                                  |
| Duck release   | `0.26 s`     | moderate recovery                                              |
| SP1200 Filter  | `Off`        | Off / Unfiltered (7 & 8) / Filtered (1 & 2) / Filtered (3 & 4) |

## Core tone controls

### Gain

* documented range in product copy: `0.0x` to `2.0x`
* default: `1.0`

Use it to set basic channel level before deeper processing.

### Saturation

* default: `0.0`

Adds harmonic density and edge. Useful for warming drums or pushing a sound forward.

### Pan

* range: left to right
* default: center

Use pan to separate elements that compete in the same frequency range.

### Width

* default: neutral

Controls stereo spread. Use carefully on material that already has strong stereo information.

### Pitch

* product copy range: about `+/-15%`
* default: `0.0`

Useful for quick tuning, character changes, and time-feel variation.

## Filters

Both filters use a 6th-order Butterworth design inspired by the Akai S950's MF6-50 switched-capacitor filter. The implementation cascades three biquad sections (Direct Form II Transposed) with the Butterworth Q factors, producing a maximally flat passband and a steep 36 dB/octave rolloff. At the cutoff frequency the signal is attenuated by exactly -3 dB; below that point the response is virtually flat, and above it the slope falls away sharply.

Cutoff changes are smoothed in the log-frequency domain with a 5 ms time constant, so sweeping the filter produces no zipper noise or clicking. This matches the behavior of the S950's clock-tracking hardware filter.

<Info title="The chip that shaped a genre">
  In 1988, Akai tucked a National Semiconductor MF6-50 into the S950 and accidentally changed music forever. The MF6-50 was a switched-capacitor filter. Instead of traditional resistors, it used tiny capacitors flipping on and off thousands of times a second, their cutoff frequency slaved to whatever clock you fed them.

  Akai wired that clock directly to the DAC playback rate through an Intel 8254 timer, so pitching a sample down would drag the filter cutoff with it. No software, no interpolation, just raw analog tracking. The result was a 6th-order Butterworth wall: maximally flat until the cutoff point, then a 36 dB/octave cliff.

  Producers in Queens and East London figured out that if you sampled a record hot, the preamp would soft-clip and the filter would round the top end off in exactly the right way. That crunch became the backbone of golden-era hip-hop, jungle, and early house.

  The math hasn't changed since. Three cascaded biquad sections, Q factors of 0.518, 0.707, and 1.932, now running in floating point instead of silicon.
</Info>

### HPF

* range: `20 Hz` to `2000 Hz`
* default: `20 Hz`

The high-pass filter removes low-frequency content. Use it to clean mud, carve space for a kick, or build tension by thinning a sound before a drop.

### LPF

* range: `200 Hz` to `20000 Hz`
* default: `20000 Hz`

The low-pass filter removes high-frequency content. Use it to darken sounds, create lo-fi textures, or shape transitions. Sweeping the LPF down mimics the classic muffled-sample sound of vintage hardware samplers.

## Reverb

* default: `0.0`

Reverb adds space and depth using Airwindows CreamCoat behind the existing channel control. The saved `0.0` to `1.0` value maps directly to CreamCoat Wetness rather than a linear dry/wet crossfade, so higher values add more ambience without treating `50%` as half-dry. Use sparingly on already dense material so it does not blur transients.

## PerformFX Airwindows presets

PerformFX includes selected Airwindows DSP compiled directly into Meiji Sampler. These are built-in performance presets, not loadable external plugins.

The A/B audition lane can compare these current presets with official permissive CLAP candidates. Official Meiji-distributed candidates must use MIT, ISC, BSD, Apache, or an equivalent permissive OSS license; proprietary, GPL, LGPL, unknown, and ambiguous licenses are excluded from official packs.

Airwindows-backed preset families include:

* space: Room Verb, Hall Throw, Cathedral, Plate Throw, Retro Verb, Bright Room, Infinite Space
* delay: Dub Echo, Ping Pong, Tape Echo, Space Echo, Pitch Echo, Four Tap Echo, Double Delay
* modulation: DP4 Lush, Daft Flanger, Stereo Chorus, Ensemble Wash, AutoPan, Tube Tremolo
* color: Vinyl Sim, Tape Warmth, Iron Oxide, Dusty Vinyl, Pocket Lofi, Bit Glitter, Derezzed, Crunchy Wear
* DJ FX: DJ Isolator, Cap Sweep, Gate Filter, Digital Gate, Beat Surge, Block Limiter, Digital Blackout
* pitch and siren throws: Glitch Shift, Nasty Pitch, Air Raid Echo, Benidub Air Raid

The exact channel HPF, LPF, SP1200, ASR-10 Boost, ducking, pan, width, S950-style filter behavior, Motor Kill, and Tape Stop controls keep their existing engines and semantics. New Airwindows PerformFX presets sit alongside them for live gestures.

## Playback offset

| Parameter | Default | Range              | Step  |
| --------- | ------- | ------------------ | ----- |
| Offset    | `+0ms`  | `+0ms` to `+100ms` | `1ms` |

Offset delays the start of pad and chop playback for one channel. It behaves like inserting silence before the sample starts, but it does not change the sample file, trim points, chop markers, or recorded loop event timing.

Offset is applied before the channel's normal audio processing. The delayed sound still goes through Gain, Sat, ASR-10 Boost, Duck, Pan, Width, Pitch, SP1200, filters, and Reverb.

For looped samples, Offset delays only the initial trigger. Once the loop starts, the repeated audio cycles at the original loop length instead of adding the offset again on every pass.

Linked pads keep independent Offset values. If channel 1 links to channel 2, channel 1 plays with channel 1's Offset and channel 2 plays with channel 2's Offset. The linked target does not inherit the source channel's Offset.

Examples:

| Setup                                    | Result                                                     |
| ---------------------------------------- | ---------------------------------------------------------- |
| Source snare `+0ms`, linked clap `+12ms` | Clap lands after the snare for a loose backbeat            |
| Source kick `+0ms`, linked sub `+0ms`    | Both layers stay locked together                           |
| Source hat `+0ms`, linked shaker `+20ms` | Shaker trails the hat while the loop event remains on-grid |
| Chop on channel 4 with Offset `+7ms`     | Every chop hit on that channel starts 7 ms later           |

<Info title="Swing and human feel">
  Offset is useful when quantization is too clean. Add small delays to selected layers to create swing, flam, and human timing differences while keeping the recorded pattern easy to edit.
</Info>

## ASR-10 Boost

* values: `Off` / `Lo` / `Mid` / `Hi`
* default: `Off`

ASR-10 Boost is a vintage signal chain modeled after the Ensoniq ASR-10's hardware Boost function. It applies three stages of processing in series at three intensity levels:

| Level | Gain   | Character                                    |
| ----- | ------ | -------------------------------------------- |
| Lo    | +4 dB  | Subtle warmth and harmonic richness          |
| Mid   | +8 dB  | Moderate saturation, the classic ASR-10 feel |
| Hi    | +12 dB | Aggressive, original hardware level          |

All three levels share the same DSP chain:

1. **Digital excitation**: gain followed by tape-style hysteresis saturation. This adds even-order harmonics and warmth without harsh digital clipping.
2. **Soft-knee limiting**: an exponential soft-clip stage that compresses peaks while adding density. Prevents output from exceeding bounds.
3. **Analog output filter**: a single-pole low-pass at 18 kHz that rounds off digital edges, emulating the analog reconstruction filter of the original hardware.

The result is a louder, warmer, denser version of the source material. Useful for drums, bass, vocals, and any channel that needs presence and weight.

ASR-10 Boost sits after Saturation and before Ducking in the signal chain:
`Gain → Sat → ASR-10 Boost → Duck → Pan → Width → Pitch → HPF → LPF → Reverb`

Cycle through levels with `Left`/`Right` in Channel Detail. Reset with `R` or `Delete`.

<Info>
  **A brief history of the Boost button.** In 1992, Ensoniq shipped the ASR-10 with a curious front-panel toggle labeled simply "Boost." Behind that one switch sat the OTTO ES5506 wavetable chip pushing samples through a fixed gain stage, an onboard ESP effects processor acting as a brick-wall limiter, and a pair of 18 kHz reconstruction filters on the analog output board. The combination was never meant to be a "sound design tool." It was a last-resort loudness hack for live performers who needed their kit to cut through a PA. But producers quickly discovered that slamming drums and bass through Boost gave them a crunchy, glued, larger-than-life character that no plugin could quite replicate. The sound became synonymous with golden-era hip-hop. DJ Premier, Pete Rock, and RZA all leaned on it heavily. To this day, people still hunt down working ASR-10 units on eBay just for that button. Now it lives in your terminal.
</Info>

## Ducking

Ducking is channel-specific and depends on selecting another channel as the source.

Persisted parameters:

* source channel
* amount
* ratio
* threshold
* attack
* release

Use ducking for:

* kick/bass separation
* rhythmic pump
* creating movement without muting

## Time-stretch

Time-stretch changes a loop's duration without shifting its pitch. It uses Signalsmith Stretch, a formant-preserving spectral stretcher suited to melodic, vocal, and mixed material.

| Parameter            | Value                    | Notes                                                                       |
| -------------------- | ------------------------ | --------------------------------------------------------------------------- |
| Ratio range          | `0.25x` – `4.0x`         | quarter speed to quadruple speed                                            |
| Near-unity threshold | `< 1%` deviation         | ratios within \~1% of `1.0` are treated as no-ops                           |
| Channels             | mono or stereo           | channels > 2 pass through unprocessed                                       |
| Processing mode      | pre-processed and cached | not real-time; result is written to disk cache                              |
| Cache schema         | `v4`                     | legacy cache entries from older versions are purged automatically on launch |

Stretch is configured per-channel by selecting a stretch source — the channel whose timing the current channel should match. The ratio is derived from the relative loop lengths. Channel Detail Sync changes wait one second before warming stretched audio; if you change Sync again during that window, only the latest value is warmed.

## Linked pad trigger

| Parameter   | Default | Range                              |
| ----------- | ------- | ---------------------------------- |
| Link target | `Off`   | Off, or any other assigned channel |

When a linked pad target is set, triggering the source channel also triggers the target channel. This applies to both live pad hits and loop-sequenced playback. Only the source channel's events are recorded into loops — the linked target fires as a side effect.

The link control is hidden when no file is assigned to the source channel. The target must also have an assigned sample for the secondary trigger to fire.

## Choke groups

| Parameter   | Default | Range                         |
| ----------- | ------- | ----------------------------- |
| Choke group | `Off`   | Off, Group 1 through Group 10 |

Channels in the same choke group are mutually exclusive — triggering one stops all others in the group that are currently playing. The choke fires before playback starts, so the newly triggered pad always sounds.

Choke groups are always visible in Channel Detail, even without a file assigned.

## SP1200 Filter

| Parameter     | Default | Range                                                                             |
| ------------- | ------- | --------------------------------------------------------------------------------- |
| SP1200 Filter | `Off`   | Off / Unfiltered (7 & 8) / Filtered (1 & 2) / Filtered (3 & 4) / Filtered (5 & 6) |

SP1200 Filter selects between four output styles from the original E-mu SP-1200 hardware. All active settings run the shared DSP chain with different output filtering: an 11.07 kHz input anti-alias filter, tape saturation, 12-bit quantization, and a 26.04 kHz multiplexed DAC and sample-and-hold stage.

### Unfiltered (7 & 8 Style)

No output filter. Raw SP-1200 character from the shared anti-alias, saturation, quantization, and DAC/S\&H stages only.

### Filtered (1 & 2 Style, SSM2044 dynamic VCF)

4-pole (24 dB/oct) ladder filter with resonant "squelchy" character and an 8-bit-stepped cutoff envelope. The envelope opens to roughly 11.5 kHz on attack (\~5 ms), then decays toward roughly 800 Hz. A small amount of upper-band bleed remains around the ladder so the imaging band does not disappear completely. Best for snares, hi-hats, and melodic content.

### Filtered (3 & 4 Style, fixed LPF)

Active multi-pole low-pass filter centered around 9 kHz with no envelope. Rolls off highs aggressively to isolate bass and kick content. Best for low-end sounds where you want a darker, thicker tone.

### Filtered (5 & 6 Style, gentle fixed LPF)

Gentle active low-pass filter in the upper treble with no envelope. It is much subtler than 3 & 4 and is meant to trim the sharpest edge while preserving presence and air. Best for pads, chords, melodic loops, and mid-range percussion where some top-end life matters.

### Common DSP stages

All active styles share these processing stages:

1. **Tape saturation**: Jiles-Atherton hysteresis model adding harmonic warmth
2. **12-bit quantization**: reduces sample depth to 4,096 discrete levels
3. **26.04 kHz DAC and sample-and-hold stage**: sample-rate reduction plus switching artifacts that create the SP-1200 imaging band

All active styles also apply an **11.07 kHz input anti-alias filter** before those three stages.

SP1200 processing sits after pitch in the per-channel effect chain, so pitch shifting changes the frequency content before the SP-1200 emulation processes it. This matches the real hardware, where the playback clock rate (pitch) fed into the DAC and output filter. Filters, reverb, and other post-pitch controls come after SP1200.

The selector is located in the **Character** section of Channel Detail.

## Sync and stem state

Additional channel state includes:

* sync multiple, shown under **Loop & Stretch** when Loop is on. With detected loop BPM and a session clock, it includes the effective post-sync BPM. Explicit filename BPM tags can supply the source tempo for trimmed loop regions when the trim length fits a normal bar count.
* selected stem
* playback offset
* mute
* solo
* linked pad target
* choke group
* optional channel name override
* optional color override

## Related pages

* [Mix](/learn/mix)
* [Mixer And Effects](/guides/mixer-and-effects)
* [Set Up Sidechain Ducking](/recipes/set-up-sidechain-ducking)
* [Time-Stretch A Loop To Tempo](/recipes/time-stretch-a-loop-to-tempo)
* [Audio And Playback Troubleshooting](/troubleshooting/audio-and-playback)