EasyEffects Presets

A collection of audio presets for EasyEffects (PipeWire).

Looking for PulseEffects presets? Switch to the pulseffects branch and install manually.

Presets

Dynamics

Vendored from droidwayin/GentleDynamics (GPLv3 upstream; carried here as data, with attribution).

• GentleDynamics — 8-band Bark-scale multiband compressor for transparent dynamic-range restoration.
• GentleDynamics Feather Loudness — same chain plus a compressor for a gentle loudness lift.
• GentleDynamics Dialogue Clarity Engine — compressors + de-esser + expander to keep film/TV dialogue intelligible.

Immersive

• Aurora Immersive — hand-tuned pure-DSP chain (EQ → bass enhancer → multiband → exciter → stereo widening → subtle reverb → crossfeed → limiter) for immersive headphone listening. It's a voiced preset, not a reference one — start ~6 dB below your usual volume.

Laptop speakers

• Cupertino Laptop Speakers — an original, from-scratch pure-DSP loudness + clarity chain (high-pass driver protection → EQ boxiness-cut & presence/air lift → psychoacoustic bass enhancer → 8-band multiband compressor → high-frequency exciter → stereo widening → brickwall limiter) hand-tuned to make built-in laptop speakers sound big, loud and refined, inspired by the tuning philosophy of MacBook speakers. Not derived from any measurement dataset — it's a heuristic engineering voicing. Aggressive and speaker-specific (not for headphones); start a few dB below your usual volume.

Utility & Effects

Original pure-DSP presets built entirely from EasyEffects' built-in plugins (no IR assets); parameters follow the EasyEffects 8.2.7 plugin schema. Each is a heuristic engineering voicing, not derived from a measurement dataset, and each deliberately exercises plugins the rest of the collection doesn't use.

• Night Listening — loudness (ISO 226-2023 equal-loudness contour) → multiband compressor → limiter, so bass and treble stay perceptually balanced at low late-night volume. Set the Loudness volume control to your actual listening level.
• Levelizer (EBU R128) — autogain normalizes any source toward a −16 LUFS integrated target (plus a safety limiter), taming wildly inconsistent track-to-track / video loudness.
• Mono Sum (Accessibility) — collapses stereo to mono via Stereo Tools (stereo-base −1), so single-sided-deaf or one-earbud listeners never lose center-panned content.
• Tiny Speaker Rescue — high-pass driver protection → psychoacoustic bass enhancer (implies a sub-bass the driver can't physically produce) → crystalizer transient clarity → limiter, for phone / laptop / small Bluetooth speakers.
• Analog Warmth — high-frequency exciter plus a short, subtle reverb for gentle tonal warmth and air.
• Concert Hall — a large-room zita reverb for spacious, ambient music listening. An effect voicing, not a reference curve.
• Movie Dialogue Boost — presence-lift EQ (~2.5 kHz) → multiband compression → de-esser to keep film/TV dialogue intelligible over music and effects.
• Reference Transparency — filter (18 Hz subsonic guard, Butterworth Q) → loudness (ISO 226-2023 equal-loudness compensation, defaulted to the −14 dBFS "K-14" home/critical-listening reference) → limiter (−1 dB ceiling, 8x oversampled for true-peak safety, 20 ms release, essentially inaudible under normal use). No EQ curve, no exciter/enhancer/compressor/reverb — the philosophy is "the best processing is the one you don't hear": only two additions to a flat pass-through, both backed by measurement standards (Butterworth alignment, ISO 226) rather than taste, plus a safety net that should never audibly engage on well-mastered source material. Note: EasyEffects' Loudness volume field directly sets the plugin's output level, not just a curve reference.

Headphone virtualization (Convolver)

• HeSuVi Virtualization Pack (41 presets) — true-stereo IRs adapted from HeSuVi's per-technology HRTF data, with +/++/+++ intensity variants. The HeSuVi DS3D + and Sound Blaster SBX 100 presets were removed as audibly identical to DS3D and SBX. See Impulse Responses and Licensing. Presets:
  • Atmos: HeSuVi Atmos
  • DTS: HeSuVi DTS Headphone X
  • GSX: HeSuVi GSX, HeSuVi GSX +, HeSuVi GSX ++
  • CMSS-3D: HeSuVi CMSS-3D Entertainment, HeSuVi CMSS-3D Game, HeSuVi CMSS-3D RX+
  • Dolby: HeSuVi Dolby Headphone, HeSuVi Dolby Headphone ++, HeSuVi Dolby Home Theater
  • DS3D: HeSuVi DS3D, HeSuVi DS3D ++, HeSuVi DS3D +++
  • DVS: HeSuVi DVS, HeSuVi DVS +
  • Nahimic / Waves / Flux: HeSuVi Nahimic, HeSuVi Waves, HeSuVi Flux HEar
  • Windows Sonic: HeSuVi Windows Sonic, HeSuVi Windows Sonic +
  • Out Of Your Head: HeSuVi Out Of Your Head, HeSuVi Out Of Your Head 2
  • Sound Blaster: HeSuVi Sound Blaster SBX, HeSuVi Sound Blaster SBX 33, HeSuVi Sound Blaster SBX 67
  • SSC: HeSuVi SSC Dublin, HeSuVi SSC New York, HeSuVi SSC New York +, HeSuVi SSC Sydney, HeSuVi SSC Sydney +, HeSuVi SSC Hù, HeSuVi SSC Hù+
  • OpenAL: HeSuVi OpenAL +, HeSuVi OpenAL ++, HeSuVi OpenAL +++, HeSuVi OpenAL CIAIR, HeSuVi OpenAL CIAIR Wide, HeSuVi OpenAL Default
  • Razer: HeSuVi Razer Surround, HeSuVi Razer Surround Bass Fix
• EFOtech MLV Pack (6 presets) — crossfeed IRs from Joe0Bloggs/EFOtech_MLV (MIT). The number is the effect depth (256–22000); start low. Presets: EFOtech MLV 00256, EFOtech MLV 00512, EFOtech MLV 01024, EFOtech MLV 02048, EFOtech MLV 04096, EFOtech MLV 22000.
• Synthetic Spherical Crossfeed — a mathematically modelled true-stereo crossfeed (rigid-sphere head, Woodworth ITD), generated by scripts/generate-synthetic-crossfeed.js.
• Synthetic Binaural Room — an "advanced" model-based true-stereo kernel generated by scripts/generate-synthetic-binaural-room.js. Fully closed-form, nothing ripped from vendor software: geometric per-ear delay/ITD from two virtual speakers at ±30°, a Brown & Duda (1998) angle-dependent structural head-shadow shelf, a pinna reflection bank (spectral notches near 8 kHz for externalization), first-order image-source early reflections from a modelled 4.2×3.6×2.6 m room, and a deterministic (seeded) exponentially-decaying diffuse tail (RT60 ≈ 260 ms, DRR ≈ 8 dB). Adds room/externalization over the drier Synthetic Spherical Crossfeed; measured ITD ≈ 250 µs, tail interaural coherence ≈ 0. Deterministic output (safe to commit/validate).
• LibreAtmos — an approximation of an object-audio mobile headphone chain (Music profile): EQ + bass enhancer + a short convolver coloration stage + limiter. The EQ/dynamics curve is hand-tuned from the chain's publicly documented behavior; provenance of the convolver kernel itself is unverified — it is currently a short 2-channel stereo IR (not the 4-channel true-stereo crossfeed format used by the other Convolver presets in this pack), so no HRTF/crossfeed spatialization is actually applied yet.
• MIT KEMAR HRTF (SOFA) — the MIT KEMAR measured HRTF (Gardner & Martin, 1995), loaded as .sofa with no conversion.
• ARI HRTF (SOFA) — a measured DTF HRTF of subject NH2 from the Acoustics Research Institute (CC BY-SA 3.0); a flatter alternative to MIT KEMAR.
• IRCAM LISTEN HRTF (Subject 1002) — the IRCAM LISTEN measured HRTF (subject IRC_1002; see THIRD-PARTY-NOTICES.md for terms), delivered as a 4-channel true-stereo .irs derived from HeSuVi's irc02 kernel (front L/R pair, head-symmetric). Unlike the SOFA presets it convolves a fixed frontal image (no live azimuth/elevation), but it loads directly without libmysofa.
• Measured HRTF/BRIR pack (3 presets) — 4-channel true-stereo .irs kernels built here from published SOFA datasets: the measurements at ±30° azimuth, 0° elevation (a standard stereo speaker pair), leading silence and the tail below −80 dB trimmed, scaled to 0 dB mono gain over 200 Hz–4 kHz (autogain off). Picked to avoid overlap with the heads already here (MIT KEMAR, ARI NH2, IRCAM 1002). Unlike the SOFA presets, these include the crossfeed from each speaker to the opposite ear.
  • TH Koeln KU100 HRTF — Neumann KU100 dummy head, anechoic, from the TH Köln HRIR_CIRC360 set (1° horizontal grid, 48 kHz, 128 taps; Bernschütz, CC BY-SA 3.0). The common reference head for binaural production.
  • FABIAN HRTF (TU Berlin) — the FABIAN head-and-torso simulator, anechoic, from FABIAN_HRIR_measured_HATO_0 (head straight ahead; Brinkmann et al., JAES 2017, CC BY 4.0). Resampled here from 44.1 kHz to 48 kHz (polyphase, 160/147). Below 200 Hz the dataset uses modelled data (per the file's metadata).
  • WDR Control Room BRIR (KU100) — a KU100 in WDR Control Room 1, Cologne (the TH Köln BRIR_CR1_KU_ROTM_L/_R sets, head at 0°; Stade, Bernschütz & Rühl 2012, CC BY-SA 3.0). A measured real room, not anechoic: about 0.27 s of reverb tail after trimming. The two speakers are within 0.6 dB of each other in level.
• Synthesized renderers — the families below are convolver → limiter chains whose kernels were synthesized here by system identification: a headphone virtualization engine was driven with test signals at 48 kHz and a linear 4-channel true-stereo (or 2-channel) kernel was fitted to its output. They are linear snapshots of adaptive or level-dependent processing, so they match the original only near the measured operating point. The names describe what each family sets out to reproduce; the measured responses and their redistribution terms are discussed in THIRD-PARTY-NOTICES.md. No head tracking.
  • FLORA (Free/Libre Object Reality Audio) — an object-based spatial audio upmix: Cinema room (spacious, reverberant) and Music room (tighter, more natural). Presets: FLORA Cinema, FLORA Music.
  • ORCHARD — a "spatialize stereo" headphone engine. The default presets include its virtual room; the Dry variants are binaural only. General is music/everyday, Game is positional, Movie has a wider stage. The Upmix variants follow the route that first upmixes stereo to surround; that upmix adapts to content, so they are a snapshot for typical music. Presets: ORCHARD General, ORCHARD General Dry, ORCHARD Game, ORCHARD Game Dry, ORCHARD Movie, ORCHARD Movie Dry, ORCHARD General Upmix, ORCHARD Movie Upmix.
  • AQUILA (Latin for eagle) — a headphone virtualizer with named modes: balanced Default, two game tunings for positional cues, a wide Movie and a subtler Music mode. Presets: AQUILA Default, AQUILA Game 1, AQUILA Game 2, AQUILA Movie, AQUILA Music.
  • DELTA (after the Dirac delta) — DELTA HD Earbud is a generic earbud correction (tonal and timing correction, not a virtualizer); DELTA Surround is a movie-mode virtualizer with bass boost off; DELTA Crossfeed is a headset crossfeed, built from extracted data and not checked against the running engine. Presets: DELTA HD Earbud, DELTA Surround, DELTA Crossfeed.
  • LibreAtmos profiles — object-audio headphone profiles, measured directly (unlike the hand-tuned LibreAtmos chain above). Each comes alone and as + Spatial, chained with a spatial-sound filter as on the original chain. Presets: LibreAtmos Dynamic, LibreAtmos Dynamic + Spatial, LibreAtmos Game, LibreAtmos Game + Spatial, LibreAtmos Default, LibreAtmos Default + Spatial, LibreAtmos On The Go, LibreAtmos On The Go + Spatial, LibreAtmos Movie, LibreAtmos Movie + Spatial, LibreAtmos Music, LibreAtmos Music + Spatial.
  • LibreSpatial — a spatial-sound headphone filter on its own: tonal shaping, no crossfeed.
  • LibreHolo — a holographic spatial-audio engine. LibreHolo Movie Quiet is Movie measured at ≤ −30 dBFS, where its level-dependent behaviour changes. Presets: LibreHolo Movie, LibreHolo Movie Quiet, LibreHolo Music.
  • LibreDecibel — a spatializer that places stereo as front speakers in a virtual room.
  • RICE Spatial — spatial audio for headphones, stereo input.

The SOFA presets convolve the whole mix as one virtual source — change azimuth/elevation live in the Convolver's SOFA controls.

Microphone (input)

Installed to <easyeffects data>/input/ (menu entry 23). Both start with an 80 Hz high-pass and RNNoise and end with a −1 dB limiter.

• Voice Noise Suppression — filter (HPF 80 Hz) → rnnoise → gate (−12 dB threshold, 250 ms release) → limiter (−1 dB).
• Voice Broadcast — filter (HPF 80 Hz) → rnnoise → de-esser (3:1) → compressor (3:1, −20 dB threshold, +3 dB makeup) → equalizer (−1.5 dB @ 250 Hz, +2 dB @ 3.5 kHz, +1 dB shelf @ 10 kHz) → limiter (−1 dB). Heuristic voicing, not derived from measurements.

Impulse Responses

irs/ holds the convolution kernels (RIFF/WAVE .irs and AES69 .sofa) used by the Convolver-based presets — every kernel here is referenced by a preset. EasyEffects' Convolver only loads mono, 2-channel, or 4-channel "true stereo" (L, LR, RL, R) kernels, so raw multi-channel HeSuVi/HRIR files have to be adapted first — that's how the HeSuVi and EFOtech kernels here were built. Inventory:

• HeSuVi (41 .irs) — HeSuVi-derived true-stereo IRs.
• EFOtech MLV (6 .irs) — EFOtech-derived true-stereo crossfeed IRs.
• Synthesized renderers (36 .irs) — FLORA (2), ORCHARD (8), AQUILA (5), DELTA (3), LibreAtmos profiles (12), LibreSpatial, LibreHolo (3), LibreDecibel, RICE Spatial; see THIRD-PARTY-NOTICES.md.
• IRCAM LISTEN (1 .irs) — Subject 1002, derived from HeSuVi's irc02.
• Measured HRTF/BRIR pack (3 .irs) — TH Koeln KU100, FABIAN (HATO 0) and WDR Control Room 1 BRIR, true-stereo kernels built from SOFA datasets (see above).
• Synthetic Spherical-Head Crossfeed and Synthetic Binaural Room (2 .irs) — generated by scripts/generate-synthetic-crossfeed.js and scripts/generate-synthetic-binaural-room.js.
• LibreAtmos (1 .irs) — short 2-channel coloration IR, unverified provenance.
• MIT KEMAR and ARI NH2 (2 .sofa) — measured HRTFs, loaded without conversion.

Kernels are trimmed of leading/trailing silence to cut Convolver CPU and latency (leading silence was trimmed from the SSC Sydney/Hù/New York, DVS and Nahimic kernels; FLORA tails were trimmed). The LibreAtmos right channel is delayed 49 samples to align with the left.

Level matching. Every Convolver preset's output gain is set so the mono-input gain averaged over 200 Hz–4 kHz is about 0 dB, accounting for EasyEffects' convolver autogain (it peak-normalises the kernel, then scales by min(1, 1/sqrt(max channel energy))). Their limiters sit at −1 dB. Switching between convolver presets should not change perceived level much.

More community IRs to experiment with: icy-comet/EasyEffects-presets-impulses.

Installation

bash -c "$(curl -fsSL https://raw.githubusercontent.com/M0Rf30/easyeffects-presets/main/install.sh)"

Press Enter for the default. The on-screen menu (install.sh --help) is authoritative if this table drifts:

#	Group	Installs
1 (default)	all	Everything (all output presets, input presets and IR/SOFA files)
2	hesuvi	HeSuVi virtualization pack (41 presets)
3	crossfeed	Synthetic Spherical Crossfeed
4	efotech	EFOtech MLV pack (6 presets)
5	kemar	MIT KEMAR HRTF (SOFA)
6	ari	ARI HRTF (SOFA)
7	gentledynamics	GentleDynamics pack (3 presets)
8	aurora	Aurora Immersive
9	cupertino	Cupertino Laptop Speakers
10	ircam	IRCAM LISTEN HRTF (Subject 1002)
11	utility	Utility & Effects pack (8 presets)
12	binaural	Synthetic Binaural Room
13	libreatmos	LibreAtmos + 12 profile presets (13 presets)
14	flora	FLORA pack (2 presets)
15	orchard	ORCHARD pack (8 presets)
16	aquila	AQUILA pack (5 presets)
17	delta	DELTA pack (3 presets)
18	libreholo	LibreHolo pack (3 presets)
19	librespatial	LibreSpatial
20	libredecibel	LibreDecibel
21	rice	RICE Spatial
22	measured	Measured HRTF/BRIR pack (3 presets: TH Koeln KU100, FABIAN, WDR Control Room)
23	input	Microphone (input) presets (2 presets)
24	uninstall	Remove everything listed in the installer manifest

curl is required (sudo apt install curl on Ubuntu).

Non-interactive use

The choice can be given as an argument (menu number, group name, all or uninstall). Without an argument and without a TTY on stdin the installer prints its usage and exits 2.

./install.sh all                  # everything
./install.sh 3                    # menu entry 3
./install.sh --ref v1.1.3 hesuvi  # pin a tag or branch
./install.sh --yes uninstall      # no confirmation prompt

Options: --ref <tag|branch> (or EASYEFFECTS_PRESETS_REF; default main), EASYEFFECTS_PRESETS_BASE_URL (overrides the whole download URL), --flatpak (use the Flatpak data directory even if a native install exists), -y/--yes (skip the uninstall confirmation), -h/--help. Files are written atomically; failed downloads are listed and make the script exit non-zero. Uninstall removes only files named in the manifest.

Manual installation

Copy the root .json files into your EasyEffects output/ directory and input/*.json into input/ — ~/.local/share/easyeffects (native) or ~/.var/app/com.github.wwmm.easyeffects/data/easyeffects (Flatpak). For Convolver presets, also copy the referenced .irs/.sofa (see each preset's kernel-name) into the sibling irs/ directory.

Licensing

The repository's own files (original presets, scripts and synthetic kernels) are MIT, see LICENSE. Bundled third-party assets are not all MIT: the GentleDynamics presets are GPLv3 upstream, the ARI SOFA file is CC BY-SA 3.0, IRCAM LISTEN and MIT KEMAR carry their own terms, and the license of the HeSuVi-derived kernels, of the LibreAtmos kernel and of the synthesized-renderer kernels could not be verified. Sources, licenses and required attribution per asset group are in THIRD-PARTY-NOTICES.md.

Contributing

Run ./scripts/validate-presets.sh (needs python3) before opening a PR. It checks canonical JSON serialization, plugin keys/values against the vendored EasyEffects schema, kernel references and headers, that every preset and kernel is listed in the install.sh manifest, and that every preset name appears in this README. CI (push to main and pull requests) also runs ShellCheck, node --check and a generator determinism check. See AGENTS.md for the full workflow.