Skip to content

AetherDSP Settings

The AetherDSP Settings dialog provides advanced configuration for all client-side noise reduction engines in AetherSDR. It includes six DSP modules: NR2, NR4, MNR, DFNR, RN2, and BNR.

Opening the dialog

  • From the RX Chain strip, double-click the ADSP tile.
  • From the VFO DSP grid, click the ADSP button.

Dialog window

The dialog has a frameless title bar with a grip glyph (โ‹ฎโ‹ฎ) on the left. The title bar includes three buttons:

  • โ€” (Minimize) โ€” Minimizes the dialog.
  • โ–ก (Maximize) โ€” Maximizes or restores the dialog.
  • ร— (Close) โ€” Closes the dialog.

Drag the title bar to move the dialog. Double-click the title bar to toggle maximize/restore. Drag any edge or corner to resize (6 px resize hit zone).

The dialog position and size are automatically persisted across sessions using the setting key AetherDspDialogGeometry.

The dialog uses theme-aware styling applied through the ThemeManager. Colors are derived from the active color theme rather than hardcoded values.

Toggle row

The top of the dialog contains six toggle buttons that serve as both tab selectors and engine enable/disable controls:

  • NR2 โ€” Musical-noise-reduction engine
  • NR4 โ€” libspecbleach spectral noise reduction
  • MNR โ€” macOS MMSE-Wiener noise reduction (dimmed on Windows/Linux)
  • DFNR โ€” DeepFilterNet3 neural noise reduction
  • RN2 โ€” RNNoise neural noise reduction (informational only, no adjustable parameters)
  • BNR โ€” NVIDIA Broadcast neural denoising (dimmed without NVIDIA Broadcast SDK)

Clicking a toggle activates that engine and selects its tab. Clicking it again deactivates the engine. Only one engine can be active at a time โ€” NR2, NR4, and DFNR are mutually exclusive. MNR and BNR may stack in some builds.

NR2 tab

Controls for the musical-noise-reduction engine.

Control Type Default Range Setting Key Behavior
Gain Method Radio button Gamma Linear, Log, Gamma, Trained NR2GainMethod Selects gain-curve mapping (stored as integer 0-3).
NPE Method Radio button OSMS OSMS, MMSE, NSTAT NR2NpeMethod Selects noise power estimator (stored as integer 0-2).
AE Filter (artifact elimination) Checkbox True โ€” NR2AeFilter Toggles the anti-artefact post-filter.
Reduction: Slider 1.50 0.50-2.00 NR2GainMax Sets maximum NR2 reduction depth.
Smoothing: Slider 0.85 0.50-0.98 NR2GainSmooth Controls how smoothly the noise estimate tracks changes.
Threshold: Slider 0.20 0.05-0.50 NR2Qspp Sets speech-presence-probability threshold.
Reset Defaults (โ†บ icon) Push button โ€” โ€” โ€” Restores NR2 defaults (Gamma/OSMS/AE on, 1.50/0.85/0.20).

Sliders on this tab use theme-aware styling via applyPrimarySliderStyle().

NR4 tab

Controls for the libspecbleach spectral noise reduction engine.

Control Type Default Range Setting Key Behavior
Noise Estimation: Radio button MMSE MMSE, Brandt, Martin NR4NoiseEstimationMethod Selects noise-floor estimator (stored as integer 0-2).
Adaptive Noise Estimation Checkbox True โ€” NR4AdaptiveNoise Enables continuous re-estimation of the noise floor.
Reduction (dB): Slider 10.0 0.0-40.0 NR4ReductionAmount Sets maximum NR4 noise reduction in dB.
Smoothing (%): Slider 0 0-100 NR4SmoothingFactor Time-domain smoothing of NR4 noise estimate.
Whitening (%): Slider 0 0-100 NR4WhiteningFactor Flattens residual noise spectral shape.
Masking Depth: Slider 0.50 0.00-1.00 NR4MaskingDepth Controls spectral-masking depth.
Suppression: Slider 0.50 0.00-1.00 NR4SuppressionStrength Overall NR4 suppression strength.
Reset Defaults (โ†บ icon) Push button โ€” โ€” โ€” Restores NR4 defaults (MMSE/adaptive on, 10 dB, 0, 0, 0.50, 0.50).

Sliders on this tab use theme-aware styling via applyPrimarySliderStyle().

Note: NR4 requires LLVM (clang-cl) on Windows. The toggle is dimmed if LLVM is not installed. Install LLVM from llvm.org and rebuild AetherSDR to enable NR4.

MNR tab

Controls for the macOS MMSE-Wiener noise reduction engine. This tab and its controls are available only on macOS builds.

Control Type Default Range Setting Key Behavior
Enable MNR (macOS only) Checkbox โ€” โ€” MnrEnabled Enables MMSE-Wiener noise reduction with asymmetric gain smoothing.
Strength Slider 100 0-100 MnrStrength Adjusts MNR aggressiveness (0 mild, 100 max).

DFNR tab

Controls for the DeepFilterNet3 neural noise reduction engine.

Note: DeepFilterNet3 is an advanced neural-network-based noise reduction system that uses deep learning to suppress noise while preserving speech quality. It may require significant CPU resources.

Control Type Default Range Setting Key Behavior
Attenuation Limit Slider 100 0-100 dB DfnrAttenLimit Sets maximum noise attenuation applied by DeepFilterNet3. 0 = passthrough; 100 = maximum suppression.
Post-Filter Beta Slider 0.00 0.00-0.30 DfnrPostFilterBeta Applies an additional post-filter for extra suppression.

RN2 tab

Selects the RNNoise neural noise reduction page. This tab is purely informational and has no adjustable parameters.

BNR tab

Selects the NVIDIA Broadcast neural denoising page. Intensity is controlled from the overlay menu. The toggle is dimmed on builds without the NVIDIA Broadcast SDK.

Bypassing all client NR engines

To quickly disable all client-side noise reduction without opening the AetherDSP Settings dialog:

  1. Locate the ADSP tile in the RX Chain strip.
  2. Double-click the ADSP tile to open AetherDSP Settings.
  3. In the toggle row at the top, click each active (lit) noise reduction toggle to deactivate it.
  4. Continue until all toggles are dimmed.

The ADSP tile updates to reflect the bypassed state. No client NR engines are now active, returning the audio to the raw sliced audio stream from the radio.

Tips

  • The six DSP toggles (NR2, NR4, MNR, DFNR, RN2, BNR) serve as both tab selectors and engine enable/disable controls.
  • NR2, NR4, and DFNR are mutually exclusive โ€” only one can be active at a time.
  • MNR and BNR may stack with other engines in some builds.
  • The Reset Defaults button (โ†บ icon) on each tab restores that engine's parameters to their default values.
  • Settings are persisted across sessions.
  • The dialog uses theme-aware styling. Colors are drawn from the active color theme rather than fixed values.
โšก Athena AetherSDR Assistant
Hi! I'm Athena, the AetherSDR AI assistant. Ask me anything about installation, configuration, or troubleshooting.