Skip to content

AetherDSP Settings

The AetherDSP Settings dialog (Settings > AetherDSP Settings...) controls the client-side noise-reduction engines built into AetherSDR: NR2, NR4, MNR, RN2, BNR, and DFNR. Open it at any time; a radio connection is not required to change these parameters. All values are saved automatically when you close the dialog.

Before you start

  • AetherSDR must be running.
  • Decide which noise-reduction engine is active for your slice before tuning its parameters. Adjusting a disabled engine has no audible effect.

Opening the dialog

  1. Click Settings > AetherDSP Settings....
  2. The AetherDSP Settings dialog opens. Select the tab for the engine you want to adjust.

NR2 tab

NR2 is the musical-noise-reduction engine. Click the NR2 tab to access its controls.

Controls

Control Kind Default Valid range Persisted key
Gain Method Radio buttons Gamma Linear | Log | Gamma | Trained NR2GainMethod
NPE Method Radio buttons OSMS OSMS | MMSE | NSTAT NR2NpeMethod
AE Filter (artifact elimination) Checkbox Enabled β€” NR2AeFilter
Reduction Depth: Slider 1.50 0.50–2.00 NR2GainMax
Smoothing: Slider 0.85 0.50–0.98 NR2GainSmooth
Voice Threshold: Slider 0.20 0.05–0.50 NR2Qspp
Reset Defaults Button β€” β€” β€”

Control descriptions

Gain Method selects the gain-curve mapping NR2 uses. NR2GainMethod is stored as an integer: Linear = 0, Log = 1, Gamma = 2, Trained = 3. Gamma suits most voice modes.

NPE Method selects the noise power estimator. NR2NpeMethod is stored as an integer: OSMS = 0, MMSE = 1, NSTAT = 2.

AE Filter (artifact elimination) toggles the anti-artefact post-filter (NR2AeFilter). Leave this enabled unless you are specifically testing raw NR2 output.

Reduction Depth: (NR2GainMax) sets the maximum reduction depth NR2 can apply. Higher values suppress more noise but can affect speech naturalness.

Smoothing: (NR2GainSmooth) controls how quickly the noise estimate tracks signal changes. Higher values produce smoother but slower tracking.

Voice Threshold: (NR2Qspp) sets the speech-presence-probability threshold below which NR2 treats audio as noise. Raise this if speech is being suppressed; lower it if noise breaks through during pauses.

Reset Defaults restores the NR2 tab to its factory values: Gain Method = Gamma, NPE Method = OSMS, AE Filter = enabled, Reduction Depth = 1.50, Smoothing = 0.85, Voice Threshold = 0.20.

Steps β€” adjust NR2 reduction depth

  1. Open Settings > AetherDSP Settings....
  2. Click the NR2 tab.
  3. Drag the Reduction Depth: slider to the desired value (default 1.50).
  4. Close the dialog. The value is saved automatically.

NR4 tab

NR4 uses the libspecbleach library. Click the NR4 tab to access its controls.

Controls

Control Kind Default Valid range Persisted key
Noise Estimation Method Radio buttons SPP-MMSE SPP-MMSE | Brandt | Martin NR4NoiseEstimationMethod
Adaptive Noise Estimation Checkbox Enabled β€” NR4AdaptiveNoise
Reduction (dB): Slider 10.0 dB 0.0–40.0 dB NR4ReductionAmount
Smoothing (%): Slider 0 0–100 NR4SmoothingFactor
Whitening (%): Slider 0 0–100 NR4WhiteningFactor
Masking Depth: Slider 0.50 0.00–1.00 NR4MaskingDepth
Suppression: Slider 0.50 0.00–1.00 NR4SuppressionStrength
Reset Defaults Button β€” β€” β€”

Control descriptions

Noise Estimation Method selects the noise-floor estimator NR4 uses. NR4NoiseEstimationMethod is stored as an integer: SPP-MMSE = 0, Brandt = 1, Martin = 2.

Adaptive Noise Estimation (NR4AdaptiveNoise) enables continuous re-estimation of the noise floor. Enable this when the noise floor varies rapidly.

Reduction (dB): (NR4ReductionAmount) sets the maximum noise reduction NR4 applies in decibels. At 0.0 dB NR4 applies no attenuation; at 40.0 dB it applies the maximum available suppression. The default of 10.0 dB suits most SSB conditions.

Smoothing (%): (NR4SmoothingFactor) applies time-domain smoothing to the NR4 noise estimate. If raising NR4ReductionAmount causes musical noise, try increasing this first.

Whitening (%): (NR4WhiteningFactor) flattens the spectral shape of residual noise after suppression.

Masking Depth: (NR4MaskingDepth) controls the spectral-masking depth.

Suppression: (NR4SuppressionStrength) sets overall NR4 suppression strength.

Reset Defaults restores the NR4 tab to its factory values: Noise Estimation Method = SPP-MMSE, Adaptive Noise Estimation = enabled, Reduction = 10.0 dB, Smoothing = 0, Whitening = 0, Masking Depth = 0.50, Suppression = 0.50.

Steps β€” adjust NR4 reduction amount

  1. Open Settings > AetherDSP Settings....
  2. Click the NR4 tab.
  3. Drag the Reduction (dB): slider left to decrease suppression or right to increase it. The current value is shown to the right of the slider.
  4. Close the dialog. The value is saved automatically.

Tips

  • Start near the default of 10.0 dB and increase in small steps while monitoring audio quality. Values above 25–30 dB can introduce processing artifacts on weaker signals.
  • Smoothing (%): and Suppression: interact with the reduction amount. If raising NR4ReductionAmount causes musical noise, try increasing smoothing first.
  • Enable Adaptive Noise Estimation when the noise floor varies rapidly so NR4 re-estimates the floor continuously.

MNR tab

MNR is the MMSE-Wiener noise-reduction engine with asymmetric gain smoothing. It is available on macOS only. Click the MNR tab to access its controls.

Controls

Control Kind Default Valid range Persisted key
Enable MNR (macOS only) Checkbox Read from audio engine β€” MnrEnabled
Strength Slider 100 0–100 MnrStrength

Control descriptions

Enable MNR (macOS only) (MnrEnabled) enables the MMSE-Wiener noise reduction engine. The initial state of the checkbox reflects the current state of the audio engine at the time the dialog opens.

Strength (MnrStrength) adjusts MNR aggressiveness. 0 is mild suppression; 100 is maximum suppression. The value is persisted internally as a normalized 0.00–1.00 range.


RN2 tab

The RN2 tab displays information about the RNNoise engine. There are no adjustable parameters on this tab.


BNR tab

The BNR tab displays information about the NVIDIA noise-reduction engine. BNR intensity is controlled from the overlay menu, not from this dialog.


DFNR tab

DFNR uses the DeepFilterNet3 neural network for noise reduction. Click the DFNR tab to access its controls.

Controls

Control Kind Default Valid range Persisted key
Attenuation Limit Slider 100 0–100 dB DfnrAttenLimit
Post-Filter Beta Slider 0.00 0.00–0.30 DfnrPostFilterBeta

Control descriptions

Attenuation Limit (DfnrAttenLimit) sets the maximum noise attenuation DeepFilterNet3 may apply. At 0 the output is passed through unchanged; at 100 the maximum attenuation is applied.

Post-Filter Beta (DfnrPostFilterBeta) applies an additional post-filter for extra suppression beyond what DeepFilterNet3 provides on its own. At 0.00 the post-filter is inactive. The value is stored internally multiplied by 100.


Troubleshooting

  • Slider has no audible effect β€” confirm that the engine whose tab you are adjusting is the active noise-reduction engine for your slice. Enabling another engine simultaneously may mask its contribution.
  • Speech sounds hollow or distorted at high NR4 reduction values β€” lower Reduction (dB): and enable Adaptive Noise Estimation so the noise floor estimate stays accurate.
  • MNR controls are grayed out β€” MNR is available on macOS only. On Linux and Windows the MNR tab is informational.
  • Changes appear to reset β€” each tab has a Reset Defaults button (NR2 and NR4). Verify you have not accidentally clicked it.

⚑ Athena AetherSDR Assistant
Hi! I'm Athena, the AetherSDR AI assistant. Ask me anything about installation, configuration, or troubleshooting.