Skip to content

USB audio

SP5XTF edited this page Nov 3, 2026 · 1 revision

USB audio

Wersja polska

AUDIO tab:

  • From radio (RX) – the radio's USB codec (usually “USB Audio Device”), Speakers / headphones – where to listen.
  • Start monitoring – receiver audio on the Mac; also needed for the AUDIO SPECTRUM tab.
  • Microphone (TX) and To radio (TX) – voice transmit from the Mac microphone over USB (“Send Mac audio to the radio over USB”).

The microphone (TX) must not be the radio's input – the app refuses to transmit, because received audio would go back into the transmitter. On the radio the SSB modulation source must be set to USB.

macOS asks for microphone access on first use – permission is also needed to read audio from the radio.

Noise reduction (DSP on the Mac) – a filter between the radio's audio output and your headphones: it separates speech from noise. Levels: Light and Medium (best speech intelligibility), Strong and Max (most silence between transmissions, speech slightly trimmed). The program learns the noise spectrum by itself in about 1 s.

  • DX voice – the OM-LSA algorithm with accurate noise tracking (MCRA): for every band it estimates the probability that speech is present – speech is kept, the background is lowered gently, without “chirping”. The best choice for weak DX stations: in measurements the speech-to-background ratio was 2–3 dB better than Medium in every trial (also for stations below the noise, SNR −6 dB).
  • AI – the RNNoise neural network (Xiph): for clearly audible stations in a noisy band it quietens the background better than the other levels (works at 48 kHz). Weak DX stations – at or below the noise level – are treated as noise and removed; measured: at SNR −6 dB AI loses about 7 dB of speech, while Medium gives a better speech-to-background ratio. For weak stations choose DX voice. The AI strength slider (30–100 %, default 70 %) mixes the network output with the original – a lower strength keeps weak signals audible.
  • ANF – whistles – automatically removes whistles and carriers (e.g. a station tuning up nearby, heterodynes); it slightly alters clean speech, so turn it on when there are whistles.
  • NB – crackles – blanks crackles and impulses (atmospheric static, sparking); speech is left alone.
  • APF (CW) – a narrow peak filter in the headphones (50 / 100 / 200 Hz); its centre follows the CW decoder tone, so you hear the station the decoder is reading. The same setting is in the CW tab. Do not use ANF with CW – it removes steady tones, including the CW signal.
  • Equaliser (EQ) – 8 bands for SSB listening (200 Hz … 3.2 kHz) with a response graph. Presets: Flat, DX – voice (less bass, clearer 1.3–2.4 kHz, less hiss – a weak station in noise), Clarity (contest) (more of the same), Warm (ragchew) (gentler for long QSOs), Less hiss (cuts hiss above the speech band). Moving a slider switches to “Custom”; double-clicking a slider resets the band, the mouse wheel changes it by 0.5 dB. The overall level is lowered by the largest boost, so the equaliser does not clip.
  • Order in the chain: NB → ANF → noise reduction (AI or Light…Max) → APF → voice band → equaliser → level equalisation.
  • Voice band 250–2900 Hz – removes noise outside the speech band.
  • Level equalisation – noise reduction makes everything quieter; this evens it out.
  • Quick choice: the NR Mac panel in the left column – level (OFF / 1–4 / DX / AI) plus ANF, NB and APF (works while monitoring is on).
  • The filter only affects listening; the spectrum, the SSTV, CW and RTTY decoders and remote operation get the audio unchanged. Delay about 21 ms. Settings are stored with “Save settings”.
  • This is a Mac filter, independent of the radio's DNR – both can be used, but usually one is enough.

Save settings stores the devices (by a permanent identifier), volumes, USB transmit and the monitoring state – the program starts with them until you save again. “Unsaved changes” means the current settings differ from the saved ones.

  • Status bar (at the bottom of the window): the speaker icon mutes and unmutes listening (the level is remembered), the slider is the same volume as in the AUDIO tab; the volume is remembered between launches.
  • Transmit equalizer (TX) – 12 bands (60, 100, 160, 250, 400, 630 Hz, 1, 1.4, 1.8, 2.3, 2.8, 3.3 kHz), ±12 dB, for the Mac microphone and the remote panel microphone (audio over USB to the radio). Presets: Flat, DX – punch (lows cut, 1.4–2.8 kHz boosted – readability in noise and pile-ups), Ragchew – full voice, Contest; moving a slider – “Custom”. The equalizer changes the tone, not the loudness (the whole is lowered by the largest boost – it does not overdrive the path); set the level with MIC GAIN in the radio. The FT-710 has its own microphone equalizer – start with it flat so the two do not add up. Watch ALC and preferably listen on a second receiver or a WebSDR.
  • Automatic monitoring restart: when macOS stops audio (e.g. after a monitor with speakers goes to sleep, after the Mac wakes or an audio device change), the program restarts monitoring itself – and with it the audio spectrum and the FT8 waterfall, which are fed by the same audio. It also checks every 5 s that audio from the radio is flowing. No need to turn monitoring off and on by hand.

← CAT Share (WSJT-X, JTDX) · Home · Remote operation →

Clone this wiki locally