Afreet Software, Inc.
     Software created by Alex Shovkoplyas, VE3NEA
Products





Morse Expert
for Android







Ham Cockpit





CW Skimmer





DX Atlas





RTTY Skimmer Server





GRITTY





HamVNA





CW Skimmer Server





Pileup Runner





DX Bulletin Reader





Voice Shaper





Band Master





Faros





Morse Runner





Ham CAP





Rocky





Morse Midi





Master.dta
Editor






Tech stuff



Programmer's
Corner





OmniRig





QRQ Profiler





RttyCompare





SSTV Tools





Toy NTP







Voice Shaper 1.0

FREEWARE




Features

Voice Shaper is a speech pre-processor for SSB transceivers. It receives the voice signal from the microphone connected to the soundcard, and sends the processed audio back to the soundcard in real time. The following processing functions are implemented in the program:
  • bandpass filter;
  • 7-band equalizer;
  • noise gate;
  • compressor;
  • RF envelope clipper.

The configuration procedure of the DSP functions is interactive: the program records your voice, then it plays the recording back through the processor and allows you to adjust the settings and immediately hear the effect of your changes. Real-time spectrum and histogram displays provide the insight into the processor operation and make the configuration process even eaier.








Main window







System Requirements


Operating system: Windows ME, Windows 2000, Windows XP or Windows 7.

CPU: Pentium 4, 1GHz or higher.

RAM: 256 Mb.

Sound Card: Any sound card with a valid WDM driver.






Quick Start

  • Download the Voice Shaper from the Downloads page and install the program.

  • Start Voice Shaper and follow the step by step instructions shown on the left panel in the program.


    1. Click on button 1 and select the input and output audio devices. These can be either two different sound cards or the same card, your choice.

    2. Click on button 2 to enable audio input, and adjust the Mic Input gain in the Windows audio control panel so that the green sector on the signal strength meter occupies about 50% of the scale.




      Input peak meter


      Note that the meter is two-sided. The positive and negative peaks are measured separately so that the DC offset in the sound card can be detected, as on the screenshot above. Try to avoid the cards with a significant DC offset.
      Use the Software Gain option only if your sound card does not have a Mic pre-amplifier.

    3. Click on button 3, click on OK to start the recording session, and spell your callsign in the microphone. The recording will stop automatically in 5 seconds.

    4. Click on button 4. The playback will start. Listen to the output audio and configure the processor as described in the Processing Functions section below. Adjust the playback volume slider for comfortable listening. Once the processor is configured, test your signal in the noise and QRM using the controls on the Playback panel.

    5. Save the audio processing profile by clicking on the Save Profile toolbutton. Multiple profiles can be created for different operating conditions.

    6. To start real-time audio processing, click on the Start Processing toolbutton, or click on the Voice Shaper icon in the system tray. The systray icon has a right-click menu with most common commands.

    7. The program behavior on startup can be changed via the Startup Options toolbutton.





Processing Functions




Linear processing


The combined frequency response of the bandpass filter and equalizer is shown on the Equalizer page, on a logarithmic frequency scale. The real-time spectrum (dark green line) and the total integrated spectrum (light green shaded area) are overlaid on the chart.

  • The bandpass filter cutoff frequencies can be set between 50 and 500 Hz (low) and 1500 and 5700 Hz (high) by dragging the yellow control points on the chart with a mouse.

  • The equalizer can be configured by dragging the blue control points. 7 octave filters, from 75 Hz to 4800 hz, can be adjusted within +/- 20 dB.





Filter and equalizer configured for the standard 200-2700 Hz SSB signal






Filter and equalizer configured for the 50-5700 Hz E-SSB signal






Non-linear processing


The amplitude response of the non-linear part of the processor is shown on the Clipper page, with the histogram of the input envelope magnitudes overlaid on the chart. The red spot indicates the average magnitude. The magnitudes of the output envelope are plotted on the right panel of the chart, on a vertical scale.

  • RF envelope clipping is the most efficient way of increasing the average power of the SSB signal without increasing the peak envelope power. Voice Shaper computes the RF envelope analythically, without actually generating RF. The clipping threshold can be adjusted by dragging the green control point. To disable clipping, drag the point all the way to the right. The clipping level, in dB, can be found as the distance between the rightmost bar of the histogram (about -11 dBFS on the chart below) and the clipping threshod (-34.3 dBFS).





    RF envelope clipper configured for 23 dB clipping






  • Compression is required to ensure that the signal strength does not change significantly when the operator turns away from the microphone. The compressor in Voice Shaper is controlled by the average RF envelope magnitude and can be configured for the compression ratios between 1:1 and 15:1. To adjust the level of compression, drag the yellow control point vertically. To disable compression, drag the yellow point all the way down until the displayed compression ratio is 1:1.





    Compressor configured for the 3.8:1 compression ratio






  • Noise Gating prevents transmission of the background noise in the pauses between the words. The noise gate in Voice Shaper works in the downward expansion mode: instead of just muting the channel, it gradually decreases the gain as the input signal strength goes down. The expansion ratio is fixed at 1:3, but the gating threshold can be changed by dragging the yellow control point horizontally. For best results, do not set the threshold higher than necessary to suppress the noise.






    Noise gate threshold at -52 dbFS






Voice Shaper on other platforms

  • Linux: Mark, N1JV uses Voice Shaper running in a virtual machine with Vbox and a Linux host computer. Here's his full configuration:

    • Intel i5-760, 8GB RAM
    • Xubuntu 16.04.1 LTS host OS
    • Virtual Box 5.1.22 using AC-97 virtual audio card, Pulse Audio host
    • Windows XP SP3 Guest OS
    • VShaper 1.0








Copyright © 1998-2020 Afreet Software, Inc.


I/O


Downloads


Registration





Author's e-mail