[Top] [All Lists]

[TenTec] Digital Speech Processing & Pegasus

To: <tentec@contesting.com>
Subject: [TenTec] Digital Speech Processing & Pegasus
From: mark@microenh.com (Mark Erbaugh)
Date: Tue, 19 Feb 2002 11:02:01 -0500

Thanks to your post, I'm now beginning to understand speech processing. If
nothing else, this is a great learing experience for me.

> Clipped audio converted to SSB RF has enormous amplitude spikes. That
> makes audio clipping and audio compression work poorly for enhancing SSB
> transmitter output. The whole reason for any compression is to allow
> driving the PA to a higher average power without driving it into an
> excessively nonlinear region (which produces excess intermodulation
> products known as splatter). Since audio processing tends to introduce
> the peaks the PA needs to not see, RF processing is necessary. And then
> it has to be careful (such as by post envelope shaping bandpass
> filtering) to not let the IMD products from the envelope shaping be
> heard as splatter.

Am I correct in assuming that the distortion is caused by the relatively
sharp corner where the speech wave is clipped? As I recall, a step change
consists of an infinite number of harmonics. I suspect that a sharp corner
caused by clipping would be seen as a step change and thus produce the

If so, would it be possible to reduce the distortion by using DSP to 'round
over' those corners?

> I've found that while the ear is insensitive to the relative phase of
> harmonics, the SSB transmitter is very much so, complicated by the
> relative phase response to harmonics of the microphone, the audio
> circuits and especially the SSB transmit filter. I have found that
> selecting the microphone can enhance the average power from the
> transmitter (as limited by the peak detecting ALC). I suspect that the
> optimum microphone will vary with the voice and especially the relative
> phase response of the transmit SSB filter. I believe it is important to
> use a microphone with limited response, similar to that of the SSB
> transmit filter so that audio stages and balanced modulator are never
> overdriven by audio components that can't pass through the SSB transmit
> filter. Hence I dislike the D-104 for its peak response is outside the
> SSB transmit filter. I get along well with the EV664 (its wideband
> response is decently flat) and for HF SSB I use a Shure R-5
> communications microphone element mounted in a D-104 case. I have a
> battle ship Turner hand microphone that also has proven to work well
> with my VHF Yaesu FT-726.

I'm currently using TT 705's with my TenTec gear. How do they stack up? Have
you evaluated the various Heil cartridges? They seem to be very popular?

Could DSP be used to tailor the input from a wide range mic to compensate
for the speakers voice characteristics. In this case, I am envisioning
recording the speaker.s callsign or other information to a .WAV file, then
using DSP in the PC to process that .WAV file into one that has just the
information needed for optimum results at the receiver. I suspect that the
needed processing would vary depending on band conditions, but with a PC you
could have serveral different processed .WAV files to use under different


<Prev in Thread] Current Thread [Next in Thread>