M-PD Voice Guard CVSD or Not?

[rmiles]

I finally received a M-PD/M-PA keyload cable for my M/A-Com IDA cryptographic keyloader. I hooked it up to my VHF M-PD Voice Guard and loaded a key. I keyed up the radio, and much to my surprise, I did not here what I expected. Over the speaker of my Thales 25, came what sounded like AEGIS. I had the same key loaded into both radios, and the same analog simplex frequency. I keyed up the Thales 25 and heard CVSD DES audio come over the M-PD speaker.

It took me a couple of seconds to put it together, and then Duh, It's called Voice Guard for a reason. So what I'm hearing from the M-PD is actually the first generation Voice Guard digital voice codec encrypted with DES. Did Ericsson/GE ever produce a CVSD DES capable radio? It's hard to believe that GE/Ericsson/Comm-Net never had a DES radio compatible with other manufacturers until they released a P25 DES/AES radio but I guess that's a fact?

 

[mitaux8030]

It might be over-simplifying things a bit, but there's three parts to a digital radio terminal:
The voice codec - the bit that turns voice into a compressed bit stream
(optional) encryption - taking the voice-approximation bit stream and jumbles it up according to the mathematics of the encryption algorithm.
and lastly the digital framing and modulation - puts the (optionally encrypted) bit stream onto an RF carrier. In doing so, it must add a layer of framing / syncronisation etc. for successful reception at the other end. Transport Layer 1 & 2 OSI stuff if you like.

Now if only just one of these 'blocks' is wrong, then no audio is recovered at the other end. Just becasue a Motorola DES radio and a GE et al DES radio both encrypt using the same algorithm, the voice codec and the modulation/framing techniques must be the same for things to all fall in line.

Having played around with Voice Guard on my MRKs, I rather suspect that it uses the very same RF modulation & framing techniques at Aegis. I also suspect that the encryption is much the same, if not identical (why re-invent the wheel?) to that used with Aegis too. The difference is in the audio codec, and this is clearly audible when comparing Voice Guard to Aegis. The tonal qualities are quite different, but I wouldn't go so far as to call Aegis superior to Voice Guard based on my quick field testing with two male voices.

I did read somewhere the voice codec used by Voice Guard, but I forget what it was...

in any case, I'm guessing that the inability to receive Voice Guard on your Thales relates back to the (proprietary) way in which the voice codec & encryption to follow is encoded onto a RF carrier, and all the framing etc etc that involves. Back in the 80s when Voice Guard was born, the mindset of companies was to market their unique product and hope to corner a profitable slice of the market. Making things 'compatible' was something that received more attention in the late 80s / early 90s.

 

[rmiles]

After using Motorola, E.F. Johnson, Thales for many years, I momentarily forgot about VG. The other manufactures use, or have used CVSD in the past. I don't believe that Ericsson/GE ever adopted that "standard" as a basis for DES. What I was really looking for was the answer you gave on tonal quality. How does VG compare to AEGIS?

I understand you can't simply encrypt an analog voice signal, it must be digitized beforehand. Many manufactures used CVSD to do this. Apparently, Ericsson/GE never went with CVSD, instead developing VG, and VGE.

I was wondering how different VG sounded when compared to AEGIS, and both compared to CVSD. I'll have to set up a little experiment.

 

[mitaux8030]

CVSD is the voice coder technology (analogue to digital), quite seperate from the encryption of course. I found the information I was initially looking for, and it didn't list the voice coder technology for either Voice Guard or Aegis, though one document I have metions SBC voice data for Aegis, and another reference here says that Aegis uses MBE vocoding, possibly an early DVSI vocoder. No idea what the Voice Guard vocoder is - being early 80s, it was probably an in-house GE creation.
I'll have to make some Voice Guard (encrypted) vs Aegis vs Aegis encrypted samples of recovered audio for you.