Insight - Why audio goes twangy, ringing an metallic in DV

Fundamentally, in simplex and repeater usage, it’s simply your SQ as received is poor and heading to what you’d call the noise floor in FM.

Where you are below threshold in FM, the received audio goes noisy, I.e. the natively heard white noise of an frequency when perceptively empty (as you’d hear with you opened the squelch on an empty frequency) starts to intrude on your weak signal, and since FM receivers have a capture effect, what’s actually heard is both empty white noise and your weak signal being swapped back and forth. So the lower your received SIG level is, the worse it becomes until you just become modulated analogue artifacts momentarily creeping through the noise.

So, how does this explain the effects in DV?

Well, if you understand DV well enough to realise that your voice is chopped and p-a digitally compressed using a very aggressive, but admittedly a clean aggressive, encoding and data reduction method to get a really small encased facsimile of your voice sent, consequentially an error impacts the quality of what’s decoded.

So given digital voice transmission, with the exception of SSB based modes, uses FSK directly or AFSK within FM (depends on the system, but mostly pure FSK), the same basic rules apply as with FM to how an FSK modulated SIG gets demuxed, and if the SNR is poor, instead of hearing noisy audio down to whispered artifacts at near the noise floor, you get various degrees of misdecoded digital audio.

At moderately poor SNR, you start to get metallic Dalek type reproduction occur, where it’s still audible and legible, but lots of the tonality of the voice (the duff bits of data) means only the main frames get demuxed properly, the intermediate differences ones don’t and the worse the condition gets, the less you actually hear clearly sounding natural.

Where you get that and dropouts accurring - that’s a sure sign that the FEC is literally discarding the missing frames entirely.

So the worse the SNR gets, the more aggressive the metallic choppy audio gets until you reach a point where there’s little or nothing. But, in fairness, the SNR level that happens at is a level where FM analogue audio modulation is long past usable.

And that’s both the essence of the tonal issues you hear, and also the essence of why DV can often make for a better mode under extremely poor SNR conditions.

But also note, there is an opposite combined choppiness and worsening readability that can occur when the AF level into the mic chain gets too high an SPL and audio dynamically compresses. Best described, I find, where the audio goes flat tonally and heads towards the audio equivalent of two-tone printing (where the images get reduced to a few times and have a low dot pitch). In this case, the ringing is the over level clipping mirror equivalent of poor SNR metallic sound.

On Sat receivers, you could (well on analogue you could) see what where known as ‘sparklies’, these indicated overly high SNR which was overloading the receiverZ’s front end and at the other extreme under poor weather conditions, literally weak reception due to a crappy SNR. So during torrential rain, you could see it go through the whole extremes at times. Why the SNR suffers so much at SHF is another topic entirely.

Note, the above is a simplified essence of what happens, the full explanation of how SNR matters in how something is resolved or not, that’s heavy reading (or in my '72 RadCom handbook, where FM was very touched on, SNR in receivers was a good few nights of refresh reads to get to grips with, as I remember as a kid). Definitely not something to reproduce here, but I’m sure if you’re motivated, you’ll find the FSK (telegraphy) and FM audio transmission aspects well documented in later RadCom and ARRL handbook editions. There are lots of even more in-depth volumes in the tech references libraries out there, some written exclusively around digital voice and images transmission.

But in essence, it’s much like what happens in analog FM but with a different frequency shifting type of modulation (SSB and AM digital voices being different again and very alien in the how’s and whys).