By now this has been posted in several locations (Reddit and CommTech Forums to name a few) however, I still feel that it is a good reference and really should be posted here as well.

1[/br]

Performed an experiment about six months ago and figured I'd finally stop sitting on the audio. This test compares the audio quality of 5% BER versus 12 dB SINAD (i.e. the points at which we spec analog and digital receiver sensitivity to) as it is reproduced by a 7/800 MHz Motorola XTS5000 with firmware release R20.50.10 on 8CALL90 (851.0125 MHz). All audio was generated by a calibrated Motorola R2670A service monitor.

The 7/800 MHz XTS5000 was chosen for several reasons:

1. It was the only easily accessible Astro 25 subscriber I had available with current firmware at the time.


2. I tend to use Motorola as the standard to which I compare everything else to.


3. I did not have access to an APX subscriber at the time.


4. The R2670A I have will support Astro (VSLEP) or P25 with it's IMBE vocoder.


5. I hadn't honestly seen my XPR6550 in almost a year at that point.


6. Unlike most APX and XPR SU's I've tested, the Astro 25 line typically hits 5% BER within 1 dB of 12 dB SINAD (not that it really matters for this experiment).


7. This one had been aligned several weeks prior.

The R2670 was chosen for several reasons:

1. The P25 encode/decode functionality of the Freedom R8000B had recently been released and my company had not purchased the upgrade yet.


2. The R8000/8100 does not have encode/decode capability for NXDN and DMR, just store and playback which made it impractical to obtain consistent base audio recordings for comparison. Maybe we'll see that in the R9000 (*cough* Doyle Wofford @ Freedom CTE *cough*).


3. I own the R2670A and my employer owns the R8000B.

P25 was chosen as the digital medium to test for several reasons.

1. See above equipment that I had access to.


2. Since FirstNet squashed the development for Phase 3 (which was focused on mobile data) and Phase 2 (TDMA) operation only applies to trunked systems, C4FM will remain the modulation method for conventional P25 operations.

A baseline recording was made using a Sure SM58 microphone, recorded at a sample rate of 192 kbps and stored as a WAV file (i.e. this is about as close as I can get to broadcast quality audio). The file was then replayed and fed into the External Modulation input of the R2670 and done so for all tests. All reproduced audio recordings were taken by directly mic-ing the speaker of the XTS5000 with the SM58. The baseline recording can be heard here.

Just as another reference point, the test was run on analog at a level that would be easily considered full quieting (~-50 dBm) at 4 kHz max deviation. Note, the frequency response has clearly changed due to the ~300Hz to 3125 Hz pass band in the radio.

Following the service manual procedure for performing 12 dB SINAD using an RLN64xx test set, 12 dB SINAD was found to occur at -120.8 dBm at 851.0125 MHz. The meter output was cabled directly to the R2670 and not an external SINAD meter such as a SAC3000. The baseline audio was then fed into the external modulation input at line levels and recorded at the 12 dB SINAD signal level (again, 4 kHz max deviation).

Finally, using Astro 25 tuner and a O.153 test pattern, 5% BER was found to occur at -121.2 dBm at 851.0125 MHz. The baseline audio was encoded via the external modulation input at 2.83 kHz max deviation (default) with C4FM modulation and the resulting recording provided here.

In conclusion, it's all personal preference. I can't stand noise and would rather listen to bit error. This is simply a comparison that can help bridge the gap between some of the common misconceptions about digital voice modes. This test should also not be used as a basis to compare the differences in receive sensitive between analog and digital as many newer radios feature receiver capable of demodulating a 5% BER level 4-6 dB below the point of 12 dB SINAD.

P25 (Digital) for the win!

When our Quantar took a crap (PA went out on a 25 watt unit) I had to run analog for a week while I waited for Sunny Comm to ship a new 110 watt Quantar, I had trouble hearing comms over the noise and loss of audio via analog in fringe areas. Guess I was spoiled on the, it works or doesn't (though with full foliage I am getting some pinched and lost bytes on fringe comms)

Using P25 is the only way to go.

I was wondering when you were going to post results. Thanks!

The one place where (high quality) digital voice really shines, in that low part of the SNR regime. But add just a few dB and IMHO analog becomes less fatiguing to listen to and more intelligible.

Of course the real motivator for digital voice is not getting a few more dB of channel performance. The real motivator is privacy and encryption.

For me the motivation is channel capacity. DMR and now p25p2 instantly doubled the capability of a system. Tetra with its four time slots could fit 32 channels (16 for full duplex) in the same bandwidth of a gsm channel that can only sustain 8.

Honestly, I'm wishing the builders of the airtime I came into control of would've just upgraded from Privacy Plus to SmartNet. Too many issues related to Capacity Plus and Connect Plus growing pains for us. (Channel capacity isn't a concern since I currently only use 15 out of the 150 channels I have slaved for my systems).