Performance expectations when transitioning from analog to digital systems

[xmo]

Earlier in this discussion, I posted a drawing which could be named exactly the topic of this discussion: "Performance expectations when transitioning from analog to digital systems"

Virtually all the radio companies have such a drawing. I chose this one (originally from MaCom) because I think it is nicer looking than others I have seen.

I also asked the question: "If you you were buying a system and the vendor's sales team showed this to you, what would you ask them to provide in the way of proof?"

I posted the same drawing on repeater-builder and asked essentially the same question. One poster observed: "there are no scales"

Bingo. As a communications professional, wouldn't you want to see numbers associated with each axis? Numbers like delivered audio quality vs signal strength. Then you would want to know where the numbers came from. The vendor's engineers? I would rather see something from an independent third party like a university thesis or a qualified testing lab.

I have seen various versions of this graph but I have never seen one with numbers from either a manufacturer or within a technical paper or research report.

 

[Ubbe]

But those kind of numbers would come from a lab test bench, having a steady and fine signal to measure. That's not a real life situation where digital recover time and other digital issues come into place when receiving fading and rapid signal loss "picket fencing", that needs the use of tripple diversity digital receivers to overcome. I have a Icom R2500 receiver with dual diversity and picket fencing are totally gone when using two antennas on my car roof.

What radio sellers regard as bad analog audio quality are when there are some background white noise but you still hear clearly what are said and the digital voice distortion sound usually sound worse than that. What drastically changes are at the lowest signal level when you can hear an analog yes or no as a respond but digital are just jumbled data bits or silence.

If you google you will probably find tests done by students from technical universities and the like that has done some more or less life like simulated tests.

/Ubbe

 

[MTS2000des]

But those kind of numbers would come from a lab test bench, having a steady and fine signal to measure. That's not a real life situation where digital recover time and other digital issues come into place when receiving fading and rapid signal loss "picket fencing", that needs the use of tripple diversity digital receivers to overcome. I have a Icom R2500 receiver with dual diversity and picket fencing are totally gone when using two antennas on my car roof.

Real world? I have 15 RF subsites, 70 percent of the subscribers are portables with low rent 1/4 wave stubby antennas. Each RFSS uses a TTA and combine/multicoupler (separated TX and RX) which is custom in P25 and other trunking systems.

The GCM8000 is not like some ancient comparator. It assembles a solid error free stream from all 15 RF subsites in real time. All of these "digital issues" hobbyists complain about on inferior hobby grade equipment like the Bowleturd radios in the video or cobbled together ham trash doesn't happen. If it did, I'd be out of a job and the vendor would not be making bank selling tons of these systems worldwide.

Hobbyist radio and pro radio are different worlds, see my analogy re the Ducati.

 

[KevinC]

Hmm…which one to believe. A YouTube video from who knows who or professionals with years of experience. That’s a tough one.

I bet I can find YT videos that prove the earth is flat also.

 

[xmo]

But those kind of numbers would come from a lab test bench, having a steady and fine signal to measure. That's not a real life situation where digital recover time and other digital issues come into place when receiving fading and rapid signal loss "picket fencing"...

OMG, the communications professionals have all been sitting in the lab and never thought about the real world!

Oh, wait. It turns out that we have thought about it:

"Wireless communication is one of the most demanding applications for the telecommunications equipment designer. Typical signals at
the receiver experience multipath fading resulting in large signal power fluctuations.

The only way to ensure that a system is able to function properly in an environment with these characteristics is through extensive
testing. . To achieve accurate and repeatable results, the RF channel must be brought into the lab through the use of channel emulation techniques."

Maybe those guys in the lab can be trusted afterall.

"fading and rapid signal loss "picket fencing" that needs the use of tripple diversity digital receivers to overcome"

Ok, let's address diversity receive in the context of land mobile two-way radio in North America and particularly the USA.

A great many systems employ receiver voting which is macro diversity. However, Ubbe is referring to micro diversity, a technique where two or more separate antennas are employed at a single receiver location. If two antennas are directly adjacent to each other, the signals received by each will be essentially identical.

By separating the antennas by a suitable spacing, the signals received by each will arrive by slightly different paths with the result that their fading profiles become de-correlated which permits a process within the receiver to recover a better demodulated composite result than would be possible from either individual signal.

There are two directions to consider: uplink from and downlink to the user. It's all well and good for hobbyists to run around with two receiving antennas on their vehicle and in the early days of AMPS mobiles there were some products that supported receive diversity.

However, cellphones quickly transitioned to portable and, correspondingly, commercial and public safety systems today must support a preponderance of portable users. Obviously, the antenna separation necessary for this type of diversity receive is impossible for a portable.

To provide the desired grade of service, the industry understands exactly how much signal is required to meet a particular DAQ requirement in the faded real-world environment and systems are designed accordingly.

The uplink is certainly a candidate for diversity receive since the lower power of portables means that system coverage is typically talk-in limited. With respect to phase one P25 systems, the industry has not embraced diversity base receive, however, the phase two uplink modulation format is slightly disadvantage compared to C4FM used for phase one uplink.

In order to support the goal of identical system coverage when transitioning to phase two, Motorola does offer a diversity receive option.

Maybe someone can comment as to other vendors such as Harris.

Finally, I am not aware of any NA/USA P25 system utilizing triple diversity.

 

[NVAGVUP]

It is likely an effort in futility, but here it goes.

P25 stream is 9600 baud. ~ 25% of the data is dedicated to error correction. (Per 1990's workshops I attended) I am not going to fact check. Knock yourself out. How much error correction does analog have?

Similar to experiences to others. This is the best example.

Oversaw a 3 site VHF simulcast (Conventional) deployment (With 5 additional RX only voted sites) Each site was built identical. Shared the same antenna system, same power levels/ERP, rx gain etc. 5 RF Channels. Qty 4 P25. Qty 1 analog (Narrowband)

Vendor Coverage testing was NOT part of vendor buildout. So a field test plan was developed internally (Using DAQ). Evaluating system talk in and system talkout (Portable) performance. After the first day of testing (3 days total), the assistant fire chief wanted to convert the only analog channel to P25.

Per xmo chart, P25 beats analog (NB) for "USABLE" recovered audio. (And by a dramatic margin). Experience above was not a lab, but real world experience. By a public safety professional who's job it is to "CLEARLY" understand what is received. For a hobbyist, it can be fun and challenging to "pick fly cr@p out of chili". For public safety responder who's life may depend upon understanding what is received, not so much fun.

Ubbe- I am more than happy to share additional real world experiences. Feel free to give me a call on your good old bagphone. (Since analog is the "best")

 

[xmo]

Ubbe wrote: "If you google you will probably find tests ...."

There is a lot of quality information published and accessible on the internet but trying a sensible sort of search like:
'two-way radio analog vs digital audio quality' will produce a lot of results from sites offering to sell you their products.

One respected entity that has done extensive audio quality research is the "The Nation’s Spectrum and Communications Lab", formally:
U.S. Department of Commerce, National Telecommunications and Information Administration, Institute for Telecommunication Sciences (ITS)

A number of years ago, ITS conducted tests of P25 radios and published the results in a report: 'Delivered Audio Quality Measurements on Project 25 Land Mobile Radios' This report says:

"The Institute for Telecommunication Sciences has compared the speech quality performance of the new generation of Project 25 land mobile radios to analog FM systems."

That comparison is exactly the topic of this discussion.

The same exact radios were tested for delivered audio quality vs signal strength in both P25 and analog FM modes. In their report, the Institute did not provide a comparison graph of the results in the format we have discussed here, however, they provided a table of DAQ vs signal results data for both formats.

It recently occurred to me to copy the Institute's data into excel and use it to produce a graph. The result is attached. It looks remarkably like the graphs used by the radio companies except with real numbers on the scales.

A couple thoughts:

DVSI vocoder performance quickly improved from the first generation that ITS tested so - if repeated today, the highest digital DAQ would be in the neighborhood of 4.2.

The irregularities in the plotted curves are probably well within the accuracy limits of their test algorythm so, if they had chosen to create such a graph they could have employed a 'smoothing' function, however, I chose not to fudge their data for the sake of appearance.

 

[kayn1n32008]

It is the complaining about audio quality vs analog that is the biggest pain in my back.

A friend of mine just returned from doing a 2 week stint doing wildland interface structure protection in Northern Canada. 4 of his fellow fire fighters were with him along with 2 apparatus that they took.

They have been using a 2 site IPSC DMR system for many years in their home jurisdiction. Dispatch and fire ground both being DMR at his department.

On this deployment all comms were analogue. Fire Line, Air to Ground and Command. His brother and sister fire fighters all complained about having to listen to analogue comms. Static, squelch crashes and widely varying audio levels. They all much prefer digital voice.

 

[TampaTyron]

Damn..... we have come full circle. Only took 16 years..... 2007-2023- RIP "analog is superior" whiners. TT

 

[kayn1n32008]

Damn..... we have come full circle. Only took 16 years..... 2007-2023- RIP "analog is superior" whiners. TT

I'm not surprised. Clear, static free audio recoverable down to -115dBm or weaker vs. static/picket fencing/squelch crashes? Analogue get tiring to listen to. I'll take digital audio all day long.

 

[dispatchgeek]

Damn..... we have come full circle. Only took 16 years..... 2007-2023- RIP "analog is superior" whiners. TT

I believe modern radios have more audio options and seem to do a better job of utilizing the vocoder to make more natural sounding audio. The voice quality certainly seemed more "robotic" on the 1st generation P25 radios we used compared with our modern equipment.

 

[Echo4Thirty]

the DSP in the radios is remarkably better than it used to be. Not to mention going from the IMBE vocoder to AMBE.

 

[ElroyJetson]

My only complaint about modern day p25 digital audio is that it's a little bit muffled. Like it needs to be run through a "de-esser" only in reverse. Add a bit of sibilance to it where needed. But generally I like the way that the audio comes through clearly with, notably, no background noise.

But it's still not on the level of the call audio out of a modern smartphone.

 

[MTS2000des]

You should try a Kenwood Viking. Modern P25 radios like that have superior audio and customizable transmit and receive response curves/profiles in software. Many folks roll with "default" settings. This is all easily adjusted by a competent programmer/system specialist.

 

[WB5UOM]

KENWOOD does have excellent audio.
That is usually the first comment I get when a new Customer hears it.

Cant say about the 800 side, but on VHF, the Atlas 1200 (straight 1200 not the crap cousins 4xxx) P25 conv excellent decoded audio down to -120dbm. (at rx input)
Couple that with a KENWOOD NX5000, hard to beat. Customers love it.
And yes the straight 1200 is a Spectra machine MX800

 

[MTS2000des]

My VP8000 can produce a DAQ of 4.0 from a -115dbm UHF repeater 20 miles away inside a building. No BER. Analog at that signal level is just white noise hash and trash. Analog FM can sound good IF the conditions are perfect and signal is strong enough, good receiver (REAL radio not a Baoturd) with quieting, and transmitter is quality (again, REAL radio not a muffled Bowelturd)- but this is rarely the case where digital voice overcomes obstacles and produces consistent, quality AF under less than perfect and adverse RF conditions. Quality subscriber radios and FNE implemented, optimized and maintained by COMPETENT technical persons make it shine.

 

[kayn1n32008]

I've had the pleasure of using all sorts of different types of radio systems both while working and Ham radio. I've used both analogue and digital voice in both services. I've used analogue and digital voice on VHF, UHF and 800MHz bands.

Analogue simplex(both Ham and LMR)
Analogue repeated(both Ham and LMR)
Analogue Type 2(LMR)
NXDN simplex clear V-Narrow and Narrow(ham)
DMR simplex AES/RC4/clear(Ham&LMR)
DMR conventional repeated AES/clear(Ham&LMR)
DMR Cap+ AES/RC4/clear(Ham&LMR)
DMR Con+ clear(LMR)
P25 simplex AES/clear(Ham&LMR)
P25 phase 1 RC4(LMR)

By far, P25 phase 1 was the nicest to use on the ears. No static, no picket fencing, just consistent, clear understandable audio. Easy to listen to for a 12 hour long shift. Consistent system access time.

Second is unencrypted DMR. Same as P25, but slightly lower audio quality. With Motorola Gen 2 and RX Audio Leveling, it is at least consistent audio.

Least favorite is analogue repeated/simplex, so many(too many?!?!) years running on radio controlled roads it's lots of static, picket fencing and VERY inconsistent radio, all dependant in the quality of the transmitting radio and the 'technique' of the user pushing the PTT. Constantly having to adjust the volume sucks ass.

I recently have been using a Motorola Type 2 system, it's a 2 site system, and has excellent coverage. Experience has been okay, but I have been using XPR6580IS/4580 subscribers with R01.10.24 Type 2 firmware. They leave ALOT to be desired. They work, access times are good BUT they are just not great radios, especially in a loud environment. Business critical vs mission critical. There is a BIG difference.

All in all, P25 has been the most enjoyable for me. Consistent audio, no static/picket fencing/beeps/boops/squelch crashes and users not doubling each other is really nice.

Listening to static/picket fencing/MDC1200/users doubling each other/squelch crashes and wildy varying audio levels inherent in analogue systems is incredibly tiring over the course of a long day.

Ultimately, a well designed P25 trunk system, with properly engineered resources and coverage with properly programmed subscriber radios is the least tiring radio system to use.

I'd never go back to analogue if I could help it.

 

[kayn1n32008]

My VP8000 can produce a DAQ of 4.0 from a -115dbm UHF repeater 20 miles away inside a building. No BER. Analog at that signal level is just white noise hash and trash.

I think the majority of radio users don't understand just how well digital formats work at these signal levels, because of FEC and just how poor an analogue signal sounds at the same signal level.

At -115dBm, Your VP8000 hasn't even reached the digital 'ledge', it's still recovering ALL the voice data, and can still receive even weaker signals, with better audio quality than am FM radio. The ledge is at a signal strength that the FM radio will be 100% unusable.

 

[kayn1n32008]

My only complaint about modern day p25 digital audio is that it's a little bit muffled... But it's still not on the level of the call audio out of a modern smsmartphone.

You realize there is a real significant difference in the size of the voice data payload between a smart phone and a narrow band digital radio right? That's an apples to oranges comparison.