I agree entirely on the RF part of your article.
But. Digital modes somehow eliminate feedback? Echo cancelation exists but a great deal of the time it fails miserably. You end up with voice-frequency delayed retransmission in the audio which IMHO is harder to understand 'through' than reasonable (ie - public safety person using lapel mic/HT to talk while inside the car with the car's radio's volume reasonably low) feedback. I personally don't understand the actual need (other than paying the FCC/Congress's bills) for all this nonsense of tightening up tx bandwidth. My scanner says theres a lot less on vhf/uhf than there ever was before as the business users all migrated to cellular. Is the 'real world' trick to just apply for 2 side-by-side 12.5k slots and run your big fat 25khz carrier down the center? ;) JS From: Repeater-Builder@yahoogroups.com [mailto:repeater-buil...@yahoogroups.com] On Behalf Of Nate Duehr Sent: Wednesday, April 08, 2009 2:25 PM To: Repeater-Builder@yahoogroups.com Subject: RE: [Repeater-Builder] (OT) APCO P25 horror stories anyone? If "it's either there or its not" is supposedly a problem, you'll never see that go away in digital comms. You can bit-stuff FEC codes by the metric truckload into a signal all day long, but at some point, it just falls out. That's digital. Period. Anyone "surprised" by this must not be paying much attention to how these technologies work. MORE IMPORTANTLY: You can't LOWER channel bandwidth and retain audio quality and still have room for tons of FEC. Good old Nyquist and his mathematical friends... Hint: THE REAL ACHILLES HEEL OF P25 AND OTHER DIGITAL SYSTEMS... is that we're trying to implement them in LESS RF BANDWIDTH than the "traditional" FM analog signals. The natural progression should have been to convert to digital in the SAME channel spacing (expensive, and no economic gain, perhaps), and then start cranking up the compression over the years. The speed at which BOTH requirements were driven in, is painful for all. You mention that hams detect things using digital techniques, far down into well below the human ear noise floor on analog receivers. What you fail to mention is that those systems REPEAT the message over and over and over again. (WSJT, for example.) This isn't possible in a real-time voice system without using MORE bandwidth. It could have been done. Making a digital system that accurately reproduces "telephone quality" audio or better, and has multiple transmissions of the same bits (error correction) is certainly a no-brainer. (The entire telephone network runs on such technology these days... try finding an analog stepper switch at your local CO! Since they're not using RF paths that fade, and have other problems, their error correction needed is miniscule compared to an RF system.) Other's have also complained that the VOCODERs chipsets from DVSI introduce too much audio delay. I personally feel that complaint is a red-herring, since prior to audio delay on digital systems, there was always the "feedback loop howl"... same problem, just different timing and our ears aren't used to it. Once you get used to either one, you learn to move away from the loudspeaker... it's informal "training", in most user's heads. Never seen any public safety radio folks go out to a parking lot and show anyone how to avoid feedback howl, but the users figured it out from seeing PA systems and people giving speeches, etc. But many probably don't easily make the "mental leap" that digital echo on-scene is the same thing. And I doubt anyone's got time to show them how to avoid the audio loop-back noise either, but if they give the users a hint that it's CAUSED by the same thing, they can apply their old head-knowledge to the new systems. (Don't stand next to the truck that has the big loudspeaker turned all the way up!) The engineering needed to deal with background noise is that it's going to take some super-duper DSP heavy hitter math to get RID of it prior to feeding the VOCODER. That's going to cost some serious bux... It's already common in the telco central office and audio/videoconferencing world, especially as packetization delays have been added by the move to VoIP on the long-haul circuits. Conferencing/wireline engineers have much easier challenges though, and can do things like "ping" the room for acoustics at the start of a call... or slowly slide in an adjustment for removal of echo as the bitstream passes by in the most advanced echo cancellers in the world. (Ever notice how most conference room speakerphones "ding" or play little snippets of "music" or similar at the beginning of a call? You think it's just another "techno-noise", but the DSP is actually sampling the echos off the room walls, glass windows, etc... and adjusting the digital filters accordingly.) Unfortunately, the typical mobile radio user's environment is continually changing, they're not willing to pay $3000 a radio (well, some did... that was dumb), and that won't be as easy for the engineers to find the noise in math, and remove it, while still retaining voice quality. It'll take a while. And that "$3000" leads to the most important concept, and why the lawsuits are flying. This stuff's expensive. I think the REAL issue is economics. Even though everyone here is a "techie" and analyzing the system's technical isuses -- the real question is... "In order to double our channel space in public safety, is it worth four times the cost?" (Ironically, cost vs. channel spacing maps out pretty closely to cost vs. RF power... it's a logarithmic function, if you ask me from a general knowledge of the price tags of these things.) VOCODERS and systems COULD have been deployed that did NOT cut the channel spacing roughly in half, and they would have sounded much better, worked much better, etc... but not only was it a system requirement to use less RF bandwidth, there are now laws in place saying "everyone's gotta use less". Less is not always more, and this is an excellent example. Let the engineers build something that uses the full 16 KHz channel mask (or more!) of a "traditional" analog FM repeater, and the digital WILL kick the analog's butt. ESPECIALLY if it can adaptively change the audio quality as the signal gets weaker. (Nothing like that exists today for most of us... I'm sure someone is doing it, but not on a large scale.) The other thing that continues to be unfortunate is that the overall standard forces a particular CODEC. DVSI is making out like bandits, but there ARE some industries (telco) that have some really heavy-hitter CODEC mathematicians who just MIGHT be able to whip their butts for audio quality, etc... and lots of CODEC choices out there that can't even be tested/tried. The standard is "stuck" with DVSI's "products". That was a mistake and will lead to newer "standards"... the manufacturers already are headed there with each creating their own proprietary systems... but many went the cheap route and continued to use DVSI's chips/CODEC's. P25's use of CODEC needed to be allowed to evolve, over time. In the conferencing world, where I work, a new audio CODEC is released that "changes everything" about once a year. (And we won't even talk about video... those CODECs change even more rapidly.) And bandwidth limitations are a large part of the problem too. It all inter-relates. Nate WY0X
<<image001.jpg>>
<<image002.jpg>>
