(Originally posted under :
viewtopic.php?f=1&t=2325 I wrote this to offer comparison between features and technical function of P25 / D-Star digital voice modes.)
When you listen to Digital systems, you will find a "different" kind of distortion than what we are accustomed to in analog FM operation.
In analog systems we typically see "picket fencing", "white noise" and sometimes another type of static resulting from "multi-path", which is a result of the receiver - yours or the repeater's - hearing the signal in addition to a strong "reflection" of the signal, that arrive at the receiver at slightly different times.
As radio-heads we become accustomed to this and in some cases don't even notice it, at least until it gets bad.
The lack of this interference is what makes us notice how "clean" digital sytems are.
Plus, the vocoders in employed in digital radios inherently ignore most of the background noises as these noises don't follow the speech patterns that the vocoders are designed to map. Thus the background noises are dropped during the mapping process leaving (mostly) just the speech.
Fire fighters have observed problems with strong background noises seen at Fire Ground ops. When background noises get too bad, it really messes with the vocoders. A few years ago a national Fire organization recommended moving to analog at firegrounds.
In comparison, the distortion that you will hear with digital systems is quite different.
Of course the most obvious "distortion" is what we experience with digital signals like our digital cell phones, and that is - you are either IN, or you are OUT. (Ever had THAT happen on your cell?)
The other, and more common digital distortion you will hear is what many refer to as a "waterfall effect". This is commonly referred to "digital distortion" and I have also heard people describe this as digital "artifact".
Whatever you call it, you will likely know it when you first hear it. You will be listening, and all of a sudden (or at the beginning) the audio will sound garbled. And I mean completely unintelligible. You will know a message is occurring, but it is like syllables that were never meant to go with one another.
This occurs when there is too much error on the receive side, and is measured as "Bit Error Rate" (BER) at the receive end.
P25 products by specification are supposed to offer intelligible audio with as much as a 5% BER on the channel.
The reason I say "kind of" when P25 was compared to D-Star is, D-Star does not always work at 5% BER like P25. Same statement in different terms, the digital distortion occurs more often in D-Star than in P25, in similar RF conditions.
The reason is quite simple. P25 utilizes an incredible amount of "Error Correction" as compared to D-Star, and in fact employs numerous methods of error correction simultaneously. D-Star does employ error correction, but not as much and not as many methods. D-Star's error correction is very basic compared to P25.
As such, D-Star is more prone to that "waterfall" effect, where you will know someone just made a transmission, but you will also know it does not sound like any language a human could ever speak.
There are several reasons for this difference. First, P-25 was designed to work in the Public Safety environment, and I think most people here know exactly what that means. The lives and safety of the users may be directly impacted by the quality (or lack thereof) of the radio channel. This should be a no-brainer to us. As such, the development of P25 led to implementation of huge amounts of error correction.
The other (and technical) reason why this so is that the channel rate (speed of the data stream) in P-25 is 9600 baud, whereas the channel rate of D-Star is half that - 4800 baud. This allows much more bandwidth available to utilize error correction.
Error correction eats up a LOT of bandwidth. This of course depends on the error correction method used, as some error correction methods use more bandwidth than others. Typically, the more robust the error correction, the more bandwidth that method will require. My memory may be off, but I believe at least 3 methods are applied concurrently in P-25 voice transmissions. The reason multiple methods were used (instead of just 1) is to balance the overall use of the data stream (bandwidth). In P25, robust error correction is applied to the "most important" parts of the data stream, whereas lesser robust methods are applied to portions of the data stream that are of lesser importance.
The IMBE vocoder employed in P-25 can output the data map of the voice in three (3) outputs, breaking out "most important" portions of speech mapping, "middle importance" and "least important". Generally speaking the "most important" portion includes the part of speech with most change in speech patterns, whereas lesser and least important have lesser changes in overall speech patterns etc...
Another way of comparing overall use of bandwidth that is afforded in P-25 vs. D-Star is to look at the combined voice / data capabilities of each mode.
In D-Star, you may stream approximately 900 baud of serial data WHILE you are talking. Some people may say this is lame, but consider this occurs in a channel that fits into a 6.25KHz channel space, wheras P-25 occupies 12.5KHz, twice as wide as D-Star, while analog wide-FM (like we use in analog HAM) takes 25KHz - which is four times as wide as a D-Star channel.
In Comparison, P-25 only allows a TOTAL of 88 baud of simultaneous serial data while you are talking. No kidding. P-25 uses DOUBLE the channel rate of D-Star, and offers only 88 BAUD data in combined voice / data operation, compared to D-Star at half the channel rate, in half the channel space, and yet offers 10 times faster data!
The reason is simple, D-Star employs much less error correction. Therefore there is more room for simultaneous data but is more prone to digital distortion.
In the very early (pre-release) stages of D-Star error correction was not employed at all. This version offered just under 2400 baud of serial data and used 2400 baud for the voice side. It was never released as it simply failed due to too much error in "real world" mobile environment (hey it worked GREAT on the bench!). The final version of D-Star that hit the streets still uses 2400 baud for the voice side, allows approx 900 baud for data, uses a little bit for "call setup", leaving approx 1400 baud for error correction. MUCH LESS than P-25, but still it has error correction which definately improves D-Star operation in the "real world" mobile environment.
If you are interested in looking at the P-25 data stream and learning more about how it is assembled, Daniels Electronics wrote a very detailed manual that breaks the entire voice data stream down to the bit level. This manual can be downloaded at the following link:
http://www.p25.com/resources/P25TrainingGuide.pdfThere is not as much information available about the composition of the D-Star data stream, but it is out there. JARL and ARRL each publish a not-so detailed document that gives basic info on the data stream construct. Not down to the bit level that the Daniels doc does. I think one needs to purchase a license for the D-Star from JARL to get that level of info.
Please do not interpret my comparison as any dissatisfaction on my part with D-Star.
I LOVE IT! I use D-Star regularly, and when one looks at the overall functionality (feature set) of D-Star it goes way beyond capabilities of "end to end" feature set of P-25.
Perhaps the most significant aspect of D-Star that blows away P25 is the "plug and play" gateway functionality that allows any of you with a D-Star radio to simply enter my call sign (N7QOR - or the callsign of person you are calling) into the "your call" field of your radio, hit PTT, and BOOM, you are connected to me (or whoever you are calling ) ANYWHERE in the gateway connected network, worldwide.
Just like NEXTHELL, (but without screwing the spectrum - like NEXTHELL!).
The D-Star gateway system "observes" where - as in what gateway-connected repeater - a call sign is last observed, and when another HAM calls that callsign, it is auto-routed to that repeater. PRETTY DANG SLICK if I say so myself.
I need to look at the numbers, but a few months ago the list of gateway connected repeaters was over 150 worldwide and growing fast.
Sorry to be so long winded, but I am a radio-head and I like to "share the wealth".
MTM was my inspiration when I was a child, his books got me started in radio.
73,
N7QOR
C U on D-Star
Login
Register
FAQ
Search
