[HCRA] FW: [VWS] Pondering a D-STAR purchase.

Sun Oct 15 16:51:51 EDT 2006



David Matthews wrote:
> Not sure why the first attempt didn't get through the mail reflector 
> OK - but here it is again
>
> -----------------------
>
> Hello all -
>
> Experimenting with newer technologies can be lots of fun, but in the end
> the questions I always seem to converge on are:
>
>    a) What is it /really/ useful for?
>
>    b) Are the practical benefits clearly worth the investment in time
>    and money (and possible loss of interoperability with hams still
>    using older equipment)?
>
> I've recently been wondering these things about D-STAR digital
> technology.  This email will just discuss the issue of *2-meter* D-STAR
> digital voice communications versus conventional 2-meter FM voice.
>
> I recently ran across an MP3 audio file on the internet which gives a
> very practical real-world example of how 2-meter D-STAR digital audio
> performs.  The original 10 minute recording is KC5ZRQ mobile as heard by
> a 2-meter D-STAR base station.  The mobile starts out at the receiver
> site and then drives into some RF-challenged areas known to cause
> problems for 2-meter FM.  To put the results in better focus, I've taken
> the liberty of editing out most of it and producing three short
> segments, each of which illustrates a specific point:
>
>    a) The first segment is 35 seconds long and using an excellent
>    signal with a clear line-of sight path.   Because the signal
>    strength is very high, the /average/ audio quality is very good, but
>    it's interesting that there is nevertheless a distinct audio glitch
>    just at around 20 seconds into the recording.
>
>        http://www.djmatthews.com/1_-_DStar_strong_mobile_signal.mp3
>
>    b) The second clip is 65 seconds long and gives a sense of how the
>    signal degrades as the mobile goes through some roadway underpasses.
>
>        http://www.djmatthews.com/2_-_DStar_weak_mobile_signal.mp3
>
>    c) The third clip is 45 seconds long and is done at the very limit
>    of 2-meter FM coverage.  The stations switch between digital and FM
>    under edge-of-coverage signal conditions.
>
>        http://www.djmatthews.com/3_-_DStar_vs_FM_weak_signal.mp3
>
>
> Based on these clips, it seems that there can be some clear advantages
> to using Digital voice instead of FM on 2 meters, particularly
> considering that this superior performance also requires less RF
> bandwidth (12.5 kHz) to accomplish.  I've heard some people say that
> using digital audio on 2 meters instead of using FM gives you the
> practical equivalent of a 10 db power gain.    However, there are also
> some disadvantages to be considered.  As the first recording showed, the
> digital vocoders can glitch and produce audio corruption even under some
> strong signal conditions where you would never have noticed a problem on
> FM.   Environments that have multipath flutter which is tolerable on FM
> may produce a much less acceptable result with digital audio.
>
> Although it's not demonstrated in this set of audio clips, the vocoders
> also tend to degrade audio much more noticeably when the source audio
> isn't clean voice... specifically, when there is high background noise
> or other background voices at the source.  Because the vocoder
> algorithms are optimized for voice, they produce warbling quantized
> audio which some people will find subjectively more distracting than the
> original noise would have been through an FM system.  From a
> psychoacoustic standpoint, the human brain is equipped to deal with
> impulses, dropouts, hiss, fades, and naturally occurring background
> noise on the source audio.  However when the vocoder tries to handle
> these imperfections and then turns them into completely uncorrelated
> audio, the listener's brain has to struggle a bit more with sorting out
> the garbage from the useful audio.
>
> I'll have to admit there's just something that bugs me about the audio
> quality under average conditions.  On the other hand,  I know that this
> is exactly what a lot of hams said about SSB when it started to displace
> AM.  Maybe I just need to get used to it.  And maybe the coding
> algorithms will improve with time.
>
> It is claimed that D-STAR is an open protocol and therefore anyone is
> free to build systems using this protocol, but this claim may be
> partially incorrect.  From what I have recently read, it appears that
> the vocoder firmware used to make D-STAR carry digital voice is a
> proprietary product owned by DVSI (see www.dvsinc.com) for which the
> manufacturer of the equipment must pay a licensing fee.  There are
> open-source (free of charge) codecs available such as SPEEX (see
> www.speex.org) which would do a good job and, if adopted by the ham
> community, might considerably lower the cost issues associated with
> going to digital voice.
>
> I suppose we can accept the fact that a commercial radio manufacturer
> already has experience with the digital protocols used for commercial
> radios, and is therefore likely to use those same protocols for the ham
> equipment it manufactures.   Now it gets interesting....  on the DVSI
> website  they're already promoting their new AMBE+ vocoder (see
> http://www.dvsinc.com/papers/toll.htm) which outperforms both IMBE (used
> by most commercial digital voice radios as part of the APCO-25 standard)
> and AMBE (used by the current Icom D-STAR implementation).   In other
> words, Icom's current line of D-STAR digital equipment - which is not
> even approaching wide usage yet - is based on a vocoder standard which
> is already facing obsolescence.
>
> Since I'm currently in the market for a new 2-meter transceiver for home
> use, the D-STAR recordings above have convinced me that there's enough
> practical benefit to digital audio to buy a radio that can later be
> upgraded to use digital audio.  However, due to the current cost of the
> digital voice modules (much of which may be embedded vocoder licensing
> fees) and the fact that it appears that standards may change again in
> the next year or two, means that I'll probably hold off for another year
> or two on buying the digital module.
>
> I'm not a design engineer.  But it seems to me that if Icom and others
> are making radios with a socket to accommodate a digital voice
> module.... and open-standard vocoder software would enable hams to build
> much lower-cost and field-upgradeable modules to fit in these sockets
> (instead of purchasing the AMBE modules from Icom), then hams have a
> great hack opportunity here.   On the other hand, Icom might code their
> firmware to detect a non-Icom digital module and refuse to work with
> it.  Some of you may recall that HP did this with some of their inkjet
> printers - - the firmware was designed to detect refilled ink cartridges
> and refuse to work with them.  If I recall correctly, the U.S.
> Department of Justice saw to it that this anti-competitive behavior
> didn't last very long.
>
> If anyone knows of any writeups of /real-world/ data applications
> currently using the 2-Meter D-Star data standard please let me know.
>
>
> 73 de Dave K3MV
>
>
> P.S. - The original 10 minute MP3 file be found at
> http://members.cox.net/ic-2200h/  but it doesn't contain anything
> significant that's not in the short excerpts above.
>
>
>
>
>
>
>
>
> --------------------
> For information on subscribing or unsubscribing to this or our other 
> mailing lists, please see the Vienna Wireless Society web site: 
> http://www.viennawireless.org/lists.php