On Sat, 02 Sep 2006 14:40:40 +0200, Mxsmanic
wrote in :

Perhaps this is a naive question, but: Why don't voice radio
communications for aviation use FM radio instead of AM radio?

I presume the reason stems from AM radio's introduction into aviation
after CW was used prior to and during WW-I. The cost of re-equipping
all aircraft with new radios is also not insignificant.

I realize there's substantial inertia in the installed base of AM
equipment, but surely one could allocate some new frequencies to FM
and use them in parallel for some years to ease the transition.

AM frequencies are currently 25 kHz wide. FM would require more
bandwidth. Regardless, where would you place these newly allocated
frequencies?

The reason I ask is that improper and misunderstood radio
communication is a leading cause of accidents,

What is the source of that questionable statistic?

and so it seems that anything that can make that communication clearer
would greatly improve safety.

Hence the popularity of Active Noise Reduction headsets.

I can barely understand what I hear on the radio.

Do you use an ANR headset?

It is true that the communication is very standardized, making it easier
to guess what is being said, but the results are pretty unpleasant if
one guesses wrong.

Request 'say again' if in doubt.

On a related note, it has occurred to me that one could develop
voice-recognition systems that understand the speech of a pilot and
then repronounce what he says in an extremely standard synthetic
voice.

What would you estimate the cost of re-equipping all aircraft with
such a system might be?

This could also improve understanding, especially for
non-Anglophone pilots who speak with heavy accents. The same systems
could clean up the speech so that it is absolutely standard, with no
missing or added words. Of course, the issue here is that the system
would be stuck if it cannot recognize what is being said, or if a
completely non-standard utterance is made by the pilot. A natural
extension of this would be systems that recognize standard phrases in
one language and translate them to another, but that would be even
more dangerous if the system ever failed.

Pilot: "Oh ****!"

Electronically rephrased: "Mayday!"

Still another idea is special training systems that listen to a
pilot's speech and transcribe it, and point out any problems with
understandability. Again, this would be most useful for
non-Anglophone pilots, but it would work for anyone. If a machine can
understand a pilot's speech clearly, then a human being should
certainly be able to understand it that much more easily.

I can understand you frustration with non-standard phraseology and
foreign accents, but given the current state of the art, such a voice
recognition/synthetic voice system as you suggest would probably be
unworkable not to mention costly and short lived. I would expect to
see data-link equipment (ACARS* or more likely ATN** or NEXCOM***)
available for GA aircraft soon.

* http://en.wikipedia.org/wiki/ACARS
** http://www.tc.faa.gov/act300/act350/
*** http://www.faa.gov/nextgen/nexcom/Publib/aboutnc2.htm