> I'd be very happy to hear your opinion on VATSIM, as I've ben trying to
> push the idea of human voic ATC within FlightGear. I _do_ have some
> ideas how voice ATC could/should be realized but my ideas didn't fall
> on prolific soil,
OK, here's the breakdown. I've used VATSIM with FS2004+FSInn and
X-Plane+XSquawkBox.
Most importantly, make sure there is no text mode. On VATSIM, you can
go online either using text or using voice, which means that there are
users who are not aware of any voice transmissions, and also that
voice users are required to read text transmissions (which is a bit
much to expect on a manual short final, for instance). Last night, in
particular, I was faced with the situation that a text-only jet
airliner popped up in front of me from below while I was 4nm final or
so -- he had no idea that I had been cleared to land on the voice
channel, and ATC had probably forgotten that my Dash7 was a lot slower
than the jet.
This also means that UNICOM must be voice, not text-only as on
VATSIM. This is most easily accomplished by creating unmanned idle ATC
posts on given frequencies across the globe, running on the ATC
server(s).
On VATSIM, pushing PTT does NOT disable radio reception on that radio,
which is unrealistic. Make sure PTT turns the corresponding COM audio
output off. On the issue of realistic distortions -- can anyone tell
me what modulation air communication uses?
Latency on the voice channel is not a problem, since radio
communication is always PTT-based. So I would go for the simplest
option of using the speex codec and a simple TCP-based protocol (maybe
using the streaming parts from IAX2?). Just make sure it traverses NAT
routers transparently! And I don't think P2P connections are required
-- all voice had better go through the servers.
Which brings me to another problem -- the VATSIM ATC client needs open
UDP ports on any firewall/router that separate it from the
internet. This is bad as it requires reconfiguration of existing
network installations, so I think it's important to make sure
everything traverses NAT and firewalls transparently.
On VATSIM, there is a general bad habit of not checking ATIS. This is
probably because ATC are not really forced to issue ATIS, and there is
usually no voice ATIS (I think it must be recorded manually). Since
FlightGear already has code for voice ATIS, it should be possible to
create voice ATIS on all published ATIS frequencies. This can (should
-- as part of the ATC logon procedure) be set by the responsible ATC,
and in its absence the current METAR can/should be relayed. The latter
is required because otherwise you have no good way of working out
QNH/base pressure for an unmanned airport.
The last issue (though not very important) is visualization. Though
not a voice issue, it would have to be dealt with as part of the
system sooner or later. VATSIM pilot clients, to the best of my
knowledge, only send information on the position and attitude of the
aircraft, along with rates of change. It wouldn't be very hard to
include second derivatives (accelerations), which would, for instance,
make take-off and landing rolls look much smoother in case of network
load problems (I have seen aircraft on landing roll jumping back by
several 100ft on VATSIM...). I also don't like the way aircraft visual
models are dealt with in VATSIM: only a code describing the aircraft
type is transmitted, leaving the client software to find a suitable
model. I'm very much leaning towards each client submitting the rough
geometry of the airframe, including reference point, to the ATC
servers so that other clients can download and show them (this would
only have to be done at logon time and can be asynchronous, so I don't
expect bandwidth problems). Only the copyright issues would have to
clarified. As a suggestion, for copyright-protected airframes, one
could transmit a few main points on the airframe only so that the
other users can anticipate the extent of the aircraft without seeing
its exact shape.
What do you think?
Andras
|