Matthew Law writes: > On 03:15 Mon 29 Dec , Ivo wrote: > > Or we could have multiple people around the world recording the sentences, > > so we'll hear the right accent when approaching for example New Delhi or > > Mexico City or Frankfurt. Maybe even bilingual, though I don't know if they > > use their native language (for example for domestic flights) or that they > > use English worldwide. > > According to the ICAO, all ATC comms should be in English. Quite rightly however, > most controllers use their native tongue unless talking to international flights. > > This sounds like a cool idea but the work involved is immense. The majority of it > is non-technical (recording sound samples etc) so it could end up being much more > authentic than MS FS if we were to use our diverse user-base :-) >
In reply to all, Currently the only ATC voice is English female speaking the non-country-specific FlightGear default ATIS, with airport names from most of the UK and the base scenery recorded. For me, the main bottleneck involved in extending this is recording, editing and indexing the sound files. If, as it seems from this thread, there is interest from others in doing some recording and editing that would be just great - I'd be happy to code up support for additional voices as a matter of urgency if folk were producing them. At the moment ATIS is the only service with a phraseology that can be easily read - just look in default.vce in the ATC dir in the base. The numbers are byte position into the sound buffer of the phrase start and byte duration of the phrase. Note that some words are run together to make phrases using underscores - currently those phrases are hardwired in. Ultimately I'd like to separate the phraseology out from the intent into xml files (Alexander Kappes started this), such that for a given intent, eg turn right heading 220, or give the weather as part of an ATIS transmission, the phraseology for a particular part of the world is looked up first, followed by the most appropriate sound file. That way sound authors could modify the phraseology without access to the code. Once again, the production of some sound files would spur progress with the code. If anyone is seriously considering doing a sound file for a given service (tower, ground, approach, ATIS currently, I'd add UNICOM if someone was recording it) then give a shout and I'll post the phraseology currently needed, and I'm sure the real pilots will add some as well, and I'd code support for it as necessary. If someone wants to do a locale specific ATIS with different phraseology that would be great as well - I'd code support in quickly. As for recording the stuff, currently we're limited to 8bit, 8KHz, mono, at which setting the voice is noticably deteriorated in quality. I believe that Bernie is working on improved sound support, so it might be worth mastering and editing at higher quality, indexing by time rather than byte location, and converting to low quality and byte position at the end. I've been cutting and pasting each phrase from the original to a new file to compress the finished sample as much as possible - it's still 5meg+ and that's at low quality for a fairly limited phraseology (interactive services like tower etc will need a lot more than pre-recorded services like ATIS). You need to produce a corresponding .vce file to go with the .wav file so the ATC system knows where to find a given phrase - see the description of the .vce file indexing a few paragraphs up. Selecting the correct voice sample for each country will be easier once country codes have been added to the airport records as proposed by David Megginson, but I could do a hack based on ICAO code for now. Note that if we get samples for multiple controller voices for the majority of countries at high quality this will easily exceed the current base package size! I don't forsee that being a problem for a while though... Cheers - Dave _______________________________________________ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
