I think it would be cool if they used some of the technology MS has in house with Cortana and applied it to the ATC in the sim so that you could speak with the ATC in a natural voice and have it understand you.

 

I like VATSIM and IVAO but you usually have to fly at peak times to get decent ATC coverage so it would be great if the ATC on top of having support for SID/STAR/holding patterns etc. supported voice recognition. 

that's where they want to go. But for me the voices still sound like a computer and don't offer the same thing.

1 minute ago, Pitbull2504 said:

that's where they want to go. But for me the voices still sound like a computer and don't offer the same thing.

 

I remember the ATC in Flight Unlimited II - those voices sounded pretty good for something released in 1997.

I hope they will build it the same way as vatsim. The new voices sounds really breathtakingly good. There will also be many new Simmers who are interested in it. If ivao and vatsim were not divided, there would also be better coverage 

58 minutes ago, highflyer2020 said:

I think it would be cool if they used some of the technology MS has in house with Cortana and applied it to the ATC in the sim so that you could speak with the ATC in a natural voice and have it understand you.

 

I like VATSIM and IVAO but you usually have to fly at peak times to get decent ATC coverage so it would be great if the ATC on top of having support for SID/STAR/holding patterns etc. supported voice recognition. 

MCE has had the capability to use voice commands to respond to default ATC in FSX/P3D for some years. Pilot2ATC as well. With Pilot2ATC you could purchase add-on voices from 3rd party vendors that very life like.

It is definitely doable. 

PS try PilotEdge for the best (and live) ATC experience.

 

I've been looking at implementing this for a few years.

The biggest hurdles (to me at least) was the voice recognition API. Windows built in speech recognition is almost useless. There are a number of 3rd party APIs that are VERY good, many working as a web API which means the more you use it the more you pay, so a monthly fee would likely be needed.

I've yet to find a standalone SDK that was any good (and less than several hundred $ per seat, even at recognising my own relatively accent free English voice), especially when you want background noise (engines, wind noise etc).

Having to pronounce each word slowly and unnaturally so that you are understood and still have to repeat multiple times is a no go.

The tech does exist now, but at a price.

44 minutes ago, b737800 said:

I've been looking at implementing this for a few years.

The biggest hurdles (to me at least) was the voice recognition API. Windows built in speech recognition is almost useless. There are a number of 3rd party APIs that are VERY good, many working as a web API which means the more you use it the more you pay, so a monthly fee would likely be needed.

I've yet to find a standalone SDK that was any good (and less than several hundred $ per seat, even at recognising my own relatively accent free English voice), especially when you want background noise (engines, wind noise etc).

Having to pronounce each word slowly and unnaturally so that you are understood and still have to repeat multiple times is a no go.

The tech does exist now, but at a price.

 

Yeah I always thought it was not possible via 3rd party devs for the reasons you described so maybe one MS can have a better chance at solving.

2 hours ago, b737800 said:

I've been looking at implementing this for a few years.

The biggest hurdles (to me at least) was the voice recognition API. Windows built in speech recognition is almost useless. There are a number of 3rd party APIs that are VERY good, many working as a web API which means the more you use it the more you pay, so a monthly fee would likely be needed.

I've yet to find a standalone SDK that was any good (and less than several hundred $ per seat, even at recognising my own relatively accent free English voice), especially when you want background noise (engines, wind noise etc).

Having to pronounce each word slowly and unnaturally so that you are understood and still have to repeat multiple times is a no go.

The tech does exist now, but at a price.

 

1 hour ago, highflyer2020 said:

 

Yeah I always thought it was not possible via 3rd party devs for the reasons you described so maybe one MS can have a better chance at solving.

 

We already have that for years. I have no problem at all talking to PF3-ATC via Multi Crew Experience, with a recognition rate of 95+ %, talking naturally as non-native english speaker.

 

So we will most probably also have that in the new sim, be it by MS / Asobo by default or by third party like MCE.

 

 

The newest TTS voices sound really good and natural. If they used enough of those to achieve sufficient variety and made them sound like they were radio transmission they would sound pretty close to real world ATC. That topped by decent voice recognition would be huge.

I did the Windows speech training thingy to teach my voice to the system twice and I can talk just as I would talking to a human being without much, if any, focus on the way I speak.

I'm all for voice synthesis for ATC, but I'm all against voice recognition if it is the only way to interact. By all means have it as an option for those that like it, but let the rest of us have our clickable menus and keyboard input for responses. I already use headphones with any games I play so that I don't disturb the household and I wouldn't want them to have to listen to half of a conversation between ATC an me.

7 hours ago, highflyer2020 said:

I think it would be cool if they used some of the technology MS has in house with Cortana and applied it to the ATC in the sim so that you could speak with the ATC in a natural voice and have it understand you.

 

I like VATSIM and IVAO but you usually have to fly at peak times to get decent ATC coverage so it would be great if the ATC on top of having support for SID/STAR/holding patterns etc. supported voice recognition. 

Nothing wrong with asking for (optional) voice recognition but please let's have an ATC that reflects the ATC present in the major flight simulation locations rather just US centric first. Voice recognition and location specific ATC accents can come in later versions.

I also agree that voice recognition would be great as long as it's not the only option.

Honesty, I've just been waiting years for someone to do ATC like Flight Unlimited III.  I would love to be able to do VFR flights with ATC from/to some class D and C airports in a modern sim.  Flight Unlimited had a great interface for "building" what you wanted to say and the responses didn't sound robotic, because it was recorded clips of real people.  With flight simulation, there is normally a (relatively small) "library" of phrases that normally gets said anyway, so recording real people's voices can work pretty well.

 

2 hours ago, Ian Box said:

I'm all for voice synthesis for ATC, but I'm all against voice recognition if it is the only way to interact. By all means have it as an option for those that like it, but let the rest of us have our clickable menus and keyboard input for responses. I already use headphones with any games I play so that I don't disturb the household and I wouldn't want them to have to listen to half of a conversation between ATC an me.

Agree entirely. A few weeks ago I decided to switch from button to voice after years using button for FS2Crew. Then it dawned on me that instead of the virtual FO being my assistant I was focusing more on the FO then properly managing the aircraft. Sometimes not getting a response because I forgot the exact phrase or not knowing for sure if he is programed to comply with a specific command. Though button control has fewer options then voice it's much more intuitive and you can be certain what you want to happen is happening making FO an assistant rather than you serving it. Sometimes it's possible to over engineer. So yes,options are always welcome.

I think ATC with Voice Recognition would be an excellent feature as an option. The tech is exists for a company like Microsoft to implement it in an exceptional way and would be great to have it as a feature option. 

48 minutes ago, PATCO LCH said:

Sometimes not getting a response because I forgot the exact phrase or not knowing for sure if he is programed to comply with a specific command. Though button control has fewer options then voice it's much more intuitive and you can be certain what you want to happen is happening making FO an assistant rather than you serving it. Sometimes it's possible to over engineer.

That's less a problem with voice recognition but rather with the design of FS2Crew. Give MCE a try. You don't need to remember exact phrases (all kinds of variations work) and you can also create your own phrases / commands.

 

Apart from that I agree that every voice recognition functionality should be optional.