KEYSPEECH PROFILES- FOR ALL CALLSIGNS!!!!
-
He’s right. He’ve actually explained that to me and that’s how I did it.
Be sure you have the last version. It won’t work otherwise.Roger…see below. I’m doing that without success…maybe it only works in the live (non-trial) version? We’ll see what support says. They are OOTO until 18 Aug, and I appreciate the fine support already, while support is on vacation. That’s service and dedication!
-
Hi Dawgboy, I confirm you that some DX buttons cant be assigned, In my case I cannot assign any button from the throttle identified by windows as id#2.
Funny enough, all the other devices works ok.I contacted the developer and he is aware, but its been more than a week since I dont get a reply for him
I am using the trial version, so yes it could be that the live version works ok.
My current workaround is to use JOytokey, to send the intended DX button as a keypress to keytospeech.
Except for this little thing, I am very happy with keytospeech and I will buy it, I am sure too that George will solve this issue very soon.
Kudos also to JCenzano, my contry fellow, for discovering me this nice app and his great profile that adds a lot to the inmersion!
Hope it helped.
S!
-
Hi Dawgboy, I confirm you that some DX buttons cant be assigned, In my case I cannot assign any button from the throttle identified by windows as id#2.
Funny enough, all the other devices works ok.I contacted the developer and he is aware, but its been more than a week since I dont get a reply for him
I am using the trial version, so yes it could be that the live version works ok.
My current workaround is to use JOytokey, to send the intended DX button as a keypress to keytospeech.
Except for this little thing, I am very happy with keytospeech and I will buy it, I am sure too that George will solve this issue very soon.
Kudos also to JCenzano, my contry fellow, for discovering me this nice app and his great profile that adds a lot to the inmersion!
Hope it helped.
S!
Thanks, ALAS! JoyToKey worked well for me to implement a stop-gap key mapping for PTT. Now, I see I need to train the voice recognition a bit better. It thinks I’m saying Century about 1/3 the time when I say Sentry. Saying AWACS worked every time, tho. I want to confirm, I always need to have the Win7 Voice Recognition listening, right?
Many thanks to you, MrWell, and especially jcenzano for laying the groundwork, as well as the KeySpeech Team. This is much more realistic and hands free.
-
I want to confirm, I always need to have the Win7 Voice Recognition listening, right?
That’s a negative.
You just have to start keyspeech.
Keyspeech will use Windows voice recognition and its profile, but it will be transparent for the user.You still need to train your VR recognition profile to enhance the accuracy, but when u are playing only keyspeech running is needed.
-
Thanks, ALAS! JoyToKey worked well for me to implement a stop-gap key mapping for PTT. Now, I see I need to train the voice recognition a bit better. It thinks I’m saying Century about 1/3 the time when I say Sentry. Saying AWACS worked every time, tho. I want to confirm, I always need to have the Win7 Voice Recognition listening, right?
Many thanks to you, MrWell, and especially jcenzano for laying the groundwork, as well as the KeySpeech Team. This is much more realistic and hands free.
Mmm I am using jcenzano profiles for keyspeech, and despite it is a huge amount of commands that keyspeech has to sort, and my non native english accent, I had a 92% of accuracy, and the rest 8% a second try is enough.
Be sure to not have a noisy environment, and also check if your microphone boom its not too close, so sounds like “S”, “CH”, don’t cause a noise.jcenzano, as you are in contact with George, could you suggest to him to enable in the profiles an alphabetic tab sorting/grouping?, so for example if I need to charge Weasel 1 profile, I could click the “W” tab and dont have to wait for the loading of all the list of profile names in order, which at least in my case is too slow to be comfortable.
S!
-
Mmm I am using jcenzano profiles for keyspeech, and despite it is a huge amount of commands that keyspeech has to sort, and my non native english accent, I had a 92% of accuracy, and the rest 8% a second try is enough.
Be sure to not have a noisy environment, and also check if your microphone boom its not too close, so sounds like “S”, “CH”, don’t cause a noise.jcenzano, as you are in contact with George, could you suggest to him to enable in the profiles an alphabetic tab sorting/grouping?, so for example if I need to charge Weasel 1 profile, I could click the “W” tab and dont have to wait for the loading of all the list of profile names in order, which at least in my case is too slow to be comfortable.
S!
He changed the way profile are loaded some months ago.
In the previous versions the profile list was there instantaneously.
My (non educated) guess is that now keyspeech is checking if each profile is “valid” before displaying it on the list.Waiting a little bit for the profiles to load is the price we have to pay if we want them to be double checked by the program.
Again, this is a guess since I don’t know exactly ho the program works.
I am sure that dev follows this thread, but I can send him an email asking why the list takes that long to load and wether he can do something to change it.
On my computer building the whole list takes around 15s. I am from the generation of C-64 with datassetes. I think that helped to develop my loading time patience.
But if he can do something do speed the process up will surely welcome.[emoji3] -
He changed the way profile are loaded some months ago.
In the previous versions the profile list was there instantaneously.
My (non educated) guess is that now keyspeech is checking if each profile is “valid” before displaying it on the list.Waiting a little bit for the profiles to load is the price we have to pay if we want them to be double checked by the program.
Again, this is a guess since I don’t know exactly ho the program works.
I am sure that dev follows this thread, but I can send him an email asking why the list takes that long to load and wether he can do something to change it.
On my computer building the whole list takes around 15s. I am from the generation of C-64 with datassetes. I think that helped to develop my loading time patience.
But if he can do something do speed the process up will surely welcome.[emoji3]OK, thanks!
My suggestions for next version:
Maybe an override check option, or adding a flag so the code so it only checks the profile the first time is loaded.
A quick search box to input the profile name manually and have it selected -
That’s a negative.
You just have to start keyspeech.
Keyspeech will use Windows voice recognition and its profile, but it will be transparent for the user.You still need to train your VR recognition profile to enhance the accuracy, but when u are playing only keyspeech running is needed.
This is good because it was driving me nuts with the VR listening to everything I was saying, even when not on a hot mic. Thanks!
-
Hello, I’m back from vacation. About the profile list, I can add validation only when profile is selected.
-
Hello, I’m back from vacation. About the profile list, I can add validation only when profile is selected.
Hope vacations were fun.[emoji3]
What are the advantages of checking the whole list?
If there are no pros on that, then it would be more “comfortable” to validate only when selected.
At least for the BMS users who are using my profiles.
Thanks for the support.
-
Edited- double post. Deleted one
-
And where is the good one? Lol
-
This post is deleted! -
Everybody uses JoyToKey ? Because I don’t need this … I’m using Saitek X52
I do not use it either. I own hotas cougar and there is no need for it with the dx support.
-
Hello, I’m back from vacation. About the profile list, I can add validation only when profile is selected.
Hello all, I have upgraded KeySpeech to 3.0.4. Notice that all profile’s files are moved into «Profile» folder.
-
Hi,
I had this tool on my todo list for far to long but final came around to set it up and tweak and play with it.
Thus far accuracy is not even in the double digits.;-)
This could be many things. I just gave Windows VR 4 lessons and this could be insufficient. Will test some more.The three things I wanted to mention are:
1 In the trail version I still need to fill in my mail address when the program is started, despite having added the trail license file which was send to me when filled in the mail address for the first time. Pretty annoying I must say.
2 In setting the accuracy correct in trying to solve the hitrate, the mouse-over window on the taskbar icon shows me STATUS: Enabled, Accuracy: 0%. This despite having set an accuracy percentage in the settings/Accuracy tab. Is this because of it being a trail version?
3 And above posts mention JoyToKey when setting PPT’s. The two options available in my version [3.0.4.0] are shown in the picture below. I guess this is an update and with JoyToKey one now has to choose Push To Talk!Now lets see if we can get the accuracy up to 95%. Am Dutch, but my English is not that bad.
Cheers.
-
It is a separate program.
So far I’m enjoying Keyspeech and having good sucess with the commands I know but I’m still learning . Overall they are very similar to my profiles but with more pbrase options for the same command so it shouldn’t take much more time. But I did notice that there is a few Sentry commands that seem to be different from the rest of Chalis, GCI, AWACs, etc. There are a few commands thst have the callsign and numbers and then “from one” so you end up saying “Sentry Avenger one one from one, picture.” This seems redundant but then again I haven’t been listening to RL comms lately.
I see the issue off the added words “from one” still present in v 2.3.5. Is this therefore a none-issue?
Example:
awacs fury one one from one declareThis can be found on all AWACS related calls: awacs - sentry - challis - cyrano - magic - cgi
In checking this further I came to learn that when correcting/deleting the “from one” words, the new phrase already excists. This resulted in 189 duplicates.
So why are these present? Is there a function to them jcenzano? Can I delete the “from one” phrases?Cheers,
-
@Ice:
Hi,
The three things I wanted to mention are:
1 In the trail version I still need to fill in my mail address when the program is started, despite having added the trail license file which was send to me when filled in the mail address for the first time. Pretty annoying I must say.
2 In setting the accuracy correct in trying to solve the hitrate, the mouse-over window on the taskbar icon shows me STATUS: Enabled, Accuracy: 0%. This despite having set an accuracy percentage in the settings/Accuracy tab. Is this because of it being a trail version?
3 And above posts mention JoyToKey when setting PPT’s. The two options available in my version [3.0.4.0] are shown in the picture below. I guess this is an update and with JoyToKey one now has to choose Push To Talk!Hi Ice,
- Once you have a license, you may just click on “Continue” button only. But you have a very good point, I will hide this box whether license is already there.
- Accuracy on the taskbar is your score of your last recognized command then you may use this score to adjust your accuracy slider.
- You don’t need JoyToKey.
Thanks
-
@Ice:
I see the issue off the added words “from one” still present in v 2.3.5. Is this therefore a none-issue?
Example:
awacs fury one one from one declareThis can be found on all AWACS related calls: awacs - sentry - challis - cyrano - magic - cgi
In checking this further I came to learn that when correcting/deleting the “from one” words, the new phrase already excists.
Not a bug. Those are some inherited commands from a sqn mate. Originally they were just planned to be
“Awacs from one, declare”
But because of the system I used to build these repetitive callsign phrases, the weird commands you are talking were built too.Just disregard them. They won’t hut you. Just use the ones that make sense.
Personally, I never use them, since the normal commands are present too.
I always try to use command like “chalis, falcon 11 request picture” or whatever callsign I have.Regarding accuracy slider, I am not native speaker but my foreign English is well above average and I found that the setting that works best for me is to fix it around 91-92%
Higher than that VR gets really really picky, even for native speakers. Keep in mind that any outside noise getting in through your Mic will distort your voice sound wave, making it more difficult for the VR to find a match.
-
Thanks KeySpeech, George, jcenzano,
Great support George. The issue mentioned is already fixed with versio 3.0.5.0. :bdance:
Still working on getting it all in 96% working order. Tests in Wordpad show promising, although I do need to talk slow and pronouce standalone words with detail. Hope this is because Windows Speech Recognition is not up to par yet on my system. Started the lessons yesterday and have done 3 sessies of the two lessons.
jcenzano. Looking at the command and phrases structure, I see inbound calls to tower on the default -request taxi- (ground)function. I have tested these but do not get a command returned on the phrases “approach say delay” and “approach say landing number” when airborne and inbound homeplate, although these suspect otherwise. I guess this is role-playing?
And I must say outstanding job. It’s all there. Thanks.