hi friends, I would like to share an experiment with you. hoping that you like it and drive it to further developement. Using "GlovePIE" , with build-in voice recognition and a self written simple script I got the dialog with a virtual raceengeneer working. As a prove a concept I made this video: http://youtu.be/UkqiCAsBQVg The advantage is, that anybody can modify the script and put more functions into without any programmer knowledge. If you like it, let me know and I will public the script.
It sounds great! If I may suggest something...take a time, finish it and release when it reaches the final stage No beta, no 0,01 version
I think that your program shows lots of potential! I'd love to see you make a finished version and post it here!
well, it is NOT MY programm. I only use available speech recognition (Win7), speech to text enginge (acapela Ryan) and controler device handling with voice processing (GlovePIE, freeware) to talk to my mate ingame. GlovePIE is running a self-written script which executes some key strokes or simply give me an answer. Lets give you an idea with an example: // Dialog to announce my next pitstop if var.sr then if said("pitstop I come in now") then keyboard.O = true // key O is assigned in rFactor to toggle pitstop wait 10 ms keyboard.O = false if var.voice then say("understood, we are waiting for you at the end of this lap") end if end if ... Ok, this is a "rule" in the script and anybody is able to modify it or invent new ones ... no mystery. So I would like to encourage you, to download my script (release soon) and start experimenting and enjoying. But in the meantime you should start to train your windows speech recognition. Invest about 10-15 minutes to improve the recognition ability of windows and you will have more fun with it (~100% recognition rate).
GlovePIE script to handle dialog with virtual race engeneer in rFactor/rfactor2 ______________________________________________________________________________ this is the readme.txt (included in download): written by GauchoRS, Brasil DISCLAIMER: THIS IS INTENDED TO ENHANCE IMMERSON UNTIL ISI RELEASES THE INGAME VERSION. Changelog: V.100 - 25.2.2012 - first english release what you need: ============== - Windows7 with trained speech recognition. You should have done the training exercises. May be that it will work in XP with MS SAPI5 installed, but it was not tested. - I recommand another text to speech engine, W7 build in "Anna" is too bad. For example "Ryan" from acapela infovox3, which I used for this demo exploit. the demo can be used 1 hour and will rework an hour after any windows restart. download here: http://www.acapela-group.com/infovox3-37-3-speech-solutions-tts.html - GlovePIE programm installed to run my script, emulating the virtual engeneer. download here: http://glovepie.org/glovepie_download.php how to prepare the virtual engeneer ========================================= Start GlovePIE, load the script and run it (click "Run") You may automatize this by using the parameter -filename_of_script.PIE in the shortcut of the program Run rFactor(2) and assure that some keys are assigned which are used in the script: O - toggle pitstop H - chatwindwow the Arrows - for manipulation in the pitstop menu TAB - toggle driver names CTRL SPACE is defined as 'toggle HUD' from the trackmap plugin by fazerbox (must be installed separately) how to use the dialog during the rFactor session: ================================================= OBS: Before start talking you have to press the joystick button 8 (top left on my G27). This may be changed in script. Release button after you finished talking. This keeps the programm from interpreting other/more commands. OBS: To initiate the dialog, your race engeneer must be wakened up with: "Ryan listen" or "hello pit" This will set the program to "voice on" and "listen for commands" state Now you may start talking to your race engeneer (press call button on wheel!) For example: "allright I go out now" just before you exit the box Ryan will answer: "ok, all systems up, now its up to you " Keep in mind that the script has predefined commands/words, and you have to talk them exatliy as defined. Up to now the dialog is linear, which means there is only one answer to your command. But this you can change easily using the command structure of GlovePIE. If you want that GlovePIE dont reacted anymore to voice commands (execpt "wake up call"), say: "pit keep quiet" "Ryan keep quiet" pitstop: =========== "pitstop I come in now" announces the pitstop to your crew. rFactor opens per default the pitstop menu (LCD at bottom right) with focus on fuel. the fuel choice is your point of reference, because any manipulation is done frome here using the arrow keys. First of all you may aks for less or more liters of fuel. I decided to use to commands: "5 less" or "3 more" just to have little amounts to adjust it. You may also make a decision of your tyre changes: "no changes" uses the down and left key (4x) und puts the tyres on "no change" "change all four tires" uses the down and right key (4x) und puts the tyres p.ex. from "dry" to "rain" After a tyre change command you can use this command to return to fuel control: "fuel" uses the up key (4x) and put the focus back to fuel administration Now you are able to modify fuel or tyre changes again or avoid repair action with: "no repairs" to command the repair of an EXISTING damage to "no repair" It is important that a damage exists, showing "Repair All" in that line If you change your mind you may cancel the pitstop "cancel pitstop" will command your crew back to the garage. more commands: ================== I implemented some more commands using keys in rFactor2: "telemetrie on" and "hud off" to toggle the hud of fazerbox´ trackmap plugin "what did he say" to toggle chat window on/off with H "display" to toggle the LCD menu with 6 "who is that" to show the driver names with TAB "bull****" to open the chat window with T = talk There are a few comands, only provocating a reply from your engeneer. They are meant to enjoy the dialog facility and increase your motivation only. Up to now the replies are linear, means that there is only one answer implemented. But this can be improved by your own. "alright I go out now" to exit the garage "talk me into position" to start a countdown 3-2-1-stop at the end of formation lap "first light" to start an advice for the start sequence "any rain during race" "who the **** is the guy behind me" "I can hear something strange" "what about the noise" "may be the tires will fail" "rock n roll" "thank you" If you want to mute the voice of the engeneer you use: "voice off" or "silent" or "Silence" or "shut up", but still the key toggle will work To switch the voice on again you use: "voice on" or "speak" or "talk to me" Copyright: ========== ok, now its your turn. start exploring the dialog facilities and enjoy it. I would like to encourage you to develop your own modifications and new facilities. greetings from Bernd in Brasil __________________________________________________________________________ Download my GlovePIE script here: http://www.mediafire.com/?cc7fhhq8yaoor9j
Bernd, nothing to say other fantastic work! One of the most useful addons I've seen on the rFactor scene. Thank-you for sharing this great addon
Holy C! - Potential here... Avast Antivirus gives me Trojan warning on download of your script from Mediafire ? Or perhaps from a popup, not sure...
I am still learning about GlovePIE ... what I am missing is an "array" like variable(x), with whole integer values to assign different answers and choose randomly. Here my first workaround to get random answers: // choose a random answer if said("hi buddy") then var.ans = Random(4) if var.voice then if var.ans = 0 then say("hi my mate, are you alright?") end if if var.ans = 1 then say("hi Bernd, you should find 2 decimals in sector 3") end if if var.ans = 2 then say("hi, we have problems with the brakes, be carefull") end if if var.ans = 3 then say("hi buddy, are you ok, what about lunch together?") end if end if end if I would like to encourage you to invest some time into GlovePIE to invent new facilities and let it work more efficient.
Googling 'free text to speech' gave me list of several free and/or open source software. Hopefully they also can be used like Acapela's infovox3.
what you need is a VOICE for your text-to-speech-engine in Win7. I have tried to find a free VOICE, but did not had success. To do this demo exploit the 1 hour free demo voice of ACAPLEA infovox was fine and it is all you need to experiment the facilities of GlovePIE. Hopefully ISI will have a 'race engeneer' or 'spotter' included in the gold version or someone will adapt the existing spotter plugin for rFactor1. In the meantime, we can have some fun with this GlovePIE scripting .... I like to use speech recognition on my daily work, invoking email and listening the content, invoking programs, writing blogs speaking my ideas to word programm, using free Windows Speech Recognition Macros. Therefore I bought a licenced voice (round about 35 US$) A good online shop offering much is: http://www.nextup.com/ have fun
ok so this looks promising and amazing... Is it possible for windows xp to do this? I mean I assume it is but I don't want to download everything only to find out I can't. Not afraid to download extra programs. I will check out if it is possible and report later...
in the meantime I have studied more in the manual of GlovePIE. There is constated that GlovePIE also works witch Microsoft SAPI5.1 in Windows XP
i try it but for me it doesn't work... i have installed the ryan's voice, restarted my pc and downloaded glovepie and your script.. i run the program of ryan's voice (infovox3 demo keyspeech it's ok?) and i clicked on run button on glovepie, but it doesn't work... i can't hear any voice after i speech... to run the program correctly, we must have the kinect?
So, to sum up, this is what we need (please correct me if I'm wrong): - Windows Voice/Speech recognition is enabled (you can talk to your computer) - Voice engine with SAPI (default from Windows will work, but yes, it's not good) - GlovePIE with custom script for rFactor (theoretically, GlovePIE will work in any game you want!) side note: I can dictate in Wordpad/MS Word, but Windows' speech recognition is almost useless in day to day activity, especially if you're not using Microsoft "standard" apps. I'm using OpenOffice suite and Windows' speech recognition is only able to open and close app. I also tried to dictate this paragraph in Mozilla Firefox but no success. Hmm, maybe I should use GlovePIE custom script for Firefox?
@taufikp: yes, you had summerized it correctly Dictation is sometimes not easy. First you have to train speech recognition very well. Than you need to use a good noise-reduction headset microfon. There should be a quiet environment und you should talk fluently .... something which helps a lot is do train the speech driven correction facilities, because the proposal to correct the wrong word(s) very often includes the correct spelling. But yeah, I use speech recognition most to open my mails, open some interesting sites in browser, doing some searching in google our youtube or wikipaedia, invoke skype or miranda or get some teasing from me "virtual friend". If have done a short video with a selfmade voice of my "virtual friend" - sorry but this in german language. http://www.youtube.com/watch?v=ft-pwE0dTTk @Zizza: you should enshure, that you have a profil in windows speech recogniton and the right language choosen (english). There is also a option to activate speech output. Its highly recommended to traing the speech recognition by executing the training programm allthere. In the chapter "text to speech" you should select a voice (Ryan) and after that you shoul listen the voice of Ryan talking a standard sample. If windows speech recognition and voice output is working, make shure that you press the correct button on your wheel (joystick.button8) or change the numer in the script. You should have the Logitech Profiler running (this garanties that windows receive the "button pressed" signal or better go into your game (rfactor) Try to command "voice on", GlovePIE should answer with "voice on" reply. The FIRST command to speak in GlovePIE is "hello box" or "Ryan listen" to wake up your buddy. Good Luck
By enabling 'dictate everywhere' in Windows Speech Recognition option, I can now dictate in any app. But man, what a troublesome dictating! This speech recognition is having great potential, those Star Trek episodes comes to mind very quickly. Actually Apple's Siri is much closer to Star Trek's computer than Microsoft's Speech Recognition. By the way, how to enable the 'audible feedback' in Microsoft Speech Recognition? I have this option checked/selected, but no audible feedback heard at all. PS Sorry for being very off topic, but I found this speech recognition is a very promising "add-on" for rFactor (and other games too, imagine giving voice command to your soldiers instead of typing/pressing keys), so that's why I want to explore more about MS SR (speech recognition) capabilities. edit: This is the sample of voice commands in American Army (from GlovePIE forum)