Very nice. This combined with Sonic's new currentcity, zip, county and state should provide a load of information for our skins. I'll take it out for a pilot shortly.
Thread title says it all.
Will allow you to use the two street names from iGuidance in any RR screen as labels. The current street is optional, which you can choose when you run the setup.
It works with all versions: v2, v3 Reg, v3 UMPC
User specified label variables (defaults are given)
5-26-07
The setup has the following changes:
1. Minor bug fixed
The OCR app has the following changes:
1. Option to only read/update upon exiting of the guidance screen (on by default)
2. Few minor bugs fixed
To turn off the reading only after the guidance screen closes, under ‘Run’ section in the INI add ‘guidance=false’
Setup OCR.zip
5-10-07
The setup has the following changes:
1. No more default letter defs or scanning for it
2. Turn arrow color message
3. If learning file is found, jumps straight to teaching then exits
3. New Icon
4. Reduced Memory usage
5. Kill OCR.exe if running when started
The OCR app has the following changes:
1. Improved check if enroute
2. Empty columns for letter space detection exposed in ini
3. Learning on-the-fly ability added ( saves unknown scanned data to file for latter teaching via setup)
4. Asks if you want to teach the unknown letters if above file found.
5. New Icon
6. Reduced Memory usage
7. Current street does get updated (if turned on)with or without a next street change.
If you notice that you have some extra spaces where there are not supposed to be, like in the middle of a street name/number, then increase the 'empty_cols' setting under the advanced settings section by one.
(this just tells the app how many empty columns count as a space)
Also included is a OCR Toggle.exe - add to a button w/ a run command if you would like to have a way to start/end the app..
known issues still:
NONE!!
Setup OCR.zip
4-22-07
Added next turn direction label ( left, merge left, right, merge right, u-turn)
Added next turn direction text to use for each direction (user specified, defaults given)
Removed speaking section
Because of major changes, please delete everything in the folder where you were using the last version, EXECPT the data folder, unless you want to re-learn your letters.
Skin labels will have to be changed, as new version uses variables so that they will be self updating when changing screens, easier to just let RR handle this instead of polling RR and checking if you changed screen and update labels on each screen.
The current label variables are (user can specify when going through setup):
Next Street: “=$_NEXTSTREET_$”
Turn Direction: “=$_TURNDIR_$”
Current Street: “=$_CURRENTSTREET_$”
example skin label line
Code:L02,244,59,152,17,255,255,0,12,"Arial:B","=$_NEXTSTREET_$","CENTER"
4-13-07
Added next street speaking, runs when next turn label gets updated
To use add these 4 lines to the OCR.ini file
speak - True/FalseCode:[Speaking] speak= command= pre street=
command - Say/Talk
pre steet - text to be say before street name..ie next turn, next street, ect
speaking beta
Edit: file removed, needs more work yet.
Seperate OCR.exe file, only use if already went through the setup.exe
Edit: file removed, will not work with new version
4-8-07
Updates labels upon changing screens.
4-1-07
Here is the next version, now fully operational!! Just place the setup in the same folder and run through again. your shouldn't have to relearn your letters if you didn't delete the files, but I've added some of the letter combinations to the letters left to learn box.
Its at the end of its beta testing in its current state.
Doesn't do yet (DONE):
Send street names to RR
Doesn't check for iG window visability
Doesn't check if you actually have a destination set
Only real requirement is that your RR is no older than than the Oct release.
Setup is best run when its in its own folder, its not that picky just easier to keep all files in one location.
Let the comments/suggestions begin...
I would also like some comments on the flow of the whole setup process!!
RideRunner...The #1 FE, PERIOD.
Current Project: DFXVoice [Beta Released]
Next in line: RRMedia v2.0
DFX 5.1.1 Skin
Appstore Link
My RideRunner Plugins
RRMedia
RRExtended
DFXVoice
Very nice. This combined with Sonic's new currentcity, zip, county and state should provide a load of information for our skins. I'll take it out for a pilot shortly.
awesomeness
Car : 00' Chrysler 300M
Speakers : Rainbow Audio SLC 265 | Subs : Diamond Audio D6 12" Subs X2
Amp's : Hifonics ZXi 6006 AB | Hifonics BXi 1606 D
CarPC: HP Laptop AMD Sempron 3000+ | 1.5GB RAM | DVD+RW | 80GB 5400RPM HD | ATI R200 Video
Cool. This is just what I've been looking for! Gonna try it out now...
1997 Corvette
looks sweet i gotta try it out
im not running an epia but what sort of cpu usages does this take
.______
| '_ |__\___
[(o)|___(o)] XB
._________
| I__I I_I|_\__I
[(o)______(o)]b VanPimpin'
LostReceptions Apps D/L Here
GPSGasoline- Rewriting
Draw- SkribblePad for Touchscreens
iGQwerty-iG3.0 Qwerty Keyboard
CarPCNetwork
Right now, not quite as much when its finished. Only CPU load is when its scanning, but that doesn't take long anyway. I have seen times from 150ms on my desktop to 2.3s on various other comps. It also depends on the length of the street names.
After I finish, I don't really think that it will take up that much...I have kept the code as small and effeicent as possible, so that anyone will be able to use it hopefully!
RideRunner...The #1 FE, PERIOD.
Current Project: DFXVoice [Beta Released]
Next in line: RRMedia v2.0
DFX 5.1.1 Skin
Appstore Link
My RideRunner Plugins
RRMedia
RRExtended
DFXVoice
very nice, hats off!!!
some comments on the setup.
the save button doesn't work, i cant click it. Using enter works though.
Sometimes the letter that its showing me to recognize, well i cant recognize it all the time. Some of the letters are cut off so i had to guess what letter it was, leading to garbled street name in some cases. For example, ocr in red text, IG in blue
EXIT 44B IL-9
EXIT 44B IL-19
KENEXNEXEDY EXPY E
KENNEDY EXPY E
scanning only seems to work when ig is visible, is this expected at this stage?
Other than that, it seems to work very good.
shouldn't the letters be the same for everybody? Just curious why you have to teach it, its not a big deal though.
nice job
01101100 01101001 01110001 01110101 01101001 01100100 01011111 01110011
01101101 01101111 01101011 01100101
beer replenishment fund
http://www.mp3car.com/vbulletin/rr-faq/
mp3car live search
i have joost invites, just hit me up for one.
Funny, I don't think that I even ever tried clicking on it, as your hands are alread on the keyboard anyway.the save button doesn't work, i cant click it. Using enter works though.
Could you post a pic, so that I can see how I might address the issue.Sometimes the letter that its showing me to recognize, well i cant recognize it all the time.
The only letter that I came across was 'D' on the v2/v3 reg, because of the way I have it scanning. It cuts it in half and so I just used the first part and skipped the second. Thats also why I have it display what it has scanned and understands as you go along.Some of the letters are cut off so i had to guess what letter it was, leading to garbled street name in some cases.
Yes, I want to get the main point of this perfected first, as the rest is a piece of cake.scanning only seems to work when ig is visible, is this expected at this stage?
No, as I found with sonics testing, differences will creap up if you have the cleartype effect on/off as one example. And I figured that if it leads to everyone doing the learning, I'd better make it easy as poss.shouldn't the letters be the same for everybody? Just curious why you have to teach it, its not a big deal though.
RideRunner...The #1 FE, PERIOD.
Current Project: DFXVoice [Beta Released]
Next in line: RRMedia v2.0
DFX 5.1.1 Skin
Appstore Link
My RideRunner Plugins
RRMedia
RRExtended
DFXVoice
Cool. Tested out and seems to run good.
My "save" button works.
When going through the "learn" process, sometimes I get "double letters", like it would show "AV" for ave, instead of doing an "A", and then a "V" one at a time. Does this makes sense? Does it supposed to do this? I guess it happened only a couple of times and everything seems to display correctly when I test it -- just wondering.
Seems to take about 2 seconds to get the street name on my system. Not sure how much cpu time it takes up during the process, I'll check that out more later.
Thanks Blue! This little app is awesome!! Can't wait to get the full version!! BTW, when can we expect it?![]()
Gonna go out now and play with it a little more....
1997 Corvette
Yes, it is working correctly. The reason that you will get double letters is b/c the letters either touch each other or there is no vertical space between them. Its the way that I have it scanning, is the reason it happens, but it was the only way that I could use in order to get the kind of accurate street name resultsWhen going through the "learn" process, sometimes I get "double letters", like it would show "AV" for ave, instead of doing an "A", and then a "V" one at a time. Does this makes sense? Does it supposed to do this? I guess it happened only a couple of times and everything seems to display correctly when I test it -- just wondering.
There are quite a few of these: TT, RA, RY, KY, TA, XA, ect...
This is also why I worded the learn window text with "enter what you see"
Not sure, as I have only been working on this for the last two weeks, so I might say within the next two weeks, we'll see.BTW, when can we expect it?![]()
RideRunner...The #1 FE, PERIOD.
Current Project: DFXVoice [Beta Released]
Next in line: RRMedia v2.0
DFX 5.1.1 Skin
Appstore Link
My RideRunner Plugins
RRMedia
RRExtended
DFXVoice
Bookmarks