Blender Pecs.
Augmented communication device

Thank you Youle for the pyttsx help!

1 Like

This could be used to make a npc say what is the thing the player points to.

win 32 only ** @ the moment** https://drive.google.com/file/d/0B29hKf6-B52ceWdGTnQ5WUk2Z2c/view?usp=sharing

Sure Lost Science - and the way youle set it up I think I can even share it at the moment (it’s a zipped folder that you unzip, and save your blend into, and then it has pyttsx support*)

but linux / android support and testing is next after I get the UI shiny

import pyttsx
engine= pyttsx.init()

engine.say('any text you want it to say')


Freezes interpreter while running
Youle has a way to run on a second thread that works for his pc but not mine,
I was thinking a speech server /socket could run in a non blocking way in almost any PC.

Did you remove the last video?I am on a windows 64 machine.I don’t think that helps me.

edit fixed video, I had uploaded the wrong video on accident,

edit2 : just download 32 bit blender, it runs on win 64 just with less ram access size, which does not apply if you dont have a ton of ram anyway I think,

Next I will setup a autocomplete list using pythons autocomplete module,

tapping a option places the remainder of the string, forming a word using what was styped + (what was selected minus what was typed)

after that I will likely need help to port to android,

edit: after I finally got around to be able to code last night I was exhausted, then there was a fire less then 2 miles from my house, and a power cut that lasted all night, fun stuff…

So predictive typing for speech I may need help getting working as well

Hey BPR, I never understand what you’re doing (my bad English understanding is not helping) but I know your a master of coding :wink:

That is awesome BPR! What would happen if you just typed random letters in there that don’t spell anything, would it try to sound them out?

yeah it uses pyttsx which grabs a local liberary that is os dependant for text to speech :slight_smile:

There is a 32 bit windows zip above that should allow you to tinker with it :slight_smile:

It may need updating*

OK, thanks!

Now, I really need help adapting a predictive speech library and then porting to Android, then we can lower the barrier for entry for a real augmented communication application.

this program already has more functionality then proloquo to go, and will be free, where as prologuo2go is 250$ + an IPAD

ha ha ha

searching ‘predictive typing’ != “Autocomplete”


Blow the dust off this thread, I am firing it back up!

with new knowledge I can finally kill old bugs!
by adding new features!


Making a system to texture the cards by creating a uv atlas material of all cards during runtime :smiley:

Ok, I have the atlas generator working in bpecs core now, next up libNew a unique mesh per card, and set uv cords using UVdictionary.

I will try and get this working tonight and start building a small library of images.


finally! I can move onto polishing and GUI for B_pecs !!!

(nonverbal speech augmentation application!)


Ok, it looks like pyttsx is hard to distribute,

Looking into swapping to this https://github.com/asrp/python-espeak
Even has a android port.