Thursday, July 10, 2014

html5 + jsc3d + face3D_chung animation demo

here you can find a demo of html5 canvas + jsc3d + face3D_chung 3D obj model animated avatar from photo (free open source) + google translate TTS text to speech + eliza like chatbot (pure javascript , no WebGL required )
(TTS function: Chrome only, Firefox & ie11 : not always)

=> _face3D_chung_say_drive.html 
=> _face3D_chung_elizabot.html    (eliza based online chatbot )
=> _face3D_chung_sashabot.html  (idem with another avatar )
=>   womanjean_elizabot.html  (idem with lowpoly avatar )

 "*" => repeat last input
  ""  => autochat
(17/07/2014) now you can enter only the first letters of words
(19/07/2014) eval(...) key added from javascript eval() function , so you can insert pieces of javascript program allowing all javascript Maths String and Conditional features in key/pattern/result (see 'whats my name' / 'my name is' example in myelizadata.js)
(20/07/2014) var elizaredir added, if not empty, resubmit with input=elizaredir (20 times max) see test example


  1. Hi Chung, I would like to invite you to where other people including myself are working on things like this. I'd really like to hear more about how you got this working. Also please could you tell me what voice you have used ? Hope to see you soon. Regards, Freddy (aka Roger)

    1. for the 3D models i use a custom version of jsc3D which loads .obj text objects (base,eyeclose,mouthopen,moutho) then interpolate to create animation , for the voice i use "google translate" tts feature and html5 audio (audio.src=googletranslate url+text)