[Bragi]

Hello everyone.
I just published a new post on my blog about the new direction that the project is heading to. 
Looks pretty cool he?

[Freddy]

Pretty cool indeed, I think you have opened things up to a much wider audience now

I'm looking forward to developments and thanks for the compliment on the visimes

P.S. liked the moving hair too 

[squarebear]

I like it a lot. Is it possible to pipe any text through her? I'm thinking along the lines of a server based AIML bot that speaks the responses.

[Bragi]

I like it a lot. Is it possible to pipe any text through her? I'm thinking along the lines of a server based AIML bot that speaks the responses.

Thanks. No not really. Pure AIML will have to be transformed a bit, though the brainfiles are in xml format, the pattern language is more like that of the PF (but more features, easier to read), so the pattern matcher works a little different then that of AIML (more specifically, no unnamed wildcards, a pattern can start anywhere in a sentence, more operators, great and easy memory functionality ...).
The idea is that you edit your bot in the windows app and then, when you are ready, upload the brain to your website. Though that bit isn't implemented yet.

Also, the current animations can still be improved upon. there is a way for the characters to move the heads while speaking, just takes lots of renders and some tweaking of the engine.

[Bragi]

I've made a new video of the second character:

Pretty cool he.
and here's a new post with some more details of the project: http://janbogaerts.name/index.php/2011/08/11/mika/

[Art]

Very nice job with Mika!! Animations were kind of brief but quite smooth and fluid.
The character has a cuteness that soon grows on the user I'd imagine. Definitely
worth developing to a further degree, especially in the AI department.

[Bragi]

Thanks. I also liked the drawings very much. Yes, the current demo doesn't show much AI, I primarily made it to show the guy who made the drawings how his work looks like in motion.

[Freddy]

Looking good   I like how you have the eyes working now too, did you recode the character engine ?

[Bragi]

Yes and no. The continuous eye movement while talking was already possible in the previous version. Under the 'backgroundAnimations', Tara has a definition for the hair animation. This can be used for the eyes as well, but with a different loop style. The hair is a 'VariableJojo', but there is also 'FrontToBack' and in the latest version 'VarTimer' which can be used for the eyes. (all still very much undocumented stuff). But I did rewrite the char-engine a couple of times (4 in total), to get the visemes working as good as possible for as many voices as possible. It turns out that, for some voices, there are subtle differences between the 'managed' and native speech api's (usualy, the native version has better timing    ).
In fact, when I think about it, every part of the app has already been rewritten a couple of times.

[Art]

Jan,

I was thinking that Mika would be an ideal character to read books, stories, poems, etc., to children.
They would very easily grow attachments to her character and personality (when developed).

Just some food for thought.

[Bragi]

Some of here expressions also reminded me of my (what's the opposite of uncle?).
Just thinking out loud: if you have the story/text in digital form, you could transform it into an xml file, with proper tags that the bot understands (topic files), add some ssml tags for the facial expressions and import that into the bot (or import, then add tags, also possible). and voila, a story telling bot.
Perhaps with each chapter or page divided as a topic, all linked, to be told in sequence, with a pause in between. I could probably create a default template topic file for that.

[mendicott]

There are lots of ePub format readers (text-to-speech) available now for tablets and mobile apps (Kindle etc).  The real challenge is making "interactive talking books" that children can ask questions to.  ;^)  And, its not just about making one book interactive, but about a system that can digest any book (ePub) and convert it into a knowledgebase for the conversational agent.

[Bragi]

True, though you have to start somewhere.

[Freddy]

Yes, don't try to run before you can walk 

[Bragi]

I've made 2 new videos. the first shows an initial AI feature: complete the sequence, which is all done using a single rule and some thesaurus look-ups, more info on my homepage. (probably best viewed in full screen, hd, so you can see the letters).

The second is a small debug session that shows how the network performs this little trick (neural code). this video is narrated (my terrible voice).