Go to content Go to navigation Go to search

T*int · 17 October 06 by Ray Crowley

I am currently working on the design of an augmented reality project. ‘Textspace’ seems to have been used for a couple of things so I am calling this project T*int (Textspace Is Not Textspace).

Essentially this will be toolset which could be deployed for no other purpose than the following artistic goal: to blur to boundaries between the discrete linguistic acts of reading|writing listening|speaking. This defamiliarization of our core linguistic competencies may lead to greater self awareness in the actants.

Here are some early design scribbles:

??The graphic environment would be purely vector text based.3d audio
processing and transmission is critical to the immersive experience. The complex element in this environment is a linguistic processor required to recognize wordstress, intonation, tone (such as sarcasm etc), class etc. Scenarios: two users in an room with 3d sound system (not collocated). Users wear head mounted displays. Speaker makes an utterance. This is processed by the linguistic processor which assigns attributes such as colour, font, size etc based on tone, volume, speed and also splits the utterance into raw grammatical elements such as pronoun, adjective, etc. This is then feed to the graphics processor which applies the graphic styling and sends this to userB’s headmounted display in Sync with the audio. UserB responds.??

This has lead me to carry out an overview of speech recognition parsing engines. Seems my original concept was a slightly ambitious :)

So now the immersive experience will be based on pre-parsed texts (extracts from Beckett, Joyce, a poem by Ginsberg, a newspaper article). Accordingly it will be primary a passive experience with the potential for basic user commands rather than half-duplex near real-time dialogue between two human users. The work continues..

this has provided a good spring board into the depths of langauge processing:

Language Technology World is an ontology-based virtual information center on the wide spectrum of technologies for dealing with human languages. It is a free service provided to the R&D community, potential users of language technologies, students and other interested parties by the German Research Center for Artificial Intelligence (DFKI).

tags: beckett, hci, linguistics, project, vr