14.10.15 Voice Recognition Test Using Kinect Micophones Built into Chassis

Thesis: I am still working on my appendices, I realised that I did not put my filmography into alphabetical order... school boy error. I tried to use word to put this in order for me but alas it just messed everything up so I am going to have to do all this manually.

Robot:

This session is focused on the vocal recognition software that will be used as human interaction interface. The response will be outputted into processing / eliza chat allowing the robot to respond to questions. I wanted to add this in as an optional extra, people always seem blown away with voice rec but I always find it difficult to implement in busy spaces. I am using the kinects internal mic set up for this and just changing the mic setting in my system prefs. The all I did was use a program 'google voice' in this case but you can use mac in built voice rec system. Then all you do is write a short bit of script that measures the pauses in the audio as a string of audio.. fox example 7 seconds of silence after an initial sentence and then you use this to trigger the enter command in processing. Again this only really works if you are in a quiet room as without silence at the end of a sentence it will not work. Audio contamination is also an issue when recording as loud noise can interrupt the voice recognition process. 

However, I really wanted to include this element as an adaption from the last project. I think once this technology becomes more adaptable and capable it will provide a very useful communication and interaction framework for animatronic character systems. 

Feelings: I am still waiting on parts from over sees and its really becoming a bit of a drag having to put things on hold whilst they are in transit. This is still making me feel pretty anxious in regards to the deadline of this project. 

Evaluation: I really like the idea of adding voice recognition into the project, but voice rec is still at its infancy stage and getting hold of software that is capable of delivering precise dictation is still .. not quite there yet. I am hoping that this will change quickly and that there will be a package available that can cope with multiple voice and contamination in time for my next project. It is a shame that this element does not quite function how I would like it to as it was covered in my thesis a possible advancement in animatronics and I really wanted to include it in this project. That said, it does work if you are in a silent room, but if the application of this system is for filmic engagement and interaction getting perfect silence is not really going to be an option. 

Analysis: The current situation of the project makes sense, the delays are sort of forcing me to think of other things that I can be prototyping and adding to the model while i am in wait of my package. 

Conclusion: In Conclusion there is most certainly use for voice recognition software in animatronics and maybe even such platforms as interactive theatre and other live action formats. However, the technology is not quite good enough yet to implement this in a working model. I am however, really please that I did try this and the fact that it does work is something that is worth expanding on in future projects.

Action Plan: I will continue to work on various parts of the project until my shipment arrives. I will look at writing the script for the hand system it should be pretty straight forward.