Friday, 18 March 2016

audio proof of concept David O Mahony

Author: David O Mahony
Team: Embedded Systems Design
Date: 18/03/2016

this post will outline the results of this sprint. the audio was broken to two parts the text to speech and the audio mask.

Text to speech:
the text to speech monitors to see when it is told to speak from the different camera setting. the text to speech is demonstrated to be working with the color detection side of the project in a blog to follow that is a combination of work between myself Luke and Kamil. while only the color speech is shown, the code for the face and object detection are simple to implement as they rely only on reading the outputs from the other detection setups. the code will also be used in the next sprent with gestures to give greater life to the robot. The improvements to come will be straight forward  as during the testing process today bugs dealing with the speech was fixed.

Audio mask:
this code was made by building upon working code from sprint one. I added in a random generator that would pick a random sound effect each time the robot moved and also was turned off when the robot would speak. the code once run with the other setups in thread code was quite buggy. one bug was it would keep playing after it had been told to stop by the Text to speech. once this was patched and re ran the code would only run once. further work needs to be implemented to discover why this is and to also lower the volume of the sound effects. i For see that fixing this code will not be to tedious now that the code has started to be combined errors that were hidden from running it solo have surfaced. the code is may likely be caused due to its reliance on a motor to tell the code if it is moving and when this is set to constant true was causing the code to lock up this will needed testing with the servo thread.

i am great full to have been given the chance to improve further the audio setup. I enjoyed the challenge  of  giving the robot more life and personality witch i belive the text to speech gives the robot life and a slight british accent to how it talks. i look forward to the day when the robot is fully together and can see how the robot acts when it is in different modes. i felt let down by my code for the audio mask and will work further in the next sprint to fix the errors that have arisen. i do not see this as a problem and will be fixed. i feel like i did my best to stick to the deadline set out by the team in the original team post with the only slip due to bugs arising due the other combination of codes. 

thanks goes to Luke and Kamil for their advice and guidance in helping me combine the code with the master threads

David O'Mahony entry 6,

No comments:

Post a Comment