After a week of getting hyped and ready for the Hackmanchester event the team travelled across the Pennines and booked in at the impressive venue of the Manchester Museum of Science and Industry.
With over 220 developers turning up on the day, including teams from We Buy Any Car, Sofology and the developers Redeye there was going to be some tough competition.
We started unpacking and fueled up on coffee from the guys at Zuto while we laid out plans for what we were going to do.
Despite pre-installing things getting connected to the internet and getting working proved to be a challenge in itself and 2 hours in some of the team where still trying to get a decent internet connection.
Our overall plan was to produce a system that could record part of a conversation and then stitch together recordings from a library to play back this conversation as a whole in an interactive way, gathering video and audio from various terminals and sending it back to a server, then playing back these samples at different terminals, building more complex interactions as it went.
We hit a number of issues with this as we went along. Our Raspberry Pi server did a good job, but the transmission speed to it was very slow. Recording the audio and video separately so we could analyse the audio to extract text we found that the video would record and then the audio would record consecutively rather than concurrently.
Implementing a push switch button to start and stop the recording on the Raspberry Pi we found that the voltage floated and that we needed to implement pull downs, but were familiar with the idea and not sure what to do with it .
Our finished submission video is viewable below, but with so many features only partly working we were not happy with what we managed to produce by the deadline. We did learn some important lessons for next time and also found some interesting topics that we need to research.
With over 220 developers turning up on the day, including teams from We Buy Any Car, Sofology and the developers Redeye there was going to be some tough competition.
We started unpacking and fueled up on coffee from the guys at Zuto while we laid out plans for what we were going to do.
Despite pre-installing things getting connected to the internet and getting working proved to be a challenge in itself and 2 hours in some of the team where still trying to get a decent internet connection.
Our overall plan was to produce a system that could record part of a conversation and then stitch together recordings from a library to play back this conversation as a whole in an interactive way, gathering video and audio from various terminals and sending it back to a server, then playing back these samples at different terminals, building more complex interactions as it went.
We hit a number of issues with this as we went along. Our Raspberry Pi server did a good job, but the transmission speed to it was very slow. Recording the audio and video separately so we could analyse the audio to extract text we found that the video would record and then the audio would record consecutively rather than concurrently.
Our finished submission video is viewable below, but with so many features only partly working we were not happy with what we managed to produce by the deadline. We did learn some important lessons for next time and also found some interesting topics that we need to research.
You can look over some of the code produced on the Hellobox github page.
Comments
Post a Comment