The images provided in this post are of the final building, built in Blender. Every item in the scene has been coded with the appropriate logic and psychics, as well as the mouse triggers. As mentioned previously, this has been based on the Leeds Town Hall, as parts of it have been built using Portland stone. This was an epic build, but I was pleased with the outcome.
However to my horror, when the game is activated, the frame rate is about 4fps, as supposed to the 50-60fps that it should be. To confirm this issue, the model was uploaded onto a file sharing website, so that Jon and Chris could confirm this. This issue was confirmed and so the group began to look into both the nature of the problem and any possible solutions.
We all worked on this problem. We sent emails to people for advice, plus looked into forums, databases and manuals for information on this problem. All group members were in constant communication about any information we found or possible areas that we could tweak. We as yet have found both an official cause and a solution. The only conclusion we came up with is that there is simply too much in the game and with each component having a physics and logic script, the game simply cannot process the data to a usable speed.
This is the major problem with the final product as it is a game and requires the ability to interact with it. So as it runs at 4-6 frames per second, it is currently a game which cannot be played. As a lot of work has gone into this, we will continue to look into the problem and seek advice, then respond constructively. I feel that although this issue is a major problem, it has demonstrated a good working structure, and shows that we were all working collaboratively towards a common goal. I am aware that this is how we should be working throughout the process, but this was a personal highlight as this process actively showed it in motion.
Worst case scenario, if it is a matter there being 'too many models in the scene', we could reduce the amount of items in the scene, or recreate the final model in a much more simpler version. However, both these options are drastic and would result in a product that would not be using the game engine's full potential. However, if this is what it comes to, I would rather provide a game/product which works than a good looking product that does not.
As Chris has continued to look into the lag problem, and solutions for getting it to run, Jon has now been tasked with looking into texturing it.
No comments:
Post a Comment