The End

All good things must come to an end...
Today I attended the final presentations via Skype as I was unable to make it to Sydney. I think I severely underestimated the experience of presenting with a team via Skype. I've used video conferencing in both a one on one and team environments before, but never as a presenter with half my presenting team on the other end. Being physically removed from my team (who did a wonderful job) made it difficult for me to follow on from points or to add some more information as per the presenting style that we had worked with all semester thus far, and as a result I felt I was unable to add a couple of key points. It was also difficult to know what was going on as I could not see or participate in the demonstration of the environment. In retrospect it would have been possible for both me and my team to be in the environment together as an online multi player map and follow each other around. Something obvious that I realised as we were demonstrating.

That aside, we posted the final environment, two videos and ten images to our Team Blog. Overall I was happy with how it all came together, however there were still many many things that I would have liked to have either added or improved upon.

To recap my contribution to this assignment I was focused, primarily, with the over all environment the building was situated in. This included constructing and texturing the terrain, vegetation placement, roads, surrounding buildings, AI and interactivity placement and flow graphing and capturing raw footage for the final videos.

As we had a clear placement point for our building, I was able to work on the environment, sans the building, until the other team members had finished constructing it on a blank map. At that point it was just a matter of exporting the building as a series of levels, importing them into my working environment and then grouping the building and dropping it onto it's final placement point. At which point I was able to add the extra layer of interactivity to the building to finish it off.

This allowed us to work more independently, a work flow that we discovered was necessary when working long distances with a project of this nature. Without the immediacy and quality of collaboration close physical proximity offers, working fluidly in a single project space which requires constant and careful version management meant that hundreds of megabytes of data would need to have been traded, almost constantly, and kept track of.

One of the major dangers of a project of this nature is information sharing. The nature of a Crysis project means that everybody needs the most up to date information, all the time. This is fine when working asynchronously on projects where either file sizes are smaller, or it is multi document in nature and easily split into separate areas. Crysis does not tend to this mode of sharing. There is no shared repository with version control which allows you to check in and out data. You have to share the whole thing, all the time, with everyone, and make sure that everybody is working with the same information. Otherwise inconsistencies and conflict can enter and potentially increase project time.

Dividing up the task into independent parts allowed us to work concurrently on our specific areas and then bring them together at major milestones thus reducing the potential risks of missing data, conflicting information or miscommunication.

This meant that we all held a clear piece of a puzzle, which could be put together faster and simpler rather than disjointedly taking turns to stir the same pot and trying to leave messages on what to do next. Or worse, realising we were stirring yesterdays leftovers.

Thankfully we never encountered any major problems with our work flow, the exchange, flow and incorporation of information was almost seamless. Albeit not as optimised as it could have been if we had been working desks, rather than states, apart.

Overall, I feel that even with the limitations of steep learning curves and time zones we were able to successfully collaborate and produce a piece of quality work that we can be proud of.

I would like to thank Michael, Joanne and Pui Pui for all their hard work on this project. Jeremy and Graham for giving us such a challenging and interesting course. And finally the rest of the class for all their shared knowledge, support and assistance throughout the semester which would otherwise have made this project many times harder.

-Gordon

Assignment 3 - Contribution

Again, as with the previous assignments I focussed more on the Terrain and environment. Below is a breakdown of my major tasks in this collaboration.

• Creating new Heightmap and terrain textures
• Importing some previous content from the Assignment 2 environment (roads, sounds, assets... etc)
• Importing final building model from blank environment
• Creation of new surrounding building assets.
• Adding vegetation
• Adding detail and assets to the environment
• Adding AI to environment and the building
• Adding and flowgraphing lights around the environment
• Capturing raw footage for inclusion in the presentation videos

Assignment 3 Work - Exporting previous assets

Assignment 3 Work - Video Capture

In discussion with Joanne and Michael I captured, compressed and sent a total of six separate clips to be included in the final presentation videos.

To capture the videos I used the following settings in Sandbox

Con_Restricted 0
r_displayinfo 0
r_nodrawnear 1
cl_hud 0
fixed_time_step 0.033333

capture_frames 1 (to start)
capture_frames 0 (to finish)

Whilst a deceptively simple task, it became very laborious with some clips having a turnaround from capture to upload of around 90 minutes. This was due mainly to the size of the environment and it's increasing use of system resources over time which led to many crashes and constant restarts. Misbehaving AI was another major problem with constant flowgraphing issues leading to many AI standing still or not behaving as desired during capture.

In the end I was able to send the completed environment back to my team who were able to assist in additional video and image capture as time constraints closed in.

All clips can be found in my MediaFire folder for this course.

Enjoy!

Assignment 3 Work - Exporting Building to New Environment

Assignment 3 Work - New Surrounding Buildings

Assignment 3 Work - Flowgraphing Lights

Flowgraphing Interior lights with proximity triggers

The goal of the interior was to reflect a Boulee's original intention of day becoming night and night be day. Unfortunately due to the limitations of Crysis' VisAreas we were unable to darken the space and remove the ambient light that is included in the environments. However, keeping with the idea of the original concept, I decided to add some proximity sensitive lights to the display panels on the upper floor. Connected to a proximity trigger the simply turn on when someone is near and then turn off again when you leave the area. The idea of this was to add a level of interactivity with the space whilst keeping in mind that those from the lower levels would probably be distracted from the a permanent light illuminating the upper level.

Attached is the flowgraph. And a demonstration of the result by video.

http://www.youtube.com/watch?v=FEPcXG0B2IQ


Flowgraphing Street lights to reflect the ToD

Adding lighting the path leading up to the building added an extra level of detail and the overall ambiance of the scene. However, street lights aren't always on, they follow timers, so to recreate that effect I used a simple flowgraph using the ToD as a trigger to turn the light and light beam particle effect on and off a sunrise and sunset

http://www.youtube.com/watch?v=IkF8PqK

Assignment 3 Work - New Heightmap

It was clear from the previous assignment that the Heightmap exhibited signs of possible exaggerated variance between some areas. The glaring example of this was the problems I had when trying to fine tune the river.



Which didn't quite line up with the reference images which suggest a more subtle fall to the land on both sides indicating that the height of each bank area is a similar height.




To look into this I went back to the original SRTM dataset and noticed something interesting. The terrain was clearly a lot flatter than the original height map seemed to demonstrate. With this knowledge I went back to Angel Island demonstration videos on Crymod and decided that the grey scale spread needed to be reduced. It seems that the original hightmap generated was using an elevation range of purely the cropped areas height range. This meant that it was dealing with a smaller range of heights, yet still mapping approximately 256 colours resulting in a more "detailed" but out of scale height map.

To fix this I regenerated the map but this time using the entire DEM's range which is usually up to around 290 meters. This resulted in a much flatter, but more ultimately more realistic terrain than before.

New Heightmap:


Previous HeightMap:


Result:

The resulting map didn't have to be tweaked an awful lot to add the water relatively accurately. The landscape only needed to be dropped or raised an average of 2m around the site of the water bodies for it to look similar to the reference images.

No more scary cliffs!

Assignment 3 Work - Adding Vegetation

Assignment 3 Work - Adding AI

Specialisation - Advanced Terrain (Posted to wiki)

I've posted my specialisation tutorial to the class wiki which outlines how to use existing data sources of real world locations to create terrain.

Originally I felt this would be a Beginner / Medium level tutorial, however I think due to the length and ability to adapt to custom application sets by focusing more on the concepts of each step I feel it is more a Medium/Advanced level tutorial.

I hope someone finds this useful in the future.

Big thanks to Zapwizard and 4k1r4 over at crymod for posting some pretty awesome tutorials which I used as the basis for this one.

http://www.mediafire.com/file/idujvmwm4xd/Advanced+Terrain.pdf