Day: April 26, 2018

When appending my assets into one file, I noticed a rather drastic change in the rendering times compared to earlier versions of my scene. While the scene is detailed, the rendering time took a drastically long time, especially since I was yet to apply textures to any of the assets present. The longest process seemed to be the synchronisation of the objects, so I turned to google to try and find some answers.

One thread I found suggested that it was an issue with RAM. Having attempted to render this mockup on my laptop, I chose to try the same on a computer with larger capabilities, but I was still left with the same issues.

Thread: 2.66, Cycles is Stuck Synchronizing Objects

After a lot of deliberation, I noticed that the values at the top of the Blender document were excessively high, something I was not sure of until I discussed the issue with my tutor. As I had made a point during my modelling processes to remove as many unwanted faces as possible, I assumed this value was correct, however, decided to check each asset again to find the source of the problem. In doing so, I found that the saucepan formed a large amount of those values, leading me to remove it from the scene.

Removing this asset radically reduced the rendering time, so my next task was to figure out how to include the saucepan into the scene without affecting the values. My first instinct was to recreate the saucepan, as the shape was rather simple and I was confident I could do it in a reasonable amount of time, but when discussing the issue with my tutor, he suggested I just import the OBJ I created in preparation for substance painter. I chose to do so, as I was rather pleased with the outcome of my model, and would have rushed to get another version finished in a reasonable amount of time.

When importing the OBJ, I was faced with some nodes I needed to remove before adding the textures I wanted. These were simple enough to deal with, as all I needed to do was delete them completely and paste the appropriate node sequence.

After making this alteration, I was able to export the following mockup in no time.

Final Mockup:

In order to gain an idea of what surroundings I wanted my final scene to have, I decided to append each asset into one document and place them in the formation I wanted them to take. When doing so, I realised there was a large problem with my saucepan, an issue I will address in another post, but for the time being, I decided to render these stills without the saucepan, as that smaller asset could be quickly added once the error is solved.

Intended Outcome:

• Finish designing and modelling the tent for my scene.

Problems I encountered and how I overcame them:

• When I began this project, I had already undergone a few testing and experimentation trials, attempting to create a simple triangular tent, and in doing so I learnt so much. When modelling the first tent, it was the first time I had attempted to UV unwrap fabric and I came across a few problems. When I first uploaded my mesh to substance painter, I was confronted with the below problem. During the modelling process, I was unaware that I needed to apply a solidify modifier to ensure the tent had normals on both sides of the sheet. Because of this, Substance painter did not acknowledge these sides, so I consulted my tutor during a masterclass and he taught me about the solidify modifier, and I was able to fix this issue.

• After solving the above issue, I felt confident in using planes to create the fabric for my tent. The initial image I used had a rather interesting designed tent, almost a gazebo type cover, which I felt fit the idea I had for my scene. Because of this, I focussed on following the reference image as closely as possible.

• To set the proportions of my tent, I began this model by designing the poles that prop the fabric up. To do this, I user rather basic tools in Blender, extruding, indenting and spinning a cylinder so that it had a hook at the top of the pole. In real life, this hook would piecer through the fabric, however as the top of the tent will not be shown in too much detail, I chose only to simulate this piece rather than replicate it exactly, and I was not confident in adding holes to the fabric that would have a cloth simulation.

• From being rather nervous to work with planes, I found that they worked just as they would if I was working with a 3D mesh, the complicated part of this process for me was pinning the fabric. I experimented heavily with pinning fabric before doing this project, however, therefore I was able to choose and apply the cloth simulation to act as I desired. I tested this with a wind force field to ensure it moved realistically.

• After adding the fabric, I added rope to simulate the pieces holding the tent up. I did not film this, therefore it is not included in my timelapse, but to summarise, I created a long thin cylinder, attaching it to the top of both larger poles, and ensured they were touching both the top of the pole and the ground, going off at an appropriate angle.

• To complete this model, I added a subdivision surface modifier to smoothen the surface, I bevelled any edges that were too harsh, and I ensured that the normals of the mesh had an auto smooth angle of 30 degrees. I also grouped together any items that would be textured the same.

Timelapse:

Final Model:

Intended Outcome:

• Create a realistic looking fire, ready to apply fire simulation.

Problems I encountered and how I overcame them:

• Creating the rocks came quite easy to me, as in my first media project ‘Changing creature in different environments’ I had experimented heavily with the sculpt tool in order to simulate natural materials. The main task for me in this portion of my model was not the creation process itself, but making the stones look similar but not identical. I was originally hoping to use my first rock as a base for the rest of them, but the resemblance was hard to change, so I decided to start each rock from a different mesh.

• Out of curiosity, I searched youtube for advice in creating rocks and stones and found the below tutorial uploaded by youtube member ‘Stym’, who used a rather complex method to create their game assets. After trying this method, I decided against it, as I found altering my rocks from an iscosphere gave me a more efficient and realistic outcome.

• After completing enough rocks to surround where the fire would be, I began modelling the centre. I chose to firstly add a large log to fit as the main fuel source, before building thinner sticks to surround it. To fill empty space and add detail, I also created a pile of soot by sculpting a circle so that it elevated unevenly, and consumed the log slightly.

• Crating the sticks that surround the centre was simple but tedious. I modelled the sticks by extruding and moving points of a cube so that it resembled the behaviour of a branch, before applying a subdivision modifier and sharpening the edges with a loop cut. I then propped them at the centre of the soot.

• To complete this model, I bevelled any edges that were too harsh, added a subdivision surface modifier, and I ensured that the normals of the mesh were correct and had an auto smooth angle of 30 degrees. I also removed any unnecessary faces/vertices and grouped parts of the model that would be textured identically.

Where to go from here:

• Create a tent to bring the scene together.

Timelapse:

Final Model: