to I went out to take photo for the next project. During setup, the sun was out and giving me amazing lighting. It took me about 5 mins to find a composition that I liked and to adjust my camera settings. But by the time I was ready to take my plates, a large cloud had pasted over.
In the meantime, I took my instant film camera around the nearby area and took some pretty cool photos.
I came running back when the sun started to peak out again. The light had shifted greatly and I had to quickly shift the scene down. Since it was a rush job, my focus and composition became slightly off. The sun did not last long. I was able to take a clean plate, and HDR before it was gone again.
Maybe if I had grabbed the grey sphere, these photos could have been usable. But the second setup isn't as strong as the first. I like the closeness of the first composition where it's like I'm about to draw a doodle of my CG object.
I did wait around for a couple more minutes. Maybe, just maybe, there would be one more chance to see the sun again. But, alas, the clouds became a dark shades of grey and thunder was booming from a distance. I quickly packed up. I do plan on having a reshoot in a few day to try to recreate the first scene.
Bonus: Behind the Scenes
Notes from Class
One issue I'm having that hasn't quit been figured out is with my anti-aliasing. No matter how high I increase the camera AA samples, the render looks the same. The only difference is in the render time.
Due to time constraints and seeing no visible change, I opted to keep the samples low.
Today, I had a pretty successful work session. Between this post and the last one, there will be a huge jump in progress. But really, I've been making really rough progress all week. Now, I have perfected them.
Compositing in Nuke
This final composite looks pretty good other than the shadow. The Shadow feels a little out of place.
the diffuse color, but I know what the real sphere looks like and that's a light grey color. So when I took the picture, the sphere is actually being blown out by the sun. In the end, I favored making the CG sphere feel more integrated over completely matching the light. For the next class, I will see what suggestions I should try to get them to match better.
Problems to Solve
way around this would be to paint the image used for this projection. The other option is to not have the sphere come in from the back.
There's no better feeling than aligning the camera with the cube plate. This is after a long battle with the CG camera focal length, CG Cube scale, and slight tilt on the Z-axis.
The focal length of the cube plate is slightly different than what I used in my clean plate. The cube plate has a focal length of 24mm while the clean plate has a focal length of 20mm. Otherwise, the angle of camera is exactly the same. I think that in the future, I might have an issue with the scale of my objects.
I had an interesting time taking my plates. I went out a mostly cloudy day knowing that the main goal of the shoot was to get hard shadows. I would have sunlight for about 30 seconds before a large cloud came over and blocked the sun for several minutes. Each time the sun was out, I'd scramble to get the right exposure settings on my camera and taking each photo. It was actually interesting seeing just how fast the sun was changing.
I used an auto-bracketing feature on my camera to get several images with taken with different shutter speeds. Since it was so bright outside, most of the picture became blown out. I could have adjusted my camera to a higher ISO. I did not because I had already finished taking the other photos, and I was afraid of loosing the quickly shifting sun again. I used Photoshop's 'Merge to HDR Pro' to combine all the images into one.
Behind the Scenes Setup
And then I bumped the tripod... But here's more a scope of the location.
Project 2 is all about matching a texture from a stone to a complex CG model and compositing it into a scene. Many of the examples use the Stanford Dragon. Fortunately, I know of a similar website that have several amazing scans of statues.
Deux Chiens de Meute à L’attache is a model found on Three D Scans that I believe features a lot of interesting complexity. (Click here for more information). The scan is comes from the artist Georges Lucien Vacossin crafted in 1911 .
The download is a OBJ file of a 2,499,989 tris of great detail and no UVs. The next step is to export the model to zBrush to convert the mesh to polys and reduce the face count. One idea to help bring the details but keep the face count low would also be to use zBrush to make bump and displacement maps.
I'm excited to start this project.
A new school year, brings new projects. This is the start of my breakdown blog for Technical Direction/Compositing or Tech Comp for short.
For our first assignment, the goal is to composite a CG object into a photographed or filmed scene. For the background plate, strong and varied shadows are required and the CG object must take up at least 25% of the frame. The idea in this project is to learn how to composite CG shadows with shadows from the plate.
This model was found on TurboSquid and was created by Rocket Sales Studio. The model was created in 3Ds Max back in 2013. Click here to visit the page for more information. I choose used the FBX. For a free model, this wasn't too terrible of a mesh to fix up. There are holes in the mesh that did not connect and the UVs are interesting...
The fixes were fairly easy to fix. I selected 3 nearby vertices and used the merge tool to bring them together. When projecting UVs, the model was created in a way that each strip's face could be easily selected.
The Next Steps
After the first class, I have much better idea of the goals of the project. For now, I will be focused on making a simple texture (probably made in Substance Painter) and finding a location to shoot my background plate.