Week 36: Texturing Design Ideas

 As the guys and I discussed the creature's texture, we thought it was would be good to give him a pattern (as discussed last semester) to emphasise his fictional/magical nature.


Throughout the process of creating the texture, I whipped up a few designs in mudbox to mull over with the guys:

We additionally discussed stripes, so I decided to try this out firstly. Neither of our hearts rested on this design, however.

In this, I was looking at simplicity in the design. I wanted tit to be striking but not so much as to take away from the character's design. The symbol on the back appealed to me and I felt something similar to this could be a possible winner.


In the early character designs, we looked at a more complicated design on the character.

In this case, I feel the use of more softer colours will be necessary. I did try to maintain a touch of the minimal approach as I had in the previous example. However, this design naturally required a little more detail. The team and I felt that the simpler the design the better.









With this one I was going for something a little different from the others. A little more futuristic, still simple, but somewhat more suggestive of another world. We did however, still feel that the triangle design, in its simple but edgy style made more of a statement, without taking too much from the character. So this is the route I decided to go down for the final design.








For the final design, I stuck with simple, but tried softer, more organic shapes. It was unanimously agreed this was the design to take on to the next stage.

Week 33: Test Shot Final

The Completed Test shot for the Mastronauts project:

Still image.

Node Tree.

Shuffling elements from EXR.

Merging the elements on top of the footage.

Writing out the the composition.





Final Test.

Week 31: Compositing the renders

I did eventually manage to find out the issue I was having with the shadow layers. Although the Final Gather was off, the IBL set up was still affecting the alpha, which I needed to get the shadows. After running through the possible factors causing the issues in my head, it became clear this was the culprit, and after a little connection breaking, the problem was resolved.

So I moved into Nuke to try compositing the shots together...I've been trying to get my head around the 'shuffle' and 'shuffle copy' nodes and trying to understand the most effective process to take when rendering for the compositing stage.

Clearly, I am still trying to get the hang of the node system and understanding exactly how each node affects the next. Although this rendering/comp stage is certainly a challenge, I am actually enjoying it. I just need to get the hang of Nuke, and then I can get the test finished. We need to start progress on actual production soon, but this has definitely put me in good stead for actual production and I am feeling a great deal more confident about the tasks I will be undertaking next semester.

Week 31: Mastronauts- Rendering the Test

I've been having a few issues with the Shadow layers primarily. The past few days have been trial and error, but I think I am starting to understand what the issue is. I have a feeling it has something to do with the IBL set up because the issue lies within the render cam only, to which, I have the IBL plugged in.

Nevertheless, this is the rendering progress so far:

Creating the Render Layers

Some of the Render Layers:

Ground AO

Character AO

ID Matte

Master Layer

Obviously not all of the layers are done yet, but I am feeling much more confident about it now. I'm certainly finding that my participation in the rendering of the Going Live project has made this process a lot easier for me to understand.

Week 29: Mastronauts - projection continue

So I tried yet another tutorial and was getting similar effects...However, unlike being stretched vertically,it appeared to be the opposite. 

So I broke down the possible factors that could be contributing to this. I had gone through most of them in my head in the last tests but because this was slightly different from the others, I decided to go through them again.

It turned out that it was a simple error on my part (this time round) in that although the projection was coming from the projection camera, the render was viewing the projection from the perspective camera - despite the actual shot being from the projection cam...confusing. Nevertheless a very simple fix. Although I had checked this in the last tests, I had neglected to check it this time round.

Lambert texture again - affected by the light in the scene.

Surface shader - unaffected by lights in the scene.

Although I was naive in underestimating the task, it was not overly complicated (third time round). Nevertheless, what it has done is confirmed to me that this will be a rather complicated effect for the shot we will require. I'm unsure, with the complication of the shot, that we will manage to use this sort of method to do it successfully as there will be a moving camera, that moves through the scene. This means that the perspective of objects in the room will be changing, meaning a great deal of cameras and projections will need to be matched up to recreate the entire scene in this way.

The guys and I are meeting on Thursday though, so we will likely discuss this shot in more detail.

Week 28: Mastronauts - Projection tests

For the transition between the library and the exterior shot, Tom, Kieran and I think that we may need to recreate the library with projected photo-realistic textures and I have been looking into this in further detail (after having done a great deal of it during the Going Live project).

We discussed simply having simplistic shapes to represent elements in the room and just projecting textures.

I've found myself following a number of tutorials to try and get this right, and although I know most of the issues I'm encountering must have simple solutions, none of the tutorials seem to cover them.

The first tutorial required the use of a locator (camera rig) to try and move the cuboid into place and match it to the image-plane as closely as possible.

Then came the projection stage:

I started off using a basic lambert shader but then realised that the light was messing with the believability.

I then moved on to using surfaces shaders as these are unaffected by lights in the scene.

The only issue I've had is getting the projected texture to sit correctly on the object. I think this will require a little further investigation. 

Holiday Shmoliday: Going Live - Ink Reveal with Rendered Layers

So, After I managed to have a rendering malfunction on the scene and it had to be sent of for a re-render, I used some of the existing layers to try out the inky/water colour reveal on the (what should have been) final renders of the shot.

After making a couple of mattes for the reveal, I tested the scattered-reveal first.

This was a more of a sweeping-reveal and was the preferred version for the director.

Holiday Shmoliday: Going Live - 3D Watercolour Reveal Tests

Here are a couple tests of the watercolour reveal in 3D.

The first one is the entire image sort of appearing from blackness. I don't think this is what was intended but I kind of liked it. It was also mentioned that there would be seconds of blackness in the beginning anyway, so maybe this could help fit the reveal in...just a thought really. This one is a little fast, but the speed can easily be changed.

The second is more like what I feel was expected from the reveal. The paper/ground would be present and then the text, lines and drawings of the map would appear in the watercolour style...(the idea is that the 3D will pop up into the scene so ignore that these elements are revealed in this too).

I could combine the both of these tests so that the paper could reveal and then afterwards the ink drawings and text on top, if this is preferred. Again, this is the first test using the same matte to reveal the elements, so there's plenty room for me to test this out a little further and do a bit of tweaking. (:

Week 26: Going Live - A few more mountain textures

    Does whatever a Tiger Mountain does...


                                                            First test - horizontal lines.

                                     
                                         
                                                 Second test - vertical lines. There are more
                                                 lines in this test. I'm unsure if this is too much
                                                 or not.

        

        

I may try another vertical line test, more in the style of the first test with less lines. I also feel that perhaps more tests involving simpler and straighter lines may provide a bit more variety. If it calls for it, this is something I will look into further.

Week 24: Going Live - The forgotten tests

So I meant to throw this up here a while back. Long ago, around the time of the first age, before the 3D fires were alive and burning, I began a rough 2D fire comp test...it turns out it's supposed to be played super fast - LIKE LIGHTNING...but fire instead, so it's not really what we need anyway. Also, there's a few strange artifacts going on that need tweaking obviously. I was also going to look in to having the paper blackening before the fire as well, but I see no reason for it now; thankfully the 3D fire looks like the way forward!

So there super slight differences in the timing here. When it's too slow it doesn't look right, it's supposed to be a fast wipe sort of deal.

This one's slightly slower but it's not hugely obvious.



So here's a few stills to help you get the idea.


                               

Week 21: Once For A Whole Day - updated timing

Well it took me a while but I eventually got round to updating the timing on the last tests:

If I'm honest, although the timing is quicker I felt that, after looking at Sheng's example, the ground texture should appear at a slightly slower rate than the building. Also because of the speed increase of the movement of the entire thing, the spreading of the layers at the end of the shot is lost and does not match the spreading of the layers at the beginning. This will require a little bit of tweaking and moving of layers.

Week 18: Once For A Whole Day - Watercolour effect

After the first test, Sheng explained that this is similar to the style she wanted but she needed the effect to be more hazy in style. She gave me an example of how she wanted the effects to flow.

3D水墨动画（恬）

                                                                                       (Style example)

 She also told me that she needed the movement to progress from the right-hand side to the left-hand side of the screen (appearing with the character as she runs across the screen) and made me a quick example of what she meant.


                                                                      (Li, 2014)


With the character movement looking like this:



                                                                      (Li, 2014)

This was really useful and really helped me to understand more about how the style of the effects could really enhance the feeling that she was trying to create.

So I edited the existing test to try and match this particular style and movement.




This was much more along the lines of what Sheng needed although it is really rather slow. I will likely need to do another quick test with more accurate timing.

Week 18: Once For A Whole Day - Painterly Effect reveal

Alongside her team project, Sheng also plans to do a short personal project. She explained to me that she wanted her environments to appear on the screen in a painterly way and so I tried to do a test in After-Effects to create the style she was looking for.

Her environments have a watercolour feel to them and we discussed and agreed that the transition should mirror this style.

Week 16: Mastronauts - Compositing Test

The festivities may have ended a while back, but the testing continues...

A bit bored of tracking, and really wanting to try my hand at Nuke, I decided to try a compositing test. It was suggested to me that I should try this simple Digital Tutors project as it really touches on a lot of compositing elements I will most definitely encounter in the production of our short film.

Again, using the student version of nuke (hence the watermark), I followed the tutorial to get to grips with nodal-based compositing.

The tutorial touched on subjects such as:
Tracking, rotoscoping, colour-correction, roto-painting, lens flare and simulating depth of field for example and I was looking at these basic compositing processes through node graphs/trees. It is, once you start getting used to it, a very organic way of working....however I feel once the graphs/trees start getting quite busy and the compositing becomes more intense, that the process will be a lot more complicated.

Holiday Shmoliday: Track test 02

So as I fiddled with the shoe track, I decided I wanted to try and track a moving object in a scene, using the background shots from a Digital Tutors project, I tracked the scene and tried to insert a moving ball in the shot. As you can see, I'm not award-winning animator. 



I tried to have the ball look as though it was interacting with the table and then moving within the eye-line of the actors (somewhat). A very basic test, and after the first two with the shoe, I know that I will need to be more aware of slipping and sliding of the objects/characters when tracking.

Holiday Shmoliday: tracking tests

So, my fellow Mastronauts (Tom and Kieran) and I have been doing a few wee tests over the festive period to make ourselves a little more confident in the processes we will be undertaking in the production of our short film.

Using a student version of nuke I've been primarily focusing on tracking.
I used the footage from a Digital Tutors tutorial I was following and tracked the shoes I modelled and textured last semester into the shot.

The coloured noise on the top of the tests is the watermark for the student version of Nuke (as the uni was closed I had to download it onto my laptop)....which in itself doesn't make viewing it very easy, but it was good practice nonetheless.

It wasn't until I got the shot out of Matchmover and into Nuke that I noticed that shoes slide slightly. I found it rather tricky to spot small and gradual sliding in Matchmover.





So in this shot, to test the tracking in nuke itself, and to see if I could make a better track (being locked out of uni which prevented me from being able to re-track and re-render it). I took an image of the shoes and tracked it to the software using Nuke and found that the shoes finally stopped sliding. The track is still not perfect as it wobbles/shakes ever-so-slightly in the shot but this is still an improvement from the last shot.