TAXI & MEGA BUS
LONDON EYE
London Eye
Taxi & Mega Bus
Bring the London Eye WOW and OMG impact to the wider London publics. Peri has been leading on this fun 3D London Eye project, while working closely with the retoucher and artworker at elvis to produce a set of 10K high quality CG renders for them to create these beautiful design on the iconic London black cap, and mega bus.
Client: Merlin Entertainment, London Eye
Worked with: Elvis Communication
01
Investigation + Planning
02
Position + Texturing
03
Rendering
Before showing the final animation,
shall we talk about process and what's behind the scene?
H
MY
G SH
G
Investigation
& Resource Planning
When my work agency first received this project, I have been working with elvis head of studio and head of production to work out what was required for the project. This includes resources planning and processes from a motion and technical POV.
For a side note, my work agency has previously produced a set of London Eye 3D renders in the past, so there is an existing model. However the issue was that the original renders were in much lower resoultuion (around 2000x1000px), and it has a locked background behind the glass from the 3D render. The new ask is to produce a set of 10K+ renders, along with placing on a new 10K+ London background image. This requires not only upscaling, but making sure the wheel seat perfectly with the WOW/ OMG on a specific position on the new background image. Also there are various deliverables for different size of taxis and buses, which means having a specific background through the glass will cause many issues. Ideally is to produce both wheels (Landscape and Portrait) with a transparent background, so that the retoucher and designer are able to apply the 3D wheel on top of different part of the background image, without going back to 3D back and forth for all the deliverables. I believe this is the most time and cost effective way to execute this brief
Conclusion
After some digging and investigation within the C4D project file, I have found that there were only 3 existing pods that has material and people added to the scene. some of the other pods are not in correct position, also most of the wheel doesn't have any pods on it. The previous wheel were rendered in a much lower resolution. In order to apply the 3D London Eye on the new 10K image, we will need to zoom out from the camera to review more part of the wheel.
This also mean for a 10K resolution new render, we will require more pods to adapt both Landscape and Portrait scene, and need to update texture, and camera angle in order to see more of the wheel while keeping the wheel at the correction position + rotation to reflect the same WOW and OMG on top of the new 10K image.
The process will be:
1. Adjust camera position and angle to make sure the wheel fit the correct position and rotation on top of the new image. This can't be changed dramatically, as the main pod has to at the correct position to reflect the letter 'O' for both 'WOW' and 'OMG' next to the parliament of westminster on the new image.
2. Adding x3 extra pods for both Landscape and Portrait wheel. Make sure all the pods are in the correct angles.
3. Adding people in the new pods, and for client to sign off the new wheel position, and people.
4. Applying the correct texture, fixing Geometry, glass reflection, glass background (Ideally transparent so that it will be easier to adapt for multiple deliverables)
5. Optimise rendering settings, create AOV masks, set files up for retoucher to work on retouching the CG renders.
New position
Below is an example of the new 10K wheel and new 10K image.
(Non-textured)
Challenges
Work produced from a freelancer
As a back story, a freelance senior 3D designer has first picked up the project as I was on holiday at the time. The freelancer has worked on adding more pods for both landscape and portrait scene, and have added people into the pods. He has also pull out the camera to review more part of the wheel as requested, and added a fish eye filter to matches the London eye wheel position to where the wheel should be on the new background image.
The issue arises after this stage, as the new texture to the new wheel was very flat, some people texture were missing, the glass reflection was invisible, also with 10K resolution, we can start to see many problems on the wheel geometry itself. This includes the glass windows, doors, frames, floor, table. Also the lighting has changed that required new setup. When the freelancer handover his file, the wheel was looking very fake and very CG when applying to the new London background. The render time the freelancer claim was massive, it has taken the freelancer 3 days to render 2 wheels before giving any initial feedbacks. Which has given a knock on effect and a risk of delay for all the deliverables.
So... I was back and let's fix this.
Just a few screengrab of what were the issues. The issue not only around the texture and geometry, but overall the wheel just doesn't look appealing.
I came back from holiday, and have picked this project up to work on fixing the issues, reducing rendering time, and making sure the wheel looks as realistic as it possibly can be within the given time frame, which has been x2 days for landscape and x2 days for portrait.
Previous render
New render
Previous render
New render
Previous render
New render
Master Image
After fixing all the texture and geometry, I have then move on to adjusting the glass window with transparent background while keeping the glass reflection around the pods. This was quite challenging as the glass is actually reflecting the BG plate, even though I have hided the BG plate from render, the BG were showing through the pod and when I removed the BG, the reflection of the glass changed. The challenge here, is to keep the same reflection of the cityscape, white having alpha BG through the glass. This way, we can use the same CG render across various deliverables and different formats.
Below is a few slides to showcase how I manged to improve the previous CG renders by applying new texture, improving geometry, and created those pods with alpha background while keeping an adjustable glass reflection around the pod. Everything has become more procedural, and have given the retoucher and artworker full control to adjust and utilise the 3D model.
Renders
As mentioned above, I have rendered a set of the fully textured London EYE Wheel for both landscape and portrait, along with AOV masks and people masks for the retoucher and artworker to apply the wheel on multiple assets for deliverables. In which, I have also prepared x2 master photoshop file which contains separated layers including the frame itself, thinner glass pods, thicker glass pods, reflection mask (to adjust reflection strength), people, so that we have full control for any extra post-production work. This has worked perfectly for all the deliverables, and the resulting works are incredible.
Don't forget to also check out the London Eye Outernet project!
Which is using this 3D wheel to create an animation from London sunny day time transforming to a beautiful sun set.
Other projects you might like: