Proposal | Version 02

A Change in Research Interests
June 26th, 2018

filmic blender, unreal engine 4, eevee, realtime, pbr, colour management

Following the submission of my Proposal document a few weeks ago, a number of things have changed or come to light that have motivated my reasons for changing the direction of my research aims slightly.

Chief among these are: the scaling back of the breadth of the project so as to focus on specific details and production values; and, to move away from realtime technologies to focus on Scene Referred/Linear pipelines due to the scarcity of information; instability of support and relative accessibility of technology regarding Real Time.

  • Validity of Research 80%
  • Complexity of Content 80%
  • BENEFIT OF ACTIVITY 100%
Above is an image taken from the ‘AR Wall’ project

Secondary Research

REAL TIME VFX INSIDE UNREAL AND BLENDER?

Firstly, there is no doubt that the introduction of software-based real time VFX solutions inside Unreal Engine 4 and Blender (eeVee) is exciting and could have a noticeable impact on the industry.

Unreal has shown a number of case studies and examples recently that hint at the power of this technology and its application with a filmic rather than game environment. Of particular interest have been ‘The Race’ – a partnership with Unreal and VFX company ‘The Mill’ and ‘The AR Wall’, which uses the Unreal ENgine alongside a real time camera feed to back project background CG plates in real time.

This said, many of the illustrative examples, are still using technology and or software outside of the free tools to complete the workflow. For example, Unreal currently has no option for matchmove and tracking internally within the software, relying on a very specific FBX keyframe export/import method of camera, tracking, solve and reconstruction data being supplied from either 3D Equalizer or BouJou, both of which are very expensive and in turn would bely the core principles of what I am trying to achieve with The Arcane Welder project.

Secondly, Blender is also in the process of developing a software-based, real time solution called eeVee. Again, the eeVee Physically-Based Rendering engine has produced a number of entirely CG generated case studies that offer an insight into realtime raytracing approximations for the future. This said, this functionality remains only in BETA form. To this end, it is highly unstable, but more importantly does not currently communicate accurately with the internal colour management within Blender and the all of the composting nodes. To this end, it would be nearly impossible to build an effective composting pipeline under these conditions.

To this end, I have returned to an earlier idea – and more importantly skills gap – that I believe is the final step to completing a fully professional, freely accessible VFX pipeline inside of free software: Scene Referred/Linear Colour Management.

Core findings

UNREAL IS POWERFUL, BUT RELIES ON THIRD PARTY, EXPENSIVE TOOLS TO PROVIDE THE FUNCTIONS I NEED

Even Though the innovations inside Unreal Engine 4 are striking and powerful, to have full matchmove functionality, it relies on third party matchmove programmes, with specific export options for FBX keyframes. Blender can not offer this functionality, and thus the high cost of these tools mitigates the overall spirit of the Arcane Welder project.

EEVEE PBR VIEWPORT RENDERING IN BLENDER IS INCOMPLETE AND UNSTABLE

Unfortunately, the realtime PBR rendering functionality in the BETA Blender 2.8 build is incomplete,with limited functionality for CG and Live Action integration. With little or no colour management options available and with the compositing nodes not talking to EEVEE at all when live action is included then this again, is not an option at this time.

SCENE REFERRED COLOUR MANAGEMENT IS THE ANSWER

To this end, I am going to move back to standard Ray Tracing rendering pipelines inside Blender, but aim to add the last core skill that I have yet to cover: Scene Referred/Linear colour management; the largest mitigating factor of Blender being classed as a Professional option for VFX. With the new addition of Troy Sobotka’s ‘Filmic Blender’ inside 2.79, this is finally an option that can be explored.

PLANNING

A REALLOCATION OF TIME, SCALE AND FOCUS  

Another factor that I had not originally considered at the point of submission of my proposal, was time with my family and workload from my full-time employment and freelance film work. Somewhat naively I was excited to explore the technology and extend the scale of the outputs, but this was unrealistic.

The commitments of my full time work continue well into June and July, with additional duties as external examiner for another institution also impacting on my ability to achieve the schedule and scope of my initial proposal.

Also, my family and I have seen little of one another and to this end have allocated two short holidays in the summer period, where family time will take priority. This again, was not factored into the initial schedule.

To this end, my initial schedule and thus size of my project will be amended to reflect not only the pressures on my time, but the need to focus on the core focus of The Arcane Welder project – VFX core skills and dissemination to young pele. Lastly, my shift to exploring a scene referred colour inside of blender, is going to require a great deal of research, practice and experiments to make it function correctly. Something which I will gain a great deal from, but is unknown territory for me and many others.

The Curious Engine

My Professional Portfolio
Go There Now

Recent Posts

Here are some recent posts that I have written as part of the Digital Media Practice 3 project – ‘FUSE’