Park Road’s head of tecnhology Phil Oatley talks about how 3D is evolving and the intricacies of the stereoscopic post process.
Based in Wellington, New Zealand, Park Road Post provides post-production services for film features. The company worked on last year’s The Hobbit: An Unexpected Journey, with other credits including District 9 and The Last Samurai.
Phil Oatley has been the head of technology at Park Road Post Production since 2008. Originally hailing from the UK, Oatley began his career in post production systems and integration in Soho in 1997, before moving to New Zealand in 2003. He initially led Park Road’s technical operations in the Digital Intermediate department, before being promoted to head of technology, overseeing all aspects of technology within the Sound and Picture departments,
How has the way that 3D is used in films changed over the past couple of years?
PO: Contemporary stereoscopic acquisition has continued to evolve as filmmakers become even more comfortable with the visual language of the medium.
With the advent of relatively easy digital 3D exhibition, and a clear audience appetite, there were undoubtedly some gimmick-driven stereo projects utilising negative space for an ‘in-your-face’ experience much like the most excessive of the less technically accomplished movies of the 1950s.
It wasn’t natural 3D, and it wasn’t always comfortable. There was – probably naturally – something of a swing in the other direction, towards very conservative stereo, which not only diminished the stereo effect but also diminished the value of using 3D for the cinema audience.
Throughout this natural ‘settling in’ period, great stereoscopic 3D work has continued to be produced. A technical and visual language with a deep understanding of creative depth continued to be promoted by teams like the one at Park Road. This rich and deep approach satisfies the sensation of depth and excitement alongside technical excellence and guarantees absolute comfort for audiences. Native stereo shoots (as opposed to post-converted work) continue to excel both at the top-end of acquisition and here in finishing.
What are the main differences between a 2D postproduction workflow and a 3D workflow?
The biggest difference between traditional 2D work, and Park Road’s pipeline for 3D projects is during the preparatory stage where we undertake the technical stereoscopy, balancing the difference between left and right eyes, so that there is zero colour or geometric misalignment.
For those projects where we have provided stereo rushes and dailies we’re actually a step ahead of the traditional online, because we have already completed a very usable stereo colour balance and geometric alignment, and this is carried as metadata directly into the online.
Then, of course, is the final creative convergence pass, where we adjust the depth placement during and between shots for the narrative drive in parallel with the final grade decisions (colour timing and stereoscopy need to work hand-in-hand). Again, we are always striving to achieve the best stereo expression while maintaining comfort; we never want any kind of audience fatigue.
What have been the biggest improvements in stereoscopic post production tools over the past couple of years?
Deploying SGO’s Mistika DI system has been an amazing evolution in terms of the 3D toolset. It has made worrying about tough colour and geometric misalignments a thing of the past.
The ease and speed of their geometric toolset is astounding, but even more impressive is the optical flow colour balance tool: with this tool we have been able to fix even the most challenging of colour differences between left and right eyes. Without this tool we would certainly be back to doing more complex and time consuming compositing fixes to achieve the same result as the optical flow colour balance tool.
Beyond that, the systems we have built into our workflow here to track and carry forward stereo and grade information with each and every clip, regardless of how that clip ends up being used, is staggeringly efficient, and really frees up resource for creative 3D work – which can only be good for our filmmakers and their audiences!
What impact have those tools had on the role of the stereoscopic artist? Have they allowed for an increase in the ‘fix it in post’ mentality? Or have they allowed for more ambitious 3D shots?
The “fix it in Post” mentality will always be there, regardless of the quality of tools onset or in post. Certainly from our perspective we are never shy to tackle anything thrown at us, and smart experienced producers understand the impact of making that call as opposed to a re-shoot.
We have been able to offer a lot of advice even before any fixing is done, right through the shoot and in to the final VFX and post production. We have always tried to work with productions in this manner, and in the case of 3D productions that team-based approach has certainly given our directors, DoPs, stereographers and other partners more opportunity to achieve more ambitious 3D shots.
Which aspect of stereoscopic post is the most time consuming? Can you envisage any tools being developed to speed up or automate the process?
The technical colour and geometric alignment is always the most time consuming – and vital – part of the process. No matter how automated the toolset is, the human eye will always be your best tool. It’s a part of the process we take very very seriously, even during the shoot when we are able, because bad alignment means fatigue and discomfort – which is something we just don’t do.
What do productions need to take into account when capturing shots that will be 2D to 3D converted in post?
That all important on-going conversation with the production is still the core part of a great 2D to 3D conversion. Ultimately it’s working with the visual layering within the shot. The composition of the framing from the DoP becomes part of that conversation and, if the shot is VFX, how many elements can be supplied has a huge impact on the quality of the conversion.
It’s important that everyone understands the strengths and weaknesses of 3D and 3D acquisition and that a razor sharp eye is maintained on z-space placement of dimensualised objects in particular, that’s probably the most challenging thing.
What impact does working at 48fps have on the stereoscopic post process?
We’d have to say that there is no negative impact at all – in fact the higher temporal resolution really helps with our work on fast action and quick pans in particular – something we sometimes have to do some clever compensation for, such as optical flow effects, in the older 24fps world.
There’s obviously a data overhead, but all our machines live on a high performance SAN here and we have HFR screening in our suites, theatres and cinema, so it’s just a really cool new aesthetic for our team to get creative with.