Dan Macarin worked at Animal Logic and Rhythm & Hues before joining Weta Digital in 2008. He has worked on many films such as HAPPY FEET, THE GOLDEN COMPASS, AVATAR, DAWN OF THE PLANET OF THE APES, THE BFG and MAZE RUNNER: THE DEATH CURE.
What is your background?
I worked for 6 years in NY doing television and commercials before moving into feature films. After working in Sydney and LA for a few years I moved to New Zealand to work on AVATAR and have been working with Weta ever since.
How did you get involved on this show?
I had been working with VFX Supervisor Charlie Tait on the first DEADPOOL movie and then on to GUARDIANS OF THE GALAXY VOL 2. Charlie was already on AVENGERS: INFINITY WAR when DEADPOOL 2 came in so Weta asked that I take on the show.
What was your feeling to be in the Deadpool universe?
I was really excited to work on DEADPOOL. I’ve worked on a lot of comic book movies in the past, but rarely do you get to work on a character that speaks directly to the audience.
How was this collaboration with director David Leitch and VFX Supervisor Dan Glass?
Dan and David were extremely easy to work with. They knew what we did on the first film and really trusted that we could give them the performance they were after.
What was their expectations about the visual effects?
To keep true to the character. To make sure Deadpool didn’t read as being an animated character. Making sure the audience feels the emotion through the mask.
How did you organize the work with your VFX Producer?
Divide and conquer. We started by breaking down all the pieces that need to be done on Deadpool. Starting with facial, costume additions, costume continuity and then we move to the immediate environment around Deadpool. This included adding blood to weapons, the ground around him or on different characters. After that we move to the wider environment doing background replacements or additions, paintwork, etc. We had teams specializing in each part, working together to complete each shot.
What are the sequences made by Weta Digital?
Almost anytime you see Deadpool in a mask, outside of the James Bond opening credits and when he gets blown up, you are seeing our work in action.
How did you used the Weta experience on the first movie?
We learned a lot on the first movie as far as what expressions are more successful than others, but as there was a new director, producer, and client side VFX supervisor, we had to make sure those lessons still worked with the new team. Weta had a good setup from the first movie, but like most sequels, clients want twice as much in half the time. We had to make our system more efficient. Our Comp Supervisor, Florian Schroeder, redesigned the entire system. To start a shot we only required the plate, the camera and a matchmove of Deadpool’s head as inputs. “Koba”, our in house 2D assembly tool, would pull in these assets and handle the rig setup for the artist. This way we spent minimum time on shot assembly and maximum time on being creative. We removed the need to redo lighting in Maya as we had done for the first film and developed a better relighting pipeline in Nuke.
Can you explain in details about the creation of the facial rig?
We started out with the facial rig that we built for DEADPOOL 1. We had 32 blend shape poses with which we could create Ryan Reynolds’ facial expressions. Those blend shapes were converted from 3D models into unwrapped 2D UV layout space. Once in 2D space, we animated the face by combining and layering pose transformations. In addition to the shapes, we also unwrapped and mixed point normals so we could use them to guide light response based on changes in the mask topography and properly relight the face. This approach gave us a procedural approach to maintaining the set lighting, so we only had to correct for the lighting differences that a new expression would cause rather than introducing an artificial/additional light. Staying extremely true to the on-set lighting also meant we didn’t have to dial color or intensity – even the shadow density in his face was procedurally extracted. Once the animation was complete, we converted the UV and relighting maps so they fit back onto the plate photography. Only then did we perform the final warp and lighting adjustments on the plate to keep picture quality as high as possible.
Can you tell us more about the animation challenges?
When we are trying to match Ryan’s performance, it’s not just about matching his expressions. Some of the faces Ryan is able to make, you can’t believably make in a mask. We have to readjust the expressions to hit the same feeling, but keep it grounded. Speed is another issue. Ryan can move his brows quite quickly, but when we match that speed in the mask, it quickly looks cartoony. The movements had to be slowed down and given weight so that it felt like a person underneath the mask was making those movements.
Did you received specific indications and references for the face expressions?
We worked with Ryan so that he could send us clips of him doing the scene without the mask. This also allowed him to give us reference for alternative jokes and different takes the director might want to try out. We built a library of facial expressions from his reference so that if I told an artist to go from “concerned dad” into “angry bunny”, it would make sense to them.
Deadpool face is seen often in close-up. How does that affects your work?
Deadpool’s mask seems simple, but the small texture detail on the black sections and the checker pattern on the red makes it very important we avoid losing detail or stretching the patterns out so it looks warpy. When close up we need to recreate texture detail and and mask certain sections of the mask to prevent warping.
How did you handle the various lighting conditions?
One of the benefits of using the plate is that we don’t have to make a lot of adjustments to the lighting. We add lighting to help shape the eyes better and to read wrinkles in the mask. This is done with a relighting system we have created in Nuke.
What was the most complicated face expression to do and why?
Shock and surprise are the most complicated for us. One of the key features we always keep in mind is to make sure that Deadpool’s expressions are believable. If we push the mask too far it looks animated in a cartoony sense. That causes the audience to lose the character. The distance between being able to read the emotion and looking silly can at times be incredibly small.
How did you work with the other vendors?
For most shots we would work on the raw plate and supply our work back to other vendors as if the dialogue and facial movements were done on set. For other vendors our work became the raw plate for them. When shots that required retiming or had significant changes to the original scan, we were supplied with the vendor’s final comp and we worked directly on top of those.
Can you tell us more about your work beside Deadpool mask?
In several situations and for various reasons, Ryan was unable to wear his katanas. The sheath was long and so anytime he’s sitting or on the ground, or in a car we needed to digitally add the katanas and their sheath back in. This required matchmoving the body and tracking in 3d rendered assets. During several points of the film Deadpool’s costume ran into continuity issues. Things like bullet wounds, blood spots and costume tears needed to be added back in or removed. This work was done using image-based motion vector tracking and a lot of manual fine tuning. We also did several driving sequences where we integrated the car and characters into a variety of environments. Some sequences required us to add blood on the katanas, chainsaws, and fence posts.
How did you created the various environments for your driving sequences?
The backgrounds were filmed in real world environments. We had to take those backgrounds and align them to the cameras for each shot, keeping in mind the horizon lines and perspective differences. The exposure and levels needed to be adjusted to make sure that it matched what they would be if you actually shot the scene outside. Our background plates were limited so we had to remove and adjust each shot to make sure buildings, trees, and cars were not seen in repetition. We added smudges and dirt to the windows so that you got a separation between the car and the outside world. Putting the background plate into world space in Nuke allowed us to create reflections for car windows and mirrors.
Which sequence or shot was the most complicated to create and why?
At one point in the movie Deadpool gets ripped in half by the Juggernaut. A different studio was doing the shots and for technical reasons needed Deadpool to be all CG. Although the integration of the digi-double was done very well, it didn’t feel like Ryan in the suit. The client asked that we use as much of the original performance as we could while combining it with what the other studio had done and apply our facial work on top. At this point it becomes tricky as we need to work with the other studio to get all their layers and request renders be adjusted to help us with the integration. You don’t want to cause anybody extra work and you want the shot to run as if two studios are working as one studio. Luckily, this isn’t the first time this has come up and we have a great working relationship with the other studio, which made the process much easier.
Is there something specific that gives you some really short nights?
This show was setup incredibly well. We had a team that was able to adapt and react so well that it made for a very low stress show.
What is your favorite shot or sequence?
It would have to be a tie between two shots. The first is near the end of the film DP goes to hug Cable and he whispers, “the kids call this docking”. I’m pretty sure I watched it a dozen times and still laughed. The second is a shot on the plane when X-Force is getting ready to skydive and DP gives a speech about how proud he is of his team. The expressions we were able to add brought the scene to life and gave it a real emotional feeling.
What is your best memory on this show?
A few shots got turned over where Deadpool had been impaled in the head. There was no dialogue on the shots and in slow motion he is making some genuinely inappropriate gestures towards Colossus. We assigned the shots out and one of the artists on the shots came into the office and asked what I would like Deadpool to be doing exactly. The comp sup, the project manager, and myself were trying to act as professionally as we could as we discussed what we can do until one of us went into demonstrating an expression. At that point the project manager left the room and as we continued down this path I remember I tried acting out the shots and about mid-way through I saw the artist’s face and the expression was priceless. I don’t think I’ll forget it.
How long have you worked on this show?
What’s the VFX shots count?
507 shots total, 56 dvd shots, 119 marketing shots including commercials, a music video, behind the scenes music video, trailers, and movie promos.
What was the size of your team?
around 60 people.
What is your next project?
Yet to be announced.
A big thanks for your time.
// WANT TO KNOW MORE?
Weta Digital: Official website of Weta Digital.
© Vincent Frei – The Art of VFX – 2018