One of the key visual effects components of Alex Kurtzman’s The Mummy was the Mummy herself – Princess Ahmanet – seen in 5 various stages as she re-generates from her entombed and emaciated state into a more human-esque appearance.
Some shots of the creature were fully digital creations completed by ILM, others were meticulous CG augmentations of actress Sofia Boutella done by MPC, and some were Boutella in make-up and prosthetics.
In the first of two stories about the creature work, overall visual effects supervisor Erik Nash tells vfxblog about the five stages. In a follow-up piece, we talk to Nash’s fellow MPC visual effects supervisor Greg Butler to dive in on the specifics of augmenting Boutella.
Stages 1 and 2 – ‘A bag of bones’
Erik Nash: The first two phases are when Princess Ahmanet’s first free from her sarcophagus under the pier and then in the alley with the rats. At that stage, she is completely emaciated. Under the pier, her body is very disjointed, with joints out of place and she’s basically a bag of bones. And because she was so emaciated and had to have range of motion that no human could, those first two stages are the only stages where she was completely digital throughout. Those two sequences were done by ILM.
The way that we arrived at the model for that was, we scanned Sophia. She’s probably been scanned more than any actress in the past couple of years! I think at one point we had scanned her over 40 times because the wardrobe kept changing and every time they came up with a new version we would have to do a full body scan of her wearing it. We started with a scan of Sophia, and then the asset team at MPC put a skeleton inside that scan of her and then shrunk all her soft tissue down on to that skeleton that was fit inside her. It was rigged, and it was rigged with sort of inhuman range of motion capabilities because it needed to bend backwards and do all kinds of unnatural things.
We knew we wanted a lighting maquette, so once they had modelled and rigged this stage one version of the Mummy we had them pose it to match one of the concept illustrations that Crash McCreery had done. And it was a pose where she is on all fours but with a 180 degree twist in her spine so that her pelvis is 180 degrees out of alignment from her shoulders, but she’s on her knees and it was a really cool illustration and they posed the CG model to match. Which we then exported and gave to the prop department at full scale – actually props and the make-up effects team worked together to basically make a maquette from this digital asset that had been posed and she was dubbed Amy. I can’t remember why, ‘Amy.’ I guess it’s somewhat related to Ahmanet and we took her everywhere. Whenever we were doing a shot that ultimately would involve this CG Ahmanet, we would trot that out as part of our reference.
The differences between stage one and stage two are pretty subtle, you see them sort of take place under the pier as she adds body mass, she gets her joints back in alignment and gets to the point where she can stand upright and move like a human. But she is still emaciated enough that in ‘rat alley’, where she is in stage two, she still had to be completely digital. She had all these loose bandages hanging off of her.
The animation process for this stage 1/2 was always a three stage process where we would animate just the body, get approval on that and then tech anim would come in and do their thing and then the last part would be to do a cloth sim for all of the loose bandages that have to react to all of her movements. So, it was the first sequence turned over and that was back in October and the last shot of that sequence didn’t final until the last week of the post.
While it was keyframe animated, we did have a contortionist, who did some reference that we used for under the pier. We’d block the shot with that contortionist and ILM did refer to that footage in some regards but the animation was ultimately all keyframe.
Stage 3 – More life
The stage 2 to stage 3 transformation happened at the Abbey. There’s a shot where she sucks the life out of another victim and then we see her on screen transitioning from stage 2 to stage 3. That was done by MPC and we shot Sophia doing all the action, even though the original intention was she would be CG and then hand off to augmented live action. Ultimately, when we did the shot, she’s all digital the whole time. There was ultimately no benefit to have her transitioning to Sophia, so even though she was in the plate we wound up painting her out and never bringing her back in that shot.
But from that shot forward, other than a few instances where we used a digi-double because of the action required, everything else was Sophia in costume and makeup where we would augment and that augmentation was big hunks of missing flesh on both cheeks, a chunk of her nose was missing, and her shoulders were replaced completely with an exposed tendon and bone. Her hands, when her hands are featured, are completely CG because Alex wanted them to be very boney and emaciated. For example, the scene where she has Nick on the altar in the church and is inspecting him, all of the hand and arms you see there in that close-up are all CG, with with bandages and all of the stuff that goes along with that.
The beauty of having at least the real actor in the plate is two-fold. You’ve got the performance from the actor that is retained and at least as important, if not more so, you are committed to something. As soon as you go fully digital, now, everybody knows, not least of which is the director, that they can change anything at any time, whereas when your process is committed to working on top of the photography, it’s a different feel. The take that you chose, the performance that is there – all of those decisions now are locked in and you can focus on the hard work of making the visual effects work with that.
And I much prefer that, because it’s inherently photographic and it takes all those variables out of the equation, which allows you to really focus and concentrate on the visual effects themselves, and you’re not chasing some elusive performance that may or may not be achievable.
Back from my motion control days, I remember coming to the realisation somewhere along the line that there’s something inherently different about framing a shot using key frames, which is a static position which you are assigning to a point in time. As opposed to a live action camera move where it’s happening in real time. The camera operator is reacting to what’s happening in the view finder – you’re not making a beautiful frame on frame one and frame 100 and then letting the computer spline in between them and it gives you a very different feeling for the camera motion.
The audience may not know why it feels different, but I think most of the audience does recognise that it does feel different. And, to me that’s one thing that right off the bat is going to work against visual effects not appearing as visual effects.
Stages 4 and 5 – It’s all in the eyes
Stages 3, 4 and 5 are essentially all the same approach – it’s basically just less and less augmentation as we go from 3, to 4, to 5. In stage 4, which is how she appears after she escapes from Prodigium, and she’s on the London streets, is a little bit of augmentation. I think there was a little bit of facial work and again, a toned down modification of her shoulders.
In stage 5, all we were doing was replacing Sophia’s eyes because the one constant through all phases is the split pupils. They are the signature Mummy look that we established. So by the time we get into the third act where they are in the Crusader chamber and Nick and Ahmanet have their battle, that was Sophia as photographed, with the exception of her split pupil eyes.
The eyes required a precision with the tracking required. But also, with the split pupils there were a lot of issues we had with eye lines being correct or appropriate and avoiding her looking like she’s cross eyed or wall-eyed. So there was a fair amount of tweaking, soft of the alignment of her eyes when we did the split pupil work in all the stages.