animation world magazine article

8/18/2019 Animation World Magazine article

1/7

February 1998 34ANIMATION WORLD MAGAZINE

Motion-capture is such a newform of animation thateven the name for this ani-

mation techn ique is still in flux.Medialab has constant ly been onthe cutting edge of developing per-formance animation technology,announcing ma jor advances to itsproprietary software at regular inter-vals. Since the company was found-ed in 1989 in Paris, they have cre-ated some 30 characters includingNickelodeon U.K.’s Bert the Fish andElvis, Pepe the Cricket from SteveBarron ’s T h e Ad v e ntu r e s o f P i no cc h i o and Cleo, who appearson Canal+’s C y b e r f l a s h , a showabout cyber-culture. Medialab spe-cializes in computer puppetry,which is a subset of the motion-cap-ture by computer field. Computerpuppetry differs from motion-cap-ture in that the results of human

body motion are fully rendered inreal-time, as the mot ion is per-formed. Therefore, animation direc-tors and performers can see the per-formance instantaneously and canthen apply immediate corrections ifneeded. Medialab creates comput-er puppetry by combining this real-time capability with sophisticateddevices to track not only humanbody motion but also facial expres-

sions and lip synchronization.One may have seen the

workings of motion-capture before:a computer generated character ismoved by an actor in a suit, who isconnected to a renderer which inturn moves the CG character.However, we are going to take you

through theprocess of creat-ing a believable,computer gener-ated character bygo ing behindthe scenes atMedialab StudioLA in LosA n g e l e s ,California. We arego ing to meetDrew Massey, ap e r f o r m e r ,Marcus Clarke,who has specifi-cally trained people to work in themotion-capture industry and one ofMedialab’s foremost technical devel-opers, Herve Tardif.

TheActor’s RoleNaturally a ma jor factor of

performance animation is the per-former. However, I found the typicalbackground of such a performer tobe a surprise. One such example ispuppeteer and performance ani-

mator, Drew Massey. Massey recent-ly performance animated “Broz” forVDTV at the Western Cable Showand interacted live with audiences.

Heather Kenyon: Your bio saysyou’re both a puppeteer and a per-formance animator. Can you explain

the differences between these twodifferent professions?

Drew Massey: Actually, w ith thetechnology at Medialab there’s nota lot of d ifference. It’s pretty muchperforming with real-time comput-er generated puppets.

HK: ‘Performance animator’: doesthat term apply to both the pup-peteer working on the body andthe face of a character?

DM: It’s all the same thing. I do a lotof traditional hand puppetry as well,muppet style. With that you’re con-trolling the head and the body of acharacter. No matter which way

you split it up, hopefully, you areblending the performances of bothpeople and making one believablecharacter.

HK: What is your background?

DM: Standard puppetry. I’ve been

by Heather Kenyon

© Medialab

Motion-capture is a great out-

let for a traditional puppeteer. -Drew Massey


2/7


3/7


involved in several movies, like T h e F li nt s ton e s and M e n i n B lac k . Thosemovies involved puppeteering w itheither cable-operated charactersand sometimes some traditionalhand puppetry, as well as a lot ofelectronic and servo animated char-acters.

HK: So you come from a puppetrybackground rather than an actingbackground? I am surprised youaren’t an actor.

DM: As soon as I started getting intopuppeteering, I started taking act-ing classes. I took acting in collegeand I’m an illustrator too. Theprocess of making art move hasalways been very attractive to me.

It’s all about the character, so if youdon’t have any sort of acting back-ground, it’s a lot more difficult tomake a believable character. Thefact that I’m making all of the move-ments and voice choices doesn’t getin the way. I started out as a pup-peteer but I’ve become a much bet-ter actor because of it.

HK: Why did you get involved inmotion-capture?

DM: Because its cool! I really likecomputer animation. There’s almostnothing more satisfying than see-ing computer animation respond toyour every move. It’s just a blast.When it works well, it works reallywell. Motion-capture is a great out-let for a traditional puppeteer.

HK: How much do you work forMedialab?

DM: I go in at least once a week,sometimes twice. It depends. Mostly,I experiment with the system andfigure out what I can do with it. Itturns out to be qu ite a lot. Really I

just get my own skills down to the

point where creating a believablecharacter is almost effortless.

HK: Is the demand for your servicesgrow ing?

DM: I think it is grow ing. I know alot of studios who are just doingmotion-capture. Medialab is the only

company, however, that’s really con-cerned about getting whole char-acters together and hiring pup-

peteers and actors to do it. That wasthe th ing that attracted me toMedialab in particular. They are soperformance-oriented. It seems tome a lot of other companies are hir-ing mainly mimes or people whoare specifically dancers and captur-ing their mot ion for a particularthing. Medialab is really concernedw ith bring ing the who le th ingtogether.

HK: When you’re acting and talkingto something that isn’t there. What

are your biggest challenges to makethat look real?

DM: That’s interesting. That’s some-thing you have to get used to as apuppeteer, the different portions ofthe body. When I see people thatare really into dance, and really con-cerned w ith their body, it takes

them a longer time to get used toit because they are not familiar withbeing outside their own bodies.

Typically their bod ies are the finalmedium. Every time I’m on a job I’mlooking at [the] monitors, the cam-era’s view of the puppet, and play-ing to that, so it’s not that strange forme.

HK: How do you approach playingdifferent characters?

DM: Like any acting job. I like itbecause the characters are so phys-ically different. It’s easier to get intotheir specific behaviors. It’s easier to

Actor Paul Pistore gives the voice and facial expressions to the character “Broz.” © Medialab


4/7


5/7


HK: Do you see the demand [formotion-capture] grow ing?

MC: I hope so. I love working onthe Medialab system. Sometimeswhen a new technology comes init doesn’t have a direct applicationbecause there are certain conven-

tions already set up. Animators havesaid motion-capture isn’t very use-ful. You have people who’ve hadbad experiences with the earlydevelopment of the technology. Ittakes awh ile for people to say, ‘Thisis a useful tool. This is better.’ Whenyou have a new tool, there’s oftena little lag before it comes into com-mon acceptance. I think that’s what’shappening now.

The Technical ProcessHerve Tardif is one of the

code writers based in Paris whoseknowledge of Medialab’s proprietaryClovis system (the engine that dri-ves the real-time factor) is practical-ly unparalleled. Clovis was firstdeveloped in 1989. Tardif is nowgoing to take us through the tech-nical side of making a charactermove and, more important ly, actbelievably.

Herve Tardif: One notion that is veryimportant is the idea of skeletons.We are going to have one real per-son wear a number of sensors.These sensors measure the positionand the orientation of the segmenton which they are attached. We areworking with electro-magnetic tech-no logy, which consists of onesource emitting a field and a receiv-er measuring that field. After someprocessing, it gives information onthe position and orientation. Withthis information we are able to builda skeleton that is going to be exact-ly or very close to the skeleton ofthe real person. We will have a copy

of the real person. That copyamounts to building a skeleton ofthe real actor, and attaching the dif-ferent values received by the sen-sors to the proper segment of theskeleton we just built. You can imag-ine conceptually at that stage, wehave a skeleton that moves exactlythe same way the person moves.That kind of information is alreadyuseful for our work. For people whoare interested in the motion acqui-sition business, that is pretty muchwhat they expect: a skeleton alongwith the orientation of each of thesegments of the skeleton.

Another application, whichis our most common application, isindeed to drive virtual characters.At that stage we have a skeleton,wh ich is a copy of the real person,and another skeleton of the built,or virtual character. These skeletons

may be very differ-ent because wemay want to ani-mate a gorilla, or avery th in woman,or a very big and fatcharacter. There aremany chances thatthe skeleton of the

virtual character willdiffer greatly fromthe real actor. Atthat stage, what wedo is a mapping ofone skeleton to theother. This gets verytricky and it’s wherewe have a lot ofproprietary infor-mation. This is a

ma jor issue for allpeople involved inmotion acquisition[motion-capture].When the propor-tions are qu ite thesame, it’s just astraight adaptation.It’s easy. But when

the proportions are different, it canget pretty tricky.

Usually there are a couple ofth ings we need to insure. Thesethings are usually the location of thefeet on the ground. We do mathe-matics to insure that our virtual char-acter will always have his feet onthe ground. Starting from there wecan go up the hierarchy of the skele-ton, and take the values from thereal actor and place them on ourvirtual skeleton. Once we’ve donethat, when the real character

moves, the virtual character moves.The more different the virtual char-acter is from the real actor, the moredifferent the motion is going to be.Suppose the script says that thecharacter should scratch his head.With the two skeletons being dif-ferent, it is very un likely that whenthe real actor scratches his [own]

Actress Lydee Walsh wearing the Medialab bodysuit.The sen-sors on the suit are picked up by an electromagnetic field

being emitted from underneath the stage she is standing on. © Medialab.


6/7


7/7


ma jority of our software is propri-etary.

HK: What functions does this pro-prietary software allow you to dothat is unique to Medialab?

HT: It’s more of a whole package,an overall level. I have seen com-

panies that do a very good job interms of rendering characters or

mot ion acqu isition, but I haven’tseen companies that can really ani-mate a character the way we do inreal-time. Our software has beendesigned to do real-time animationand to be used to produce TV ani-mated series.

We are now capable ofdo ing real-time characters w ithshadows. We are also able to use

our system on a real live set. Whenit comes to compositing computergraph ics w ith real live shooting,there is one notion that is veryimportant and that is coherencybetween the real world and the vir-tual. For instance, when you havea real character talking to a virtualone, you need to pay a lot of atten-tion to your cameras. We came upwith a way of calibrating the virtu-

al camera w ith the real camera. It’sa very simple process that allows usto integrate, in a very believablemanner, the virtual character witha real environment. This processworks with fixed cameras that donot move. One very big improve-ment would be to allow for cameramotion. This gets into the field ofvirtual stud ios. Recently, we havebeen working w ith a companycalled Symahvision. They offer a sys-

tem that can track a real camera,shooting a real live scene, and thenprovide us with camera positions tomatch our virtual camera position.With this system we should be ableto integrate virtual characters witha live set. This is going to increasethe credibility of the compositing.

It’s one thing to see avirtual character talk-ing with a real person,but having these twocharacters filmed witha moving camera isreally something else.It adds a lot. It is a verylarge technical difficul-

ty. We are go ing touse that system inproduction verysoon. We are in anextensive test period. We are tryingto use the system on a show whichis being produced for a TV channelin France.

HK: When do you think we canexpect to see this?

HT: March. Recently, we switchedto a wireless system. We used to usea wired system where the actor waslinked to an electronic unit, with 16cables which really restricts themotion. That was one of the ma jordrawbacks. Now we are workingwith a company called AscensionTechnolog ies. We’ve been usingthat w ireless system for over a yearnow and it’s giving some pretty

good results. We can now have anactor walking on a stage withouthim or her being linked to anywires. Before we couldn’t roll on theground or turn around many times.Now we can do all of this very well.We even have someone do inggymnastics like backflips right nowat the studio.

HK: Where is the future of this tech-nology?

HT: Our goal is to come up with asystem that TV channels could useor even direct live television. We areworking on the camera issuebecause we know they w ill wantthat. We w ill also probably see sev-eral characters. Right now we have

one character when we shoot.When we record, we do one char-acter at a time. In the future we willhave multiple characters interact-ing. There is a huge number of dif-ficulties to get to that, combiningthe two worlds is difficult and weneed to be very precise.

HK: How far away do you think thatis?

HT: We already did that on someshows. The level of interaction islow, because it is difficult, but Ibelieve that very soon, probably thisyear, you shall see some virtual char-acters interacting. We’ve been inthe field for six years now and it’s

getting to a point where people atsome TV channels in the U.S. areready to go for it. It’s already beingused in Europe by Canal +, TF1, FR2and Nickelodeon UK.

Note: The online version of this arti-cle contains two Quicktime moviesof Medialab’s “virtual characters” inaction.http://www.awn.com/mag/ issue2.11/2.11pages /2.11kenyonmedial-

ab.html

The computer, running Medialab’s proprietary Clovis soft-ware, combines the information coming from the two

actors. © Medialab.

H ea th e r K e n y on i s E d i to r - i n - C h ie f o f An i m a t i on W o r l d M a g a z i n e .

Note: Readers may contact anyAn i m a t i on W o r l d M a g a z i n e con-tributor by sending an email to

[email protected].

animation world magazine article

Documents