On Euphoria.

by Steve Bowler on May 20, 2008 · 5 comments

in art,business,nitpicking

One!  Singular sensation!

I really, really want to get my hands on some kind of Euphoria demo, in order to either support or assuage my fears regarding the “better than rag doll” solution that Natural Motion has come out with, but I just can’t get over all of the tiny voices in the back of my head that says there’s still something wrong with it. When I try and approach people on the team in charge of making purchases or green lighting eval kits, this is what runs through my head when they ask me if we really need it:

  1. There’s a lot of things I don’t like about movement in GTA, and I wonder if it isn’t Euphoria’s fault.
  2. Every Euphoria demo pre GTA or Force Unleashed looked like pure ass.
  3. The original Lucas Arts product that Euphoria was to debut on was cancelled or shelved. No one talks about “that Indiana Jones game based on Euphoria” anymore. It’s not even listed on Euphoria’s website for games or upcoming games that use their software.
  4. Endorphin (their previous product) wasn’t very useful unless you were making a football game and needed tackles you couldn’t animate or motion capture easily.
  5. It’s not engine plug-n-play friendly. A programmer from Natural Motion supposedly needs to be embedded in your team for around six months.
  6. I can’t help but feel that they’re over-selling it in videos that show it off.

And speaking of their videos, their latest sales pitch on their website has a lot of us in the industry scratching our heads.

Now, I understand that according to their website, it’s very scale-able. You can tell it when you want it on or off, so you’re free to use your own hit reacts (thank god). But then in their video they show nothing but computer generated hit reacts and claim that they’re all “unique” and are somehow superior to traditionally animated ones or motion captured hit reacts.

This would be fine, if every single hit react they showed didn’t look just like the one they showed before it or a subtle variation thereof. These are screen grabs of the five hit reacts they showed:

First Euphoria hit react.Second Euphoria hit react.
Third Euphoria hit react.Fourth Euphoria hit react.
Fifth Euphoria hit react.

It becomes immediately apparent to any industry veteran that there’s pretty much only one hit react behavior being shown here, with a few variations on the movement. Maybe that’s what Natural Motion was going for here (showing off the subtle variety you can get from a single hit react), but from a potential consumer perspective, I now have to assume that this software only ships with one hit react behavior. One could assume that it doesn’t, but when the price tag for this software either approaches or exceeds a six figure mark, it isn’t safe to assume anything.

But more to the point, they’re showing this off as if it’s a high value, when we’ve been spending the past five years in the industry intentionally trying to avoid the very thing they’re selling in this hit react section. Subtle variations don’t focus test well. Most people think they’re seeing the same animation over and over again. If you want high value for your content, especially at the distances hit reacts are viewed in video games (far away and small on screen) what you really want is something that is immediately recognizable as different, so you want very broad and grossly different moves. Not subtlety.

The issue is exacerbated when they display two dummies side-by-side, again, presumably to show off how they look different. Only, it kinda backfires, and they wind up looking the same:

And a one and a two and a...

Continuing in this same fall and rise hit react we see they’re still mostly identical:

One...Singular sensation.

Again, this kind of “twinning” is something that we’ve been really trying to avoid in videogames for a long time. Only having a few animations in a bank to choose from means that we see a lot of situations where if five guys get hit, you almost always see the same animation play on two of them, and consumers aren’t a big fan of that (neither are those of us making the games, but you work with the hand your memory footprint deals you).

So I’m left wondering where the real value of Euphoria lies at, which without any kind of eval kit it is tough to stomach the price tag and everything that goes with it (developers from Natural Motion coding it into your engine since it’s not middle-ware). If all I want to use Euphoria for is a pose-driven rag-doll replacement and an IK foot placement solver (so if a guy stands on a curb his foot steps up), is it really worth it for the full integration? Additionally, I sometimes feel that all of the cool natural motion you see with Euphoria is wasted in a video-game environment, since the player doesn’t want control taken away from them (so stumbling is a big no-no unless you’re intentionally drunk).

It’s tough to find people who are using Euphoria who can talk about it, and I missed my demo time at GDC this year. Anyone out there use it who can talk about it? I’m not really looking for a sales pitch from Natural Motion (I could just as easily call them and get that), but someone who’s actually used it and can discuss if it’s a value-add or not a lot of bang-for-buck.


Comments on this entry are closed.

Previous post:

Next post: