After and exciting All-Star weekend, let's kick things off with a red post collection!
Was Quinn Rushed?
Ironstylus started off by leaving a reminder in a post discussing whether or not Quinn's model seemed to be rushed and not as intricate as, for example, Vi.
"I'll get back to this in detail, but yes. She was rushed.
This is honestly a bigger conversation which I'd like to share, but it's a bit too late for me to get into it now.
Hopefully tomorrow."He quickly elaborated on his previous comment, saying it wasn't a quality issue but instead a pipeline issue.
"Quick note: Personally this isn't a quality issue to me, this is a pipeline issue. That said, you will notice some stylistic differences because these models represent 3 different modelers and 3 different texture artists. There's going to be some discrepancy between them.Further elaborating on that,
"Looks bad" is a relative and fairly moot argument that is far too dependent on preference, so I'm not really going to touch that too much, but I can go into more detail about our challenges with Quinn in the ways that really mattered and how I would address them today"
"I feel it is quality. As I said, this isn't an issue of quality for me, as it's mostly an issue of pipeline. I'm fine with the quality of her models, but there's a lot I'd have done differently in hindsight."
IronStylus then proceeded to drop a knowledge bomb about Quinn's design and some problems they ran into, and changes they ended up making.
"So the deal with Quinn..
From the beginning we had some issues. I jumped on the project when we were ideating on the falconer trope. She was just called "Falconer" at the time. I had returned to a lot of skin production work post-Diana and work on Falconer involved me jumping teams for a little while. I had learned a lot from Diana. A whole lot. I had a lot of aspects of Diana's ideation and production logged in notes in order to help project some of the process onto any subsequent champions I worked on.
Rewinding a bit, Diana is viewed to be one of our more successful champions in terms of character development. She is a very well rounded champion thematically which is portrayed in all aspects of her "triangle", creative, design, and art. She also represents one of our more successful pipeline experiences in that due to the front-loaded work on creative and art, there was no iteration time needed post-approval/greenlight. The day she was approved she went into production. Production wise, she was mostly without headaches, she was straight forward, and a lot of the hard stuff was done at the beginning. At least the stuff that would cause headaches down the line in production, iteration in the 3D stage, or other tweaks that would require a lot of manpower, were avoided. So, successful in many regards.
Quinn suffered from the opposite. She stumbled in many ways at many points in the ideation phase. This is due to a lot of reasons, and I'm very much responsible as anyone. At first, in terms of art, she was actually a melee character. Two handed sword assassin. We sort of took a Batman approach to creative, that she was the secret superhero of Demacia, "The hero Demacia needed but not the one it deserved", etc. Gameplay wise, it was a lot of mark the target, follow-up.
These parameters changed a lot, obviously. We went in pre-approval with a very different direction post-approval. We didn't have as clear goals as were needed to make the production and further character development smoother. Mind you, this is not always something that's easy to track. With a lot of people working collaboratively, the noise in the room can be loud, and we can have difficulty in distilling a character's goals. Morello is exceptionally good at establishing goals for characters, so he and I have shared that we feel Quinn could have been a bit more cohesive in retrospect.
Additionally, unlike "Team Diana", where myself, Volty, Runaan, and Harrow were very much the owners of that champion from the beginning, our little group did not coalesce until later, after Quinn began production. At the time though, we weren't really wise to the fact that some of our more hazy goals from the beginning would cause some pain later on.
Her story shifted, we hammered a lot out as production continued. I think it went in a very good direction actually. I'm quite fond of Quinn the character, in that I think she's compelling and unique. Design progressed, and Volty had some really interesting ideas about her. We felt confident in that. Art was tricky. I was having some issues in making her unique. I had finished off production art of her and Valor, handed it off to 3D, and was happy, but I didn't quite realize that a problem would arise.
She ended up suffering from the exact thing Diana did not suffer from, iteration in 3D.
When we got her in game, when our model was fleshed out and animation was rocking, there were concerns that she wasn't quite different enough. I agreed. I had already moved onto relaunch so I needed to hop back onto Quinn, work with the 3D artist, and do some model tweaks. Before doing this, I met with a few artists and brainstormed what we could do. We basically went back in and did something we didn't quite hit on initially, we created a solid goal. She was a quintessential ranger when we distilled things.
Now, yeah, duh, that’s what Quinn is, a ranger. We’ve said that before. However, oddly enough, it didn’t quite dawn on us how clear this made certain art choices. I took a look at what we had, and I really cut into her. We changed a lot. We introduced some elements into her design, we created some bigger reads, etc. We did succeed in making Demacia’s version of a Ranger, but it wasn’t as smooth of a process as I would have preferred. Myself and the 3D modeller assigned to her did a fair amount of extra work. In the end, we came up with what I felt was quality, but I realized that if I had some simple questions answered up front, the process would have been a lot easier.
All that said, there was something funny that Diana and Quinn shared. Truncated timelines. Diana was pushed up in our slotting due to some pipeline changes. Quinn had the same thing happen to her. The difference is that we had the majority of the goals for Diana solved up-front, but Quinn didn’t quite have as solid of a foundation. When that foundation was poked at, we had to approach things a different way, and iterate at a stage that’s much more difficult to iterate during. Since all of Diana’s thematics were front-loaded, that early work compensated for her being pushed up in the release schedule. With Quinn, we actually had to push her back a bit to hit the notes we wanted to hit.
In the end, I do really like Quinn. I do think she’s unique, I do think she’s something different in the scope of fantasy rangers. I like the place she occupies, and I think she’s a very well fleshed out character. I do however feel that we could have more smoothly figured out her goals and made the correlation between what they were, and how they would directly impact art and other aspects.
I think she hits our quality standards really well also. If you look at the sculpt, texture, animations, etc, they’re very well done. They do show differences in respect to other characters because they were made by different artists. Additionally, Vi, the example shown here, was in development much longer and was given the benefit of more marination time. That’s not to say Quinn came out half-baked, it means that stylistically, we could have brought them more in line. Even currently we’re working on reconciling how our characters fit with each other, and with the game environment.
That’s a lot of style guide work which myself and other seniors are involved in. That work is directly applied in Relaunch, where Grumpy Monkey and I work to get our assets updated, or in-line with each other stylistically. We’re working on the principles and the practice at the same time. It’s fascinating and fun, but we do learn all the time. It can make the target move a bit here and there.
So, while yes, there were hiccups in the pipeline, we made choices in places where it was more difficult to make iterations in. I feel we hit quality, but given the choice, I would reapproach Quinn differently, with our goals better established. I would take what we know now, about goal creation and adherence to those goals, about small group investment in a character early on, about stylistic consistency and visual design across champions.
This is the more interesting and sometimes difficult part of the job, and the challenge of being a developer. These are all learning experiences. We are lucky to be in a position to learn from the processes we engage in and change how we approach things in the future. We constantly learn so that we can produce something better. Yet, we’re human. We’re going to trip up and misalign at times, so I’d be an idiot to say “EVERYTHING WE DO IS PERFECT AND WE MAKE A TEN OUT OF TEN EVERY TIME.”. We believe in what we produce, we wouldn’t put it out if we didn’t feel that way, but because we’re iterative, we have the blessing and the curse of learning over and over. We know that circumstances will change each time, goals and processes will shift. Character development can be really tricky, and sometimes you can think everything is solid and rock-hard, only to find that when you look back at what’s been created, you see a slew of other options, solutions and opportunities you didn’t notice the first time around.
Hope that helps clarify. Sorry for the essay."
In response to this, one summoner asked "where do you feel like the biggest choke points are that hold back your overall quality control?", to which IronStylus replied:
Not so much quality control, but being on the lookout for emergent issues in a character's development or physical production. We need to front-load more. We need to have small groups do big things, which is when we get our best results. When a champion idea materializes, people rally to it. That excitement causes use to really want to churn on it. This creates lots of discussion, discussion creates goals. Mind you, this is happening, this does organically come into fruition, but it's not exactly a science.
I think sometimes we don't establish those goals up front intentionally. We get a lot of awesome content and amazing characters, but sometimes those things more or less fall together. It's hard to standardize a process like that. Draven was just.. Draven.. from the beginning. Boom. How do you make that into a processes that's trackable?
The closest thing I can think of is what we did with Diana. I'm going back and again looking at what I learned from that. Morello and I have talked about it a lot, and it's something I'm keeping very much in mind as I work on the next champion I'm involved with.
There isn't one singular choke point, or a bottleneck so to speak, it's a long pipeline, with a LOT of variables. But early goal lock-down and extrapolating what those goals mean for the character is the big focus for me"
IronStylus also commented on the relaunch process as a whole, saying:
"...It's a big task, but we're up to it. The team is always growing, and it's changing. I'm taking on 3D currently, going to try to be a rounded character artist which can handle concept, modeling and texturing. We've recently nabbed a new animator. VFX is our bottleneck at the moment. They're an awesome team but VFX artists are hard to come by,
We'll grow, and we'll take on more. We currently have 2 big pushes we're on right now, and 2 more in the immediate afterwards of those. The thing is it's actually easier than champion development because we're not dealing with mechanics most of the time. The last relaunches have come with a host of kit reworks. That complicates things. When things are more explicitly VU's, that's when we have a lot of smooth sailing. Good case being Sivir. Mostly Visual changes, nothing balance related at least not that I'm aware of. The opposite example is Sion. He's a major project.
Relaunches can sometimes take less than 2 months to do. Champions can take 6 months to a year depending on how long each phase of ideation/pre-production/production is."
"She's on the list, but not the highest priority at this very moment. She's a challenge in a number of ways. Sivir and Sion are the ones we're currently chewing on."
On the topic of Irelia's hair, one summoner asked if it will ever get an update to her model's hair, particularly to be not so triangle-y and more natural looking, just like what happened during Katarina's update. IronStylus responded:
"She's not up for a VU any time soon. Not sure when the timeline will be for getting any substantial work done on her. She'd need to be re-rigged to get the same effect as on Katarina.
This will probably be fixed in subsequent skins but probably won't get any retrofitting for now. By the time we would retrofit the existing base and skin lineup she might be a candidate to do a whole VU on."