How to make an AAA game in Unity (or fail badly)

Oh god I know that feeling too, "the movie" in the head, I came to resent it so much!
1. it was too fast, it was pointless to note everything down
2. it lead to world too big and alien to others, which made me hard to connect with people because it was like I was 2 or 3 steps ahead of anything, it casted shadow on them and I was treated differently.
In fact a lot of my focus on rules was to put that to rest lol, ie to teach people and reveal their "inner creativity" and stop putting me ahead. That didn't quite work lol, revealing creativity did work though lol. Also have

But anyway, let's go back practical: In the context of thread, I thought it was much better to defend the position of how to increase quality of everyone. Those technique don't make someone into a genius. But they are not about "genius" creation either!

You don't have to be a secluded "genius" to output great works that convey the emotion. Talent is cool but not necessaries to do good works. And that's the main point! That is to increase quality of your works, these frameworks will help you, especially getting closer to the AAA ideals.

It's about offsetting the potential quality of everyone toward the best they can do! You do better with than without and that's the point. Also a lot of AAA aren't exactly creative masterpieces (not even mass effects) they mostly go into common place with enough twists on common takes to find a huge public. Mass effect is mostly inspired by KOTOR which come from star wars which itself took a lot of inspiration everywhere but notably from Joseph Campbell's "hero with thousand faces". They follow similar formula with just enough twist to distinguish from each other.



We are not talking about genius creativity and turning people into miserable artist like me lol I hope so!

 

Yay! Next page...

Alright, so here's a little video showing the strategy layer. I just spent 22 days working on this. There is a lot of stuff that I had to build and tons of problems I had to work around, including the UGUI event system being a bloody nightmare.

I'm going to do a mini post mortem on this stuff tomorrow, and show one or two of the tools I built to configure some of the data, etc.

All in all, I'm extremely pleased. My graphic design skills are finally at an acceptable level and the functionality is quite complete to a high degree of quality.

In total, again, this took 22 days to produce. I built all the ui stuff by hand (including building my own buttons because I really hate the ugui buttons). The drag and drop is entirely feature complete and does proper validation and the like. The scenes and props are good enough. I am also quite pleased with the result of the runtime portraits, as these look pretty good.

Visually I think it works, thematically I think everything fits together.

 

Good job

 

Good stuff! I'm probably not going to be the best person to judge the mechanics, since I haven't really played turn-based, but anyway...

Visually it looks good (although some variation in UI design between the panels might help). Definitely, the UI much better communicates the style of game. It doesn't (and shouldn't) look like a traditional first or third person combat game. The strategy-style UI immediately makes me think it's involves a more strategic, abstract style of play.

The music seems to drown out all the sfx or something, in any case it needs more combat sfx, clanging and bashing, and possibly would benefit from a Mortal Kombat style intro and outro to combat.

Anyway, I think the UI is a great step forward and really helps to bridge the gap between the mechanics and the visuals.

I can't really say too much more due to lack of experience with that genre but just let me know if you need a playtester!

 

I really need to spend a week really finalizing the audio - all the interactions lack audio and all i have is a single generic ambient track that really doesn't fit right.

I actually finalized the visual design two days ago, and went with the panels building out physical looking elements (as opposed to just black smears with text or something). I absolutely need a little more variety and it's easy enough to add some texture at this point.

But the audio is really what's most important. Good ambient audio will bring out those scenes - especially the training scene where you have the shadows and the trainer. Hearing the effects will really make it feel complete and bring it to life. And proper sounds on the UI clicks and the like will really solidify the ux.

 

@frosted agreed! I was about to say that when I could see a fight was going on in the background, and all I could hear was that menu-style soundtrack, it really knocked me out of the atmosphere. Looking forward to seeing and hearing more!

And great idea to do the narration, it'll probably help you out a lot to practise now so that you can do a better job of marketing it when you're done.

 

No, sorry, that's just incorrect assumption that "art is magic, because I don't know how to do this".

In general you apply rational/reasonable approach to pinpoint desired result. The less "reason" you apply, the longer it will take to make something.

Now... do you draw?

 

He makes 3D models I think that totally qualifies

 

If you tell me why it's relevant, I'll tell you if I draw.

It's not some "magic" bull, I never said that it was some mystical phenomenon. It's a simple fact, that good art tells a cohesive story, it is not simply a question of getting the object correctly drawn/modelled. Everything has to reflect a message, a mood, it has to be a snapshot of something that had a before and an after in space and time. If you take a part of one scene in a well-done painting or a drawing, or even writing, and add it to another, it almost certainly won't fit.

I'll try to say it again for clarity: there is absolutely nothing about art that is not logically/scientifically understandable. That's not the point I'm making. It's merely that art and emotion operate within a self-consistent, but not objectively logical framework, and due to the complexity and non-logicality of this framework, it's much easier to operate in it if you operate on feedback given by emotion, rather than trying to rationally learn the 'technique'. It's probably possible to do that, but it would be like applying yourself to learn a language that had little logical structure. It would be a pain in the ***. That's why I think programmers in general tend to dislike or 'not have an aptitude' for art, since they enjoy thinking logically.

 

People don't realize that "having something to say" is surprisingly hard ... though there is ready made message or emotion for anyone who want to do something and not bother about it, they are organized in genre, tropes, archetypes, formula etc ...

 

Do you make games applicable to this thread? For all the words that pass back and forth, it seems hard work and dedication never seems to pop up. I've seen plenty of great looking foundations from six week Uni projects, I believe one was called Trailed which looks like a start for something like The Last Of Us. So if they can do it, it can't be that difficult right?

@neoshaman if you have all the answers, where's the practical application situated upon your own set of issues? Every project has their own specific pitfalls, that'd be more valuable than the random amalgamation of tidbits. Whilst still valuable contextual application is always the winner.

If this thread is literally going to be about Billy, I'm not sure really what purpose it's serving any more.

 

In the spirit of adding something useful to the thread, I just ran a test with Doxygen for project documentation, and it looks pretty damn good. For a big complicated project it seems like a pretty neat way of referencing the codebase. Plus it forces you to be neat and orderly when writing code so as not to cause problems with the software.

I also made a flowchart for the asset store project that I'm working on right now. Making it really helps me and other people see where all the parts fit together.

Only wish I'd done this earlier.

 

When you have no experience with any sort of non-programming creativity (art, music, etc), it is possible to imagine all sort of stuff about the subject, and all of it will be incorrect. When you start drawing, or composing music yourself, all those incorrect ideas you had before will quickly go out of the window.

That's why I'm asking if you draw. If you don't draw, you should probably start drawing.

Or you could establish a technique that maximizes emotional impact you want. Same thing as approaching it logically.

Programmers in general don't have an aptitude for art because they almost never put in enough hours to get better at it. That's all there is to it.

Wrong.

 

Well it was more rhetorical, I know it's not that difficult so if you struggle with it you need practice. If a bunch of students can one up an experienced indie there's an issue..

 

@neginfinity I've drawn on and off since childhood, but I'm more of a writer than anything else. I haven't done it for quite a while though. I kind of moved away from artistic stuff into engineering/programming and now for games I'm kind of getting back there a bit. Anyway, if you've been following this thread, you'll have seen quite a few examples of my 3D modelling work.

As a writer, stories mean more to me in games than anything else. And artwork should tell a story. I collect a lot of digital art, and I rarely see screenshots of any game (AAA to ZZZ) that really tells a story, but if you collect a lot of concept art (as I do) some of it is extremely powerful, it really sparks your imagination. I can only hope to make a game that sort of does the same thing - not just through the 3D art itself but also through the gameplay, backstory and characters.

In any case, I think it's a waste of time to try to become an artist through 'rational means', you'll simply end up making stuff that is unsubtle and doesn't evoke anything beyond being a semi-realistic portrayal of reality. Agree to disagree I suppose.

 

@ShadowK

I have a very simple example from a very simple project I could show (this one https://dl.dropboxusercontent.com/u/24530447/divers/Quest_Of_Dream_guide_chapitre_1[1] (2).pdf ) but it's in french, it's a direct application and you can a see a how a small synopsis (not provided) is translated in a whole gameplay and it took me 30 mn to make and validate. It's a first pass, but not only it brought extra clarity and simple communication, it was also prescriptive on where to ameliorate the whole flow of the story while keeping it in line with gameplay.

In my current project, I use it to explore extra dimension I wouldn't have with those tools. I'm still trying to get those "exact sentiments" in depth, which lead me to actually contract a woman because the last tidbits is not something I think I can find by myself (the theme had evolve into very feminine territory on top of the cultural sensitivity).

I can't share the whole extended pitch and synopsis (33 pages), I wonder if those extracts will do, note that they contain extra reference that only "locals" from my place will fully get (all clans name are symbolic surnames applied to woman here), there is also a bunch of thing I don't know how to properly translate into english (veil might be a close approximation, not sure), every problem is based on local expression too, I'm making sure the overall structure is understandable to "foreigner". There is subtlety like the main character is not supposed to save each clans (unlike traditional narrative), but provide support so they can overcome their own problem.

In red you have the question I need to answer, from a structure point of view, the work is complete, I can put symbolic approximation and be done. But since I want this work to lay foundation for future works I prefer to find the right sentiment applied. Also the problem I have won't spill over plot point later, they are all contain within a bigger thematic establish ealier (the concept of clan) so if I decide to change the concept of a part he would still be under the functional goal of the story beat which is solid.

A lot of work was put simply in documentation (mythical creatures from tales like youglanglan) and find match for the situation (daddy pig are demon of gluttony, youglanglan was in a tales about gluttony and the whole clan concept is about measured used of resources).

That's the main difficulty with other project I worked on, I can't just imagine stuff, I must lego reference within the constrain of having entertainment and spectacle on a culturally true and sensitive framework. In fact right now I don't care that much about how the final scene will be played, those concept are the parameter that decide their contents.

Also while this part is later in the story, defining them will also define the content of earlier scene that are currently undefined and which are the exposition that lead to that. Since the core identity is found in those later scenes, I don't go and waste time creating those exposition, though I have random note that constitute potential idea to use for later, the good one will be promoted in time. The story is simple, it's all about the symbol and the meaning of the journey.

For example here is the pitch of one of the clan:

and an illustrative story that goes with it:

It really just a big puzzle of references (Josephine was the wife of Napoleon, she had live on my island and had some negative impact, and some stuff happen to her statue here, there is a whole mythos). Technique helped me keeping all of them in check and concentrate on the minute details rather than the big structure and mundane progression. Former careless attempt had fallen appart because I didn't control the flow of things, using the framework I propose this one only got denser and surprise me, only going better over time and taking a live of its own.

 

You are our new art game teacher

 

@neoshaman cool, so you've written out a fundamental basis. I think the main point is, I'm seeing little in the way of foundations / basics / prototypes or anything that could apply to an actual applicable real world game someone could play.

What we're essentially doing is covering the fundamentals of large game design, which is fine but only really applicable once you do something. AAA often find out what works in theory, often doesn't work in practice.. They under-estimate time scales etc. and the impact of said choices multiplies as your resources lower. Also factoring in, you can't actually make a game design document for your specific project without at least a fringe understanding every component from A - Z. So the theory part of it has little relevance until you've got pre-requisite experience.

If attempting even a mid level PC game, again you can't spend every two minutes to every one learning new principles.. There are pre-requisites like you know how to make art, you know how to make music, you know at least an intermediate level of coding and you've previously created a prototype or two in the same vein allowing you to find toolsets that work best for you as a lone wolf or as a team.

Which is getting the experience and messing about is the fun part, if you want a career at some point you have to take it seriously which is the dull bit.

With that said, most things seems a little speculatory. Ultimatley we're talking about AAA but when there's a lack of even low / mid level PC game prototypes shown here (actually in terms of quality, everywhere) never mind large budget titles, so it appears some priorities are a little out of whack. Today I'd be more concerned with what some of the stronger indie's are able to do with basic stuff available to them..

Scoping out some games like "Pamela" which I'd class as a mid level PC game, that's more along the lines of my competition not whatever Nathan Drake is up to.. I've been told the MMO I've been prattling on about is getting released in June. An MMO made by a team of two? Originally and expanded to a team of eight..! Again, it's not AAA I'm concerned about.

P.S they said the UI is getting an overhaul.

 

You forgot to say "Except for @frosted who is doing sick work and actually making a game".

 

I actually did, then I revised it and forgot you.! Sorry buddy!.

 

Just to give a sense for the scope involved with putting together the strategy layer shown in the video above, it required a pretty immense amount of labor and S***loads of code.

An outline for the code required is as follows:

120 files worth of C# to cover the UI elements: This includes everything from hand writing a button, to dealing with the drag & drop system, to building out a "style" system, to object pooling, tweening. The core drag and drop logic requires 15 classes, there are over 20 classes devoted to specific drop slot types associated with draggable items.

30 files worth of C# to define and configure meshes: This includes some tools for defining unit types and associate those unit types to some set of mesh + range of textures + ranges of accessories. This also covers basic stuff like configuring offsets for hand slots and the like as well as equipment prefabs. Because I'm working with a random mishmash of meshes there is a ton of variation in what's possible to change and what happens to be compatible.

40 files worth of C# data defining stuff that requires serialization for the game save: This includes everything from saving characters and equipment to inventory, mission types, equipment properties, abilities, upgrades, temporary conditions, etc. This also includes stuff like providing adequate data to correctly reassemble the character mesh and reliably generate portraits at runtime that remain exactly consistent from scene to scene.

It took something in the range of 2-5 iterations on absolutely anything visual in order to produce a pleasing result, and I reorganized the basic functionality a few times (how stuff like selection works, what stuff retains selection, what stuff competes for it).

I'm sure that given some revisions to the code base I could strip 1 out of 2 or perhaps 2 out of 3 classes, consolidate, refactor, etc. But I don't really care. Doing this in 22 days from the the UX and UI to figuring out exactly how the game systems work:
- the different types of jobs each hire can have
- how the leveling system works (randomized rolls, minimal options)
- the different kinds of missions and side missions
- the different kinds of character states that persist through the game.
- at least an initial pass at defining out the visual characteristics of the character classes.

It was a pretty insane amount of labor.

 

Nice, i like the blocking move and the Dark souls like stamina bar
On what parts did you work on the game ?

 

Magical floating zombie hovering over the car @1:52.

 

@ShadowK
I never said my current project was AAA you should have precise that part

The design isn't large though. You asked about MY current work in which I apply this. I do have some prototype laying around, but not for this game yet. In fact not applying this type of work killed most other projects I started because they were aimless, lack coherence and I was just tossing stuff until it sticks. So I had it the other way around too. I'm working to contain the risk and right now it works well because my current project is full of uncertainty which solution aren't readily available. Art direction is its most potent risk.

I'm well aware of that problems, I worked on a AAA games around 2008 (before the subprime crash which hurt the financing of the project and ultimately cause the demise of the company), it was never released (Totems by Elsewhere Entertainement) The project suffered greatly from inappropriate handling of the story (which basically is the planning for place, character and props), so much that, even though I wasn't an original member, I was one of the rare employees who could understand the scope and put it back together, using the very technique.

For proof those kind of technique work, I shared video where they praise these kind of technique, I can also point to DeMarle's talk handling of deus ex Human revolution (there is two talk http://www.gdcvault.com/play/1015027/Building-the-Story-driven-Experience I don't which one is it, but it is essentially one before and one after release of the game)

It work for any scale, it just mean "know what you are doing". I'm basically grounding the works, then I will have no problem expending on it. It's scope agnostic and absolutely not speculatory. It really do help. Planning matter. And it allow me to keep things flexible to adapt to production issue and isolate creative problem so they don't spread elsewhere. It turn production more into a grind toward a result rather than solving creative stuff on the fly, and since you have definitive targets, it work better, you only have production issues to solve.

Now the prerequisite are fine, but the discussion was never about (individual) proof of that, but documenting (I think) solutions to look at.

But fair, let's show some stuff, I don't think you will be appease by those! Like I said, where I'm I don't have much to show right now, most of what I show come from leftover into a dropbox. So bear with me It does not prove anything, but I think it will be cool if I can show stuff so it's not just talk.

Spoiler: Showing preliminary stuff about my project

The game I'm making Is a simple adventure game, I evaluate some risk and investment as high so the scope is rather small (story is very short). All the risk are in the art direction, it ask for significant documentation, studying and solving.

None of the stuff shown below was made to be shared and shown, they are internal stuff to me, so the quality is not where it should be for that.

Here is a low poly, unlit character use in early test visual aspects (unlit visual target), it's not the best works, it was just a quick proxy to validate basic visual direction (I want something inspired by stereotypical Caribbean painting with silhouetted characters), also NOT photo realistic. No AAA target.



I also had random fun decimating the topology to see if I could use a minimalist star rig (ie skipping some bone on the thigh, arm, fingers, etc ...) and have something decent, not for production model. Eyes is very bad, very very bad (I didn't use concept art and improvised while modeling directly).

Spoiler: If you like rig problem

I think I got a good range even without the thigh bone, still limited, you can see the compromise made in the topology in that region. The rest of the topology aren't teh best either, but that wasn't the point (though I corrected some volume in further revision lol even though it's a toss away model). Might be useful for background character or LOD who have limited movement can bear being a bit stiffer.

It taught me I have extra work to do with black characters (not painting them brown and call it a day), I kinda learn the ugly truth about black representation; black (and not brown) skin and afro hair are under represented and under documented, and when they are the HUGE lack of care is astonishing.

This is an extra risky investment to make, I wrote about it there: https://forums.tigsource.com/index.php?topic=43003.0

There is no shortcut possible I need to recreate a full visual language that address their specificity, they need a full investigation. Which had to the burden of the project. If you don't know what "4C hair" mean, you have no business telling me how to do black hair lol (unless you have 4c hair).



I made a mock up to show a skin target:
Typical black character in illustration (brown skin):


Actual skin target I want:


Also random early concept art, note that I'm now a game designer, I dropped drawing 10 years ago I need to go back to practice

I know about the neck and overall anatomy problem lol, I didn't know what I was drawing when I started, I fumble my way solving it, and in the end many small decision made there was discarded or kept. I learn that I needed a 360° study on afro hair because the usual visual tropes I kept using before won't work in this project I have much more diversity to cover than the crappy and sloppy bubble hair everyone make.

Translating it in 3D is even worse, I had no idea I was doing and reference where all so bad! Only Japanese cared to look at what black hair actually was and tried to make ideal version of it (Sazh and dazh notably, in ff13, look at how they have different textures to show off their character's personality). Say no to endless hair helmet and cornrows (who actually don't know how hair behave with them anyway) Though the last mafia 3 has some impressive hair on his main.

Here is some shader test, this one is to go per pixel with a typical hair shader (lux)



Using the same shader I experimented with using noise to approximate the tightness of afro curl, and idealize the hair as "glitter" effect.



Turns out there isn't much tool to draw anisotropic curl anyway I need to make one from scratch just to test hair stuff.

Afro hair have a specific hair diffusion profile anyway (I mean at the viewing scale, their volumetric property scatter light differently). It's also hard to find good professional photo of afro hair to use as a reference, those that exist tend to be fashion photo who only know one or two haircut and heavily processed in a way that destroy the hair characteristics

So much work to keep that cultural sensitivity! Stuff you can skip if you do mass effect like scifi works where everything is possible! Here I'm constrain because those thing are the main subjects of the game.

 

Neither did I .. Taking a vested interest in these mega scope games, doing them is another matter.

Not necessarily just yours, it's cool to have an opinion on it. In extremely rare cases you might hit on something most too close to their projects didn't think of. Again not you here, but I always find someone's logic flawed when they state an opinion as fact without having done it or not really having the experience themselves.

P.S We've all been there, this is the problem with making large games and one reason why some companies closed off the dev teams. It's a good idea to realise that your project is lacking before you're too deep down the rabbit hole.

That's cool, so have quite a few of us on this forum. I've found it's very much project dependant, you can swap out a team leader / project manager / lead developer / lead story writer and because of that the core basis won't flow properly. Some times it's straight forward, whilst certain techniques work others work better and they don't always require an essay's worth of research to deploy.

Well that's the problem in a nutshell with these sorts of threads, most embarking on a journey like this will of most likely made a few prototypes / games of a smaller scale and it's just a matter of expanding. It's not particularily difficult, just time consuming.. You'll already know roughly how to plan a game out, what's involved and roughly how long it will take you.

I honestly believe one of the hardest part when you have experience is idea's, good one's and I'm not talking just about indie's here either. The amount of padding in some AAA games, like fetch quests etc. or repetetive cycles to keep play time longer is astonishing.. That doesn't just apply to the mechanics either, we're talking art and it is actually quite a difficult thing to get around some times, that's where a team comes in handy if nothing more than just to bounce idea's off them.

Secondly polishing everything up, the amount of innate detail / polish in large budget games is amazing. Have you ever tried optimising a large game for console? Polishing some silly little mechanic like Batman's batclaw and all the connectors / side positions and all the little weird behaviour that could cause?

No it's cool and you obviously have tons of artistic talent (more than me), I'm not much of a drawer.. I say I suck at drawing, although it's more about being out of practice than anything else. I can't say it's ever really hindered me in 3D, when I was younger I drew pictures of FF characters / architecture a lot and I did take classes.

Anyway, it's the whole walk before you run kinda thing..

@zenGarden

I've just been generically helping out when I can really, back from when they were using Unity having issues with running such a big game in Unity 4.X. As the game expanded, the owner of the company got more staff in and they developed relationships with publishers etc. so I went down my own merry path and focussed back on my own stuff.

 

That said, I'm a non coder, I know I can't make a AAA on my own right away, in fact all that planning talk help me narrow down what I need to learn from code too! I have start coding though, I'm just slow and non optimized, architectures issues scares me the most. That's why I Have more interest on scope than polish "now", I don't have the chop to handle the optimization that would make that polish works.

At the same time if we forgo the solo AAA for indie AAA, making janky stuff with great asset that is your forte can attract talent that will help on the janky stuff!

Anyway here is the mass effect 2 level planning and production, there is surely lesson to take from them:

 

This asset was also brougt to my attention
https://www.assetstore.unity3d.com/en/#!/content/56762
seems to have the right kind of tools and seems rather complete

 

I actually read through most of your discussion on tigsource. You're clearly a completely obsessive lunatic, but I very much respect the completeness with which you approach exploring and experimenting with this subject.

I think the way you explored expressiveness and ebony complexion is really pretty fascinating. Dealing with dark skin and expressiveness with simple graphics appears to be really complex.

 

Lol it's generally what is expected when you do art in artschool lol (well meaning oriented artschool, the beaux art type). The thing is there is surely people who are more qualified that me to do this job faster, because I have seen some artist doing it without giving much thought, though they don't have a gap of 10 years in practice and might have been doing dark complexion for much longer.

And those concern are mostly typical graphical concern you ask for every artworks anyway (readability is always important).

I mean artist are trained to split hair in four, look at that blog post about "tangent", most bystander would say why bother? If you work for a big established animation studio, well it will matter lol (oh well he have updated with a disclaimer lol).

http://willfinn.blogspot.com/2010/04/art-school-101-tangents.html

Now there is a thing like good enough!

People think that artist just sit down making pretty picture with "talent", actual skill is heavily analytic. Sometime I don't get why the divide between artist and programmer, both are so deep in logical assessment! (shhhh! .... I know why, art don't break, code do). Most good animation blog that talk about skills are massive hair splitter.

I got deep in dark complexion mostly because of carelessness (I mean absence of documentation or footnote that don't acknowledge the specifity), so to establish precedent and diffuse whatever I learned. Also because it turns out that how things are set up I will have a lot of black characters, so I should get it right first. It's a long term investment. It's just that around me there is a lot of black person, and seeing 3D model or character, most of the time, their complexion just feel "wrong".

 

@neoshaman

There is such a thing as a tech artist, where the dividing line between artist and coder is non existent. Generally they are multi-talented guru's that could easily build an entire game by themselves without really struggling and take on advanced lighting / shader / API concepts with little issue.

To companies, they are worth their weight in gold. As a lone wolf, that's who you have to be and in a small team the divide is slim.. We have one "dedicated" coder, the rest are multi-skilled.

 

While somewhat true due to their multi-role nature, where I worked, software engineers still got paid significantly more than technical artists.

I will not argue that that's who you have to be as a solo dev though.

 

Umm, kinda depends.. Y'know what it's like in big companies, a technical "artist" was actually more of a junior role last place I worked with the highest rank in that field being technical authority. They got paid probably about three times as much as a software developer / engineer. Although that was again dependant on which department they worked in (engine development, art team etc.)

In other places technical artists were lead roles, all I really know is a good TA is a talented fudger.

 

Looks great! I think I noticed you being less active on the forum during the last few weeks ^^. Your hard work sure has paid off! The only advice I have at the moment is to clearly communicate that you need to drag'n'drop the items into the slots. Maybe make a popup explaining it, that only goes away once you've drag'n'dropped the "OK" button in the "I understand" slot. Otherwise I'm sure some people will only try clicking on things, and from the video it wasn't clear to me if you've implemented that as an alternative input method.

I did look around on freesound.org a bit to see if I could find some sound loops that could be adapted for the training background ambience, but to be honest most of the content on that site is crap in my opinion. Many recordings even have clipping in the recorded signal because someone didn't tweak the levels properly for the recording or crushed it to death in editing.

And holy crap, at the amount of your files just related to interface stuff! Seems like I wasn't too far off with my estimate that this stuff is some of the most tedious and time consuming stuff. I'll try to minimize the GUI stuff that I need for my project.

I skimmed the thread and it doesn't look too crazy to me. I think people are underestimating what "proper research" for an artist means. E.g. if you're a concept artist designing ships for a pirate game, at the end of your research you'll basically know how those things work and what the purpose of each visible part is. The limiting factor for the research that a committed artist does, often is the budget of time and money available for the task.
A teacher at my university once told us "Designers need to know everything" and he was dead serious. If you read biographies of accomplished illustrators you see quite many that actually didn't just receive a traditional art education but studied something like architecture, geography, history or similar.


@neoshaman: Interesting stuff. I have a total lack of cultural sensitivity and I probably should paint more people with darker skin color. You've collected some good references. About the two core problems of very dark skin tones being underrepresented in art and games and the problems you describe with making them expressive enough in spite of lack of contrast between eyebrows and skin etc.... maybe those are connected. Maybe the reason the extremely dark skintones are being avoided is that they bring so many complications with then. You have to pick your battles and in any generic AAA project I'd personally consider that a potentially bad time investment, if those super dark skin colors only will be a small subset of the characters in the game. If that cultural aspect is a big part of your game, sure, go for it. I'm just not sure if there is a great solution for the problem you're facing.



Does anyone use the "projector" component in Unity? I can't get it to work properly. It seems to color the whole meshes it touches instead of projecting the texture. I'm using a material with one of the two projector shaders. Are there any known bugs?

 

Projector is a very weird component - AFAIK it redraws what's being rendered (although i believe there is some optimization so it's usable on terrain). The biggest problem is generally with tiling textures - in order to make the projection work you generally want to clamp it. I remember trying to use it like 3 times - the first time I couldn't figure out how to use it. The second time I did and was like "oh, that's not hard". By the time the third time rolled around, I again, couldn't figure out how to use it.

So it's got some weird stuff about it that makes it not at all intuitive, but once you figure out how it works it's very simple.

Sound is really tough - there's just a massive chasm of difference between quality audio and most of what you find. sounddogs.com has a pretty wide selection and is usually my go to for scanning through audio (although to be honest, I've never ended up actually using anything from them).

I think I will probably be better off generating ambient audio in fmod using small bits and pieces than trying to find something precooked.

Again, I'm sure you could cut those numbers in half without too much work. But there is a lot of small detail, and more detail generally demands handling more exceptional circumstances with explicit code. UI code in general is the most tedious code one writes.

 

Thanks! Maybe I'll stick with just using lights instead. I can get ~90% the effect that I want that way.
I had the texture set to "clamp", I googled for others having problems with it and I just couldn't get it to look remotely correct. I could try again in a fresh project, but I'm not sure it's worth the time. The projector sounds to me like one of those components that never got a robust implementation because it's too niche.

 

So been checking out Houdini, first impressions are quite positive although I'll go more into depth later. Here's something I came across on my travels:

Also this is an interesting watch for any artists doing the A / AA thing:

 

Epic has done the tech, and it is available to users, all hard work has been done for you, this is the advantage of 3D engines like UE4.

 

Well I wouldn't go that far, sure I'll save three or four weeks on skin / hair shaders.. But it's the character creation system and artwork that'll be the hard part.

Still, not complaining.. I'll take whatever I can get.

 

More than that, if you are someone that does not know how lightening works or does not know shaders , they bring you realistic rendering out of the box.

Like they braught new animations possibilities
www.youtube.com/watch?v=1UOY-FMm-xo
Or many improvments
www.youtube.com/watch?v=BXcw2IrIinc

You can concentrate more on the game content and let more of the AAA high tech and tools to Epic.

 

Mmmm, but I do know how lighting and shaders work.. Yeah if you're new it'll take a fair while, remember the first time I tried to make an ocean shader eeeshh.!

Still it's one of many things I don't have to do, so again I'm not complaining..

Personally I always thought that was the point in using a pre-made engine, to make your life easier not harder. Although I've learnt over the years, that isn't always the case.

 

You take a team of level designers and basic 3D artists that does not know lightening or deep engine and shader programming , but that know how to code gameplay, for this team that is not interested in learning what is inside the black box , but needs the maximum functionnalities out of the box, UE4 is good.

It is like a car, you can have someone very good driving it perhaps winning races, but he does not know how work deeply
the engine and could not customize it and let that to the constructor , and ask teh constructor for better engine or pieces.

 

@Martin_H

To be frank I didn't come first from a purely "sensitivity" perspective, I came from a purely visual design, the setting enforced it on me. The thing is once I start trying to envision the characters I had trouble finding the visual language and pattern to express it. Which lead to research and which lead inevitably to cultural sensitivity since representation is a huge deal in this domain.

Because ultimately it's a chicken and egg problem and it force you to walk on eggs, there is obvious historical reason why there was no care, and this carelessness lead to a lack of focus on those specificity which lead to scarce resource to solving them (it was a big deal in early photography, emulsion where not design with black in mind, despite being a huge market, wood and chocolate advertising is what forced change in the tech history is funny! ). I think I need to build resources.

Spoiler: Skin in photography (found doing my homework on skin)

In AAA nowaday it's less a problem because now you can literally just measure it! And you start to see a lot of black character for that reason, they are mostly just scanned! And skin is way less a problem than afro hair! Hair is difficult and that's just flat, dull straight hair, they align in a plane, now afro hair is a volumetric structure with complex diffuse subsurface scattering that change depending on hair cuts, view angles, location in the head and environment condition Only Japanese can do it apparently

The truth is I'm just doing my job as usual, no agenda.


edit:
BTW if someone where I can buy good scan of dark complexion skin?

 

I don't, but maybe you can ask someone who offers such products to create some good scans.
www.texturing.xyz
They sell things on the substance store too.

 

@Martin_H

Wow that's a good find for looking at and referencing details, too bad they only serve white skin there I'm worried that substance would not be legitimate reference because it is procedural.

 

Just send them an email and ask them nicely to scan non-white people of varying shades too. What do you think they'll reply? "Sorry, we scan white people only"? I'm sure their selection only happens to be whites-only because they happen to have those people at hand and willing to be scanned and would need to put in more effort to find black scan models.

https://store.allegorithmic.com/libraries?by_category_type_id=3

They sell the same assets in the substance store, seems like they have expanded beyond 100% procedural, by partnering up with other companies, because such scan data could be used in Substance Painter.

 

It's more like I was thinking: "will I be able to afford custom demand?" Though speaking of cultural sensitivity I realize that may be, in an afterthought, using "serve only" was a bit cringy, I didn't make it on purpose, I was a bit careless .

Also it seem these are reference not texture.

However dropping specificity of skin in the search and using google image to eyedrop results I found that:
http://www.3dscanstore.com/index.php?route=product/product&product_id=538
It's the only one with albedo and not diffuse/painted, it's not even dark enough.

I think it would be simpler to create my own 3D scanners.

 

Turn down the Lightness in the color settings?

 

@Billy4184

You would assume I would ALREADY tested that too



It's a customized lux Shader, I optimized it by approximating the LOD blur to make it more versatile and less expensive for very little differences(I don't really notice it).

But there is skin variation over the body and different behaviors, It's not as easy as darkening a skin (yes they belong to the same chromoscopie range), there is many type of black skin too and it's not uniformly darker. It feels wrong, so I need to investigate that.


Look at the guy at the extreme left and the guy behind him.


I need to do it like kojima, measure the albedo directly through many reference.

I'm also aware of state of art paper on the subjects (Jensen's research notably, Devebec, Jimenez and Penner), notably this one for "global" colors:
http://graphics.ucsd.edu/papers/egsr2006skin/egsr2006skin.pdf
THough I hav eno idea how to use it. There is a lot of stuff I'm trying to peruse that are a bit above my level

 

Then you have the material editor which allows you to experiment without knowing how to write a shader normally. Not to mention copy shaders back and forth between Unreal, Shader Forge, and other material editors.

 

I think this tools will be a big contributor of the director's style creation that Billy want. parametric structure editor.

http://forum.unity3d.com/threads/wi...d-parametric-modeler-for-unity.330869/page-13

Unrealeased but very promising.
http://forum.unity3d.com/threads/wi...-modeler-for-unity.330869/page-8#post-2582742



About collider generation
http://forum.unity3d.com/threads/wi...-modeler-for-unity.330869/page-6#post-2510548

Spoiler: More samples