Search Unity

Does Unity make game dev too difficult?

Discussion in 'General Discussion' started by GarBenjamin, May 18, 2017.

Thread Status:
Not open for further replies.
  1. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    I know what you're saying and sure there were always times of frustration but overall for me it was always fun. That's why I had it as a hobby. I looked forward to the time spent on it. For some reason the past couple of years it has become more like a chore. That's why I am checking out other things. Because the biggest difference I can see is I've been trying to force myself to get into Unity.

    I am no longer doing that. I may play around with Unity from time to time but I will primarily be using other apis and perhaps even other languages. The way it is architected, the game objects, etc it's just not the way I want to develop. It is a great game engine just not for me. I don't think any game engine is really.

    At least with Unity I did complete some games and I spent a good amount of time working in it not enjoying it all that much but still I did it. Other game engines I rarely get past the screenshots and documentation sections of their respective websites. And I'm still visiting here and participating in discussions and almost certainly will continue to.

    Anyway yes for me game dev was always fun. It had its frustrating moments and at times was quite challenging but I enjoyed it. I guess if I still find game dev feels like a second job even when using other stuff then I will have to think maybe I am just burnt out on it and need to stop it completely. All I know is if it is fun again I will spend some of my free time doing it. If it is not there is no point in doing it. It would be silly to.
     
    Martin_H likes this.
  2. zenGarden

    zenGarden

    Joined:
    Mar 30, 2013
    Posts:
    4,538
    30 min for a character is too low.
    Putting some days week in Voxel and animating a character i consider it as fast , once it's done you work on everything else (level design, gameplay).

    Like game level assets (house, trees), they will be copy and paste everywhere on your level, so they deserve minimum work and attention, like some hours or some days. Below it's irrealistic or you must lower a lot the details and go very simplistic style.


    Magica has tools like rectangle tool to fill or delete rectangle regions to speed up your work, and an image importer to make a plane voxel base to start with and you add volume.

    What would suit you would be some voxel 3D game with a voxel level editor like Minecraft.
    Making assets or characters is using voxel cubes , even a kid can make content.
     
    GarBenjamin likes this.
  3. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    Interesting you say that because MagicaVoxel can and is used to make models for Minecraft as well as ultra low poly games in general and I have also been searching for 3d modeling software for kids/children/everyone/non artists/ etc because that is what I am after I think.

    I come across a lot of different content that is at least a little interesting. Last night I searched for game jam tricks and tips because some of them make some damn nice art in an obviously short period of time.

    I guess I relate more to these developers than I do the ones focused on taking so much time. What he accomplishes in a day or two here I think is what the norm should be. I wish he had specified hours instead of days. 2 of his days may well be the equivalent hours of 2 weeks or more for me and some other people.



     
    Last edited: Jun 9, 2017
  4. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    4,378
    This part of the discussion was fruitful, I have an idea to combine smooth teddy with skin modifier, using stroke as way to build the skeleton of the skin modifier on screen space like smooth teddy ... as long as I figure out convex hull for complex junction ... well it's in the todo pile, next time once current task is done.
     
    GarBenjamin likes this.
  5. zenGarden

    zenGarden

    Joined:
    Mar 30, 2013
    Posts:
    4,538
    There is no norm, this is a matter of experience, if you are not experienced enought you'll put lot more time indeed.
    In these videos people are already skilled to make a game in a short time, they have already practiced a lot.

    Anyone can make basic food like anyone can make a basic game, while making top tasting food needs some practice and skills like knowing the subtle ingredients that makes the difference, perfect cooking time etc ...
    This is the same rule to make a game, the more you use your tools and Api, the better you become and faster you turn ideas in a game.
    Perhaps using Voxel tool each week you'll start to become better , you'll find it easy and you will work lot faster and start creating anything.
     
    GarBenjamin likes this.
  6. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    No doubt like anything else the more time spent doing something the better a person will become at it. To a degree anyway. A lot of it depends on if the person is deliberately practicing looking for ways to maximize speed.

    I have about a dozen more tools to check out and then I'll figure out what to do next from there.

    I can already see that if the model building was just done differently... a different workflow it would be considerably faster. It occurred to me perhaps a code-based approach would work well here as it does for level design. I mean like sketching out a character in 2D image and then processing that building 3d object OR even defining faces entirely in code like the good ole days. So I will explore those at some point.

    I think with three images it should be fairly easy to convert into a simple 3D model. Yes it may need some editing after but the bulk of the creation would be as fast as one could sketch it out in 2D and run it through the conversion process. Of course idea is to limit work not create more of it but the ideal might be to use an image of top down view, front view, bottom view and side view.

    Best results would be a view from every side and possibly that would still be faster than modeling it all but I think the TFBS would cover majority of cases fine. Just need to set ratio of pixel to unit size in model. Color coding would be used to identify separation areas and along with x or y position establish relationships between the different images. It might just work. Kind of interested in trying it out because nothing is faster than the codes man. But I do want to evaluate all of these tools I have downloaded first.
     
    Last edited: Jun 9, 2017
  7. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    4,378
    You can also do a library of objects and module over time, so you don't have redo gfx at all!
     
  8. sngdan

    sngdan

    Joined:
    Feb 7, 2014
    Posts:
    934
  9. neginfinity

    neginfinity

    Joined:
    Jan 27, 2013
    Posts:
    6,434
    Nope. 100% certain of it.

    One very important thing that needs to be understood when modeling is - "forget about precision". In a programmer mode, you (well, I) would want a precision, full control over everything, knowledge of every aspect of everything, etc. When making a model or a drawing, this approach "precision-focused" approach needs to be thrown away. And that makes it opoposite of code-based approach.

    It is a hard problem.

    Images will be misaligned, may have perspective errors, and will need to be adjusted for lens distortion. This already calls for opencv, computer vision algorithms and feature matching. Basically... photogrammetry.

    If the original picture is a sketch, then trying to use them in this fashion will result in disaster, because 2d sketches often either pay less attention to perspective than a photograph would, or, in case of a cartoon/anime characters sometimes ignore it completely and break perspective rules.

    I think It flat out won't work. You're heading towards photogrammetry. Photogrammetry is not fast and is not simple.
     
    Last edited: Jun 9, 2017
    GarBenjamin likes this.
  10. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    I'm always thinking of simple and imperfect not grand and perfection. What I am thinking of is pixel art. Like making objects in a 16x16 grid or 32x32 pixel grid. Top view, bottom view, front view and side view. Pixel art. Not freehand sketches. Not sure it would be any faster but possibly.

    Basically my first attempt would be to have it identify cubes and their relative dimensions and positioning.

    One of these tools may be the solution though. Tonight I will start installing and testing those.
     
  11. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    4,378
    That's still photogrammetry, if you can do it for 16 pixel, you can do it for all of them :D I research the faisability of this at some point before it became big and agisoft had an easy solution. At it's core it's just trigonometry but with a lot of hidden variable to be inferred and fuzzy inputs. In control environment it can be feasible but that's still as much work than modeling directly if not more, no gain.

    But what you are saying with drawing per view is basically a sketchlab like stuff
     
    GarBenjamin likes this.
  12. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    Of course, thinking about it from a programming perspective makes it clearer... maybe I have been approaching it wrong. Instead of messing around with box modeling extruding it all maybe it would be worth investigating just using primitive shapes period. Like to get a unique look I could make a shape such as a diamond or better half of that a pyramid. And then just position, scale and rotate those to build up the objects. Thought of that for 2D before but not for models.

    Another thing to test. :)

    Course nothing will compare to speed of just using rectangles and texturing them... and actually that led to another thought.

    A couple more things to try.
     
  13. neginfinity

    neginfinity

    Joined:
    Jan 27, 2013
    Posts:
    6,434
    OpenCV + twin cameras + object on chessboard grid.

    Still going to be pretty hard to do.

    Well, there are metaballs/metaparticles, which are frequently overlooked and then there are distance fields.

    Both could be used to generate geometry via marching cubes algorithm or an equivalent.
     
  14. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    4,378
    Have you seen minecraft and the army of visual clone? :p
     
    GarBenjamin likes this.
  15. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    Of course! lol I meant cubes as in literally only a single cube no arms or anything. Can slide and jump around via "magic secret ingredient".

    One nice thing is thinking about all of this stuff does renew my interest in game dev. You may be right though maybe I should just use the MC approach. It is what I would default to as a programmer. Many times I have thought how easy it would be to build a little animation tool to animate a character made completely out of primitives. It's just rotations & related translations. I could have it spit out code to copy and paste in or save to a data file.

    Well I will get started tonight evaluating all the tools I've downloaded during my research the past few days. After that is done if it isn't solved I will look into my own custom solution and say to hell with the normal modeling stuff. Build my own modeling system one way or another whether from 2d to 3d conversion or some kind of primitives-based builder maybe some kind of grid based snap together thing. I don't know yet. Probably take a few iterations to find the right approach. And of course an animation tool for it.
     
    Last edited: Jun 9, 2017
  16. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    4,378
  17. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    I've seen that in my travels around the net looking at various dev kits. Strangely one of the things I downloaded that was supposed to be a modeling tool has a name very similar (maybe the same). I hope it is a modeling tool and not this game engine. lol
     
  18. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    4,378
    It has modeling and animation tool embeded, not sure if you can export though, but any modeling tool would allow the same, no need for a specific set of tool, it's just parenting of cube objects.

    The top one is the real texas a game that was well received (and is quite old) and only use cube to express everything.
     
    GarBenjamin likes this.
  19. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    I like the animation style it is the same concept I had for that last 2D game project I was experimenting with.



    I need to get back to that at some point. Maybe winter. I was happy with this graphics & animation style. Let me get things done so I could focus on the game.
     
    Last edited: Jun 9, 2017
    neoshaman likes this.
  20. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    @neoshaman thanks for posting Craft Studio and mentioning its built-in modeling tool. Yes that is basically what I'd end up doing. And seeing it in action I don't think it would save much if any time over just using the same primitive-based approach in any modeling tool.

    So that saves me time. Scratch that off the list. I'll check out all of the tools tonight then it becomes a matter of whichever is best I will use. I think at this point the options are down to have graphics & animation, have graphics & not animation and finally don't have any graphics and show blank screen.

    Ultimately I'll end up just dropping to the most basic ultra low poly / primitive-based approach possible. Super simple graphics and animation. How simple all depends on how good these tools are. The easier they make it to create more detailed objects quickly the better odds of me doing that. Otherwise at the lowest end cube text over it "YOU", tall cube with text on it "tree", etc. I think that is too low although for protyping might be perfect.

    I don't actually need this now anyway. But when I next do a 3D game it will come in handy. Spending time now so I don't need to waste time on this when I finally am in a game dev mood.
     
    neoshaman likes this.
  21. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    @neginfinity going through all of the crap I downloaded and I found one that is interesting. You may want to check this out. I guess it is supposed to be part of Windows 10. At some point in the past couple of days I found a download for it that works on my Windows 8.1 machine though. Probably an older version than this new release shown in the videos.



     
    Last edited: Jun 10, 2017
    EternalAmbiguity likes this.
  22. neginfinity

    neginfinity

    Joined:
    Jan 27, 2013
    Posts:
    6,434
    Can't. I'm on windows 7.

    I looked through those videos before, and concluded that it ain't Teddy.

    The doodling tool is a plain old half-assed 2d extrusion. The rest is repositioninign primitives. Even though someone could quickly make a pigeon out of spheres, the sphere were still intersecting each other instead of forming a single model.
     
    GarBenjamin likes this.
  23. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    @neginfinity there has to be something out there that does it. I remember coming across one or two over the past few days but they were the "big boys" kind of companies offering them. Where you had to submit a request to check out the software and one of their reps would get back to you. So of course I didn't bother with those.

    This whole experience this past week has made me think of art in general again. I think perhaps one of the biggest differences between artists and non-artists or perhaps just between normal people and myself is I cannot relate to art the way some people can.

    I've always been interested in colors. Color theory and just making nice palettes. I don't always do it but I do find it fun and could probably waste hours just playing around with a color wheel trying to find a cool palette (not cool as in temperature but cool as in interesting and "it works").

    But like that last bit was a good point. So many times when I have spent some time in art tutorials they will say something along the lines of "think of the mood you want to convey..." and I can't relate to that. There is no mood I want to convey.

    I can understand sure there is horror/fear, happy, defeated and so forth but they have no meaning to me in games or art. "... based on the mood you wish to convey you can now choose your palette" and then they will give examples. Notice how the red tones in this example give this a warmer feeling than this next version with blues which feels colder?

    And I think umm... no... all I get from it is one is red and one is blue. And they go into red can be danger, anger and passion (and I think make up your mind because those things are all very different to me) and blue can be wholesome, calm and masculine etc. I can kind of see red for anger, yellow for enlightenment and so forth. I get those things at a technical level. But never at a "feeling" / mood level.

    This makes me think do other people actually feel something when they look at color in art? Like they actually feel something just from the colors alone automatically? Or do they just think "red, orange... hmm kind of like fire... warm"?

    The latter is more how I can relate to colors as far as them having some kind of meaning in and of themselves. In that way I can connect as far as "blue... cold" but I also think "blue... sky" and "blue... water" and for games "blue... ah mana! red... health!" I'd use red for health potions and blue for mana potions simply because gamers are used to that. They've been "trained" by games. I'd use blues and whites for wintery levels simply because again that is what games have done for decades.

    I need to ask some people out in the world... just "normal" people not in game dev, not programmers, not artists just people and see what they think. Maybe show a picture of a scene in red monochrome and ask them what they get from it and then show them the same image in blue monochrome and ask them what they get from that.

    Sorry for the ramble just thinking out loud.
     
    Last edited: Jun 10, 2017
  24. zenGarden

    zenGarden

    Joined:
    Mar 30, 2013
    Posts:
    4,538
    @GarBenjamin
    Learn Photogrammetry and you no more have to make models or textures
    www.youtube.com/watch?v=JVig3WTcV7A



    This could feet for some exploration and puzzle game.
    Buying some houses assets or some animated characters and you could start coding to make an action game without bothering about 3D art.
     
    GarBenjamin likes this.
  25. neginfinity

    neginfinity

    Joined:
    Jan 27, 2013
    Posts:
    6,434
    Haven't seen it. It wouldn't suprise me if some $1000+ software had this feature, but I honestly never heard of anyone mentioning it. As far as I know the only software that (possibly) replicated the tech was two PS2 games and then that's it.

    The goal of art is to tell a story. Or, if we simplify it further... to convey an idea, a mood, or a symbol.
    There's always something. "The mood you want to convey" is what people are supposed to feel while looking at a picture/model they made. If they aren't supposed to feel anything at all, then you have a blank canvas.Same applies to music, by the way.

    My advice would be to temporarily shut down your inner programmer. The very logical part of the brain which wants everything to be precise, perfect, reasonable and sorted according to order. This way of thinking seriously gets in the way when trying to draw or model something, and horrifically slows you down. Turn that part of the mind off when you're dealing with artwork.

    One book that allows to temporarily tap into into the proper thinking mode is Betty Edward's "Drawing on the Right side of the brain". You could try checking it out.

    The point is that this kind of thinking:
    Is not quite right, because it sounds like you want a strictly defined meaning for color.
    In context of drawing, IMO, a color has a very large array of associated meanings. Not strictly defined. Kinda like a word that means ten million different things at once. It is an extremely "fuzzy" kind of meaning. Of course, there's also the whole color theory with primary and complimentary colors as well.

    Looking at red/black monochrome scene for a long time will be tiring, and generally you shouldn't do something like that. I believe there was a terminator game a long time ago, which did something like that, and that was pretty much unplayable. Greeen black would be alright, though.
     
    Ryiah and GarBenjamin like this.
  26. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    @zenGarden after several days of researching and then all night testing different modeling software I am pretty much burnt out on that stuff again. ha ha

    Thanks for that though. I think I am basically to I will either use cubes and spheres and so forth OR just buy existing ultra low poly models OR hire an artist to create ultra low poly models. I could create ultra low poly models but even that is just more time than I want to spend on such a thing. I'm just one person doing this very part-time and I don't want to be spending the bulk of my time just on graphics instead of being able to work on an actual game (whenever I decide to work on an actual game again) every time I want to make a game.

    So... maybe I will buy a bunch of low poly assets. I don't mind spending $300 to $400 to get a decent asset library going. And that is a quick & easy solution. I was just really hoping I could find something that would make it easier not just for me but for everyone. I do believe it is possible and I think there are solutions but they seem to be high dollar stuff targeted at company level instead of individuals.

    Another solution would be just use a dang cube and the only work is to update the texture for each different type of character and object. Well and define various animations for it. But obviously it would be very simple to do some even fairly cool looking animations. I might go with something like that. I am a big believer in less is more and trying to maximize what you can get out of something.

    You know even just the lowly cube could be rotated and stand on a point and then rotated and now we have this pointy spinning enemy / obstacle. Some cubes can rise up from the floor. Other cubes can drop down from the ceiling. Some cubes can move around jumping from spot to spot, etc. I think behavior is as strong if not more so than the actual look of something. And this is another reason why I don't feel it is justified to spend so much time creating graphics. I think it is more valuable to spend that time bringing them to life in interesting ways. It is still "art"... still expression... and still visual communication... just achieved through a different route. Done with more of a focus on gameplay, interacting with the player.
     
    Last edited: Jun 11, 2017
  27. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    @neginfinity thanks for that and I definitely understand what you are saying. That really is a big part of it. I have to mentally switch gears from programmer mode to artist mode. It is a conscious thing. I do that by doodling for several minutes. Or for modeling maybe spending 5 minutes just modeling whatever just random nothingness really. And that gets my mind to switch.

    I actually bought that book long ago and went through it or at least a good part of it. I still remember that onion drawing exercise to this day. And basically that book was all about one thing in my opinion... art takes time. You have to really slow down and learn how to see what is really there. See the lines and forms that make up something. And of course that is at odds with my goal which is to not slow down but to speed up. Not take a lot of time but to spend very little time.

    I found your reply very interesting and beneficial. Thanks! I can tell that you being a programmer can relate at least to some degree.
     
  28. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    4,378
    I still don't understand, with all of what you are saying, why you don't just do teh art through code? you would have so much more possibility and not have to deal with external soft ... And all those tut you could actually implement the rules and not care about them at all ... It's like it's perfect for you ...
     
    GarBenjamin likes this.
  29. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    I will have a look at it. It comes down to.... how much work is it going to be? How much time is it going to take?

    Compared to just using a cube and / or other primitives and focusing on animations and behavior to make the visuals interesting. Compared to searching for and finding a decent set of models that a game idea can be modified to fit.
     
  30. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    I went off researching it and ended up bouncing all over and in the process entered back into the voxel stuff and found one I hadn't found over the past few days.

    Qubicle and it might actually be worth checking out. Maybe. It is nice to see that it was used to make some Indie games but that doesn't mean anything because a lot of people would spend months working in Blender or anything else and do the same. BUT... one thing that is interesting is the 2D games in the video. I wouldn't have known they were actually voxel models. Kind of interesting.



    I'll probably blow the $20 and give it a try.

    Well never mind I guess maybe I won't considering this person seems to use it a lot and still it took them well over 1 hour to make this dragon.


    That's just way too long. So figure a person spends maybe 4 to 5 hours per week on this stuff. They could make maybe 4 to 5 models once they are experienced at it. Probably 1 to 2 models like this per week in the beginning. So yeah maybe 5 models counting real simple stuff (crates, rocks, collectibles) And if even a tiny game needed like 100 models total that would be like 20 weeks just creating the models. Haven't even done the animating of the characters yet. So that would be what maybe another 6 months? Something like that.

    I think this just makes it very clear... cubes and other primitives it is! :)

    If I do ever come across something that truly is awesome and actually eliminates some of the work, has a much improved workflow, etc I will definitely let you all know. But I think short of one of us spending the time to build such a thing it might not happen. I think most people are fine spending weeks (even months & years) just making art because they really enjoy it. I actually do too but not when I am trying to build a game. Then it just gets in the way.

    Anyway... by the same token if you ever find a software that does make the modeling much faster share it please (I wouldn't be terribly surprised if some folks already know of and use such a thing but being smart they are keeping it to themselves lol). :)
     
    Last edited: Jun 11, 2017
  31. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    Well decided to check Steam and what do you know...

    There is a software 3DF Zephyr Lite on there that has great reviews for producing 3D objects out of photos.

    This might be a great way to get environment stuff. Like grass, trees, rocks and so forth. I think it would be more helpful to people who are striving to reach a very high graphics fidelity. @ShadowK and @Billy4184 might find it helpful. I don't think it would help with characters and certainly not in making cool dragons and such because at least around here I haven't seen one in a long time. But still if a person is after realistic environment probably very worthwhile especially if you get it on the Steam sale for $60 like this fella did.
     
  32. zenGarden

    zenGarden

    Joined:
    Mar 30, 2013
    Posts:
    4,538
    This is not related to Qubicle but to the 3D engine, you could do the same using Magica.
    3D engines can render Orthographic camera, you just need to create and use voxel colored planes, there is no difference between using voxel planes or using pixel images you still have to paint something.

    I think this is definitively the way for you, there are very cheap assets 2D or 3D that would suit your low poly requirement.
    1) you won't have to bother about making 2D or 3D art
    2) you won't spend time creating art, only coding
    3) you won't get frustration and directly get what assets you like and assemble them to make a game, you jump on what you like directly that is coding

    There are some very cheap ones :
    a retro Knight
    https://www.assetstore.unity3d.com/en/#!/content/39419
    Voxel characters pack
    https://www.assetstore.unity3d.com/en/#!/content/80522
    Simple voxel people pack
    https://www.assetstore.unity3d.com/en/#!/content/82661
    Town
    https://www.assetstore.unity3d.com/en/#!/content/17713
    Cartoon Biomes
    https://www.assetstore.unity3d.com/en/#!/content/29560
    Polygon adventure pack (non animated characters)
    https://www.assetstore.unity3d.com/en/#!/content/80585

    If you dig you'll find more, as they are assets you can use them with any Api, it's not restricted to Unity, so you should find what you need. You have no more escuses to not make something when you'll be again in mood of game dev :D
     
    GarBenjamin likes this.
  33. neginfinity

    neginfinity

    Joined:
    Jan 27, 2013
    Posts:
    6,434
    "If you want to have a dragon in your game... find a real dragon and scan it."

    There are several software packages of this kind (Zephyr, Agisoft, 123d catch, etc), but they have flaws.

    Issues:
    1. You need to have the object.
    2. The object is probably copyrighted, trademarked, patented or both.
    3. If you scan stuff as is, the lighting ends up being baked in.

    #3 is a big problem, actually. Basically, you'll need a lightbox that produces "directionless flat lighting" which is ridiculously hard to achieve in real world. And after that you'll still be missing normalmapping information.
     
    GarBenjamin and Martin_H like this.
  34. APSchmidt

    APSchmidt

    Joined:
    Aug 8, 2016
    Posts:
    2,494
    My view is that Unity gives the false impression that making games is easy; this is why so many people do not manage to make anything with it, as soon as they understand that making games is not easy at all and requires lots of time, dedication and learning. :)
     
    GarBenjamin, Martin_H and neginfinity like this.
  35. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    4,378
    That's exactly what I'm talking about, primitive are downright simple to do in code, so much even me have been doing that extensively ... given I'm only a 10th of the programmer you are, I'm sure you could do even more complex thing for literrally free as there is a lot of resource to do great stuff like house, organic creatures, tree on a low fi budget (generally the most complex stuff ask for a kind of convex hull implementation and some IK, once it's code, it's done). A basic code based parametric system seems to fit you! And it even work for animation.

    Here is a simple sin() based walk cycle:


    Here is another programmatic animation:


    And system coded can be generalized beyond the model, which mean that's instant win. The last video have simple coded curve function for the bounce, the same use for GUI and many other unrelated things.
     
    GarBenjamin likes this.
  36. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    4,378
    And apparently even ubisoft is investing in this old techniques (all the test here as been made by one single programmer on spare time as proof of concept, so no heavy cost involved).

    It's not new as Ken Perlin did some experiment back in the day too.
     
    GarBenjamin likes this.
  37. neginfinity

    neginfinity

    Joined:
    Jan 27, 2013
    Posts:
    6,434
    See the sliders? Let's assume that each slider can only have 10 distinct positions. 16 sliders total.

    Meaning.... 10^16 possible combination of parameters. So, instead of spending 2 hours learning (once) how to animate a walk cycle, you'll spend next two days tinkering with the parameters only to produce unusable animation. That's how experimental tech works in general.

    The problem with even something like walk is that there's more to it than it seem at glance.
    Balance. Gait. Expressing character mood and personality. And suddenly stuff that started with "just a simple sin() function' ends up being a several dissertations worth of material.

    Meanwhile, using old modeling method character walk at a bare minimum is defined by a total of two keyframes and mirrored variations of them.

    Code is fine and all, but it is important to recognize situation when you're running into an "amazingly interesting problem (tm)", which will waste a lot of your time and produce nothing useful.
     
    GarBenjamin and Deleted User like this.
  38. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    4,378
    @neginfinity

    I love you LOL :D

    1. you have proven that's one order of magnitude LESS parameter than doing by hand. using such number are meaningless.

    2. Its not experimental, it's now rather common BUT mixed with plenty animation, see the second video (which us way less parameter). It's basically a simple IK that follow a transform. And you can set up the transform using the two keyframe method, keyframe are basically just positions, which can be inputted by code. And to be frank they are still mostly blending.

    3. We are talking @GarBenjamin , he said he didn't care about mood, and those are acceptable animation for someone who don't care about mood. EVEN considering that basic mood is easy to achieve using simple transform. We are talking low poly art, maybe cube, those animations would be more than enough. If minimal graphics, why not minimal animation style?

    4. It's not so much a super interesting problem, It's been used since the day of the super fx on snes or even before. AND it's reusable, once you have done it for one character, it's generalization quite easily for no effort (have a limb system). The video wasn't meant to tell DO EXACTLY THAT, it was meant to say there is some possibility doing something good with just code. In the case of the cube, using the bouncing animation of the sphere (without the limb) is a sufficient example.

    5. Anyway I must resist the bait and stay focus on my work instead of having a diversion to code an example myself :oops:
     
  39. zenGarden

    zenGarden

    Joined:
    Mar 30, 2013
    Posts:
    4,538
    @GarBenjamin
    Something like Spore creatures is what you need lol

    It's 8 years old system and it was ported to Nintendo DS also.

    Or better "Dream" game for PS4 , you can create anything
     
    GarBenjamin likes this.
  40. neginfinity

    neginfinity

    Joined:
    Jan 27, 2013
    Posts:
    6,434
    Here's the problem;
    With hand animation you can do absolutely anything. You also have estimated 1 exaflop of computational power at your disposal to spot deviations of movement compared to patterns observed in reality.

    With procedural animation... you're constrained by the model used to drive it. The model is dumb as a brick, and cannot spot errors. If the model is incorrect, then "whoops". And THAT's the problem.

    The trouble with procedural animation is that the programmer will attempt to derive a correct model for driving the animations, miss bunch of parameters, and after wasting XYZ hours, results will look bad. Issues like that may explain why overgrowth spent so many years in development (development started in 2008, reached beta in 2017)

    If you're talking about exporting IK targets and then using them to determine limb placement in the world, then it is not exactly noteworthy. As for video, I had enough distractions in past two days, so right now I'll pass.

    Mood in art is one thing. Lack of mood in character will lead you to wooden movement, if you're lucky, or right into uncanny valley, if you're not.

    Minimal animation style is what I described. Walking loop based off two keyframes and their mirrored versions. Can't got lower than that. Trying to go procedural might kill the style, but it depends on what you're after.

    --edit--

    In the end, it is important to decide what your priorities are. If the idea is to make a game, then working on a procedural animation will be a dangerous distraction, which will most likely waste your time for no gain. If the idea is to have fun coding, sure, have fun with that.

    It is also necessary to figure out if the prolblem is even worth being automated. Walk cycles are well documented and well explained. Now, try to make an animated character do a martial arts move, and then it'll get much more complex.
     
    Last edited: Jun 11, 2017
  41. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    @neoshaman the programming approach to animation does look interesting. That's the kind of thing I've thought about doing. Only more of a straightforward thing like click set keyframe (model in just a default post).... select and rotate/translate the parts as desired click set keyframe to then store deltas for those rotations and translations. Really it is not too far from what any model animation program has except it would an interface that is more streamlined for me.

    @zenGarden something like the Spore creature builder would indeed be quite fantastic. With a massive categorized library one could create some very interesting things.

    I think both of these kind of approaches would be very helpful to a lot of people. For me they still represent more time & work than I care to do. At least for now. All of this constant clicking & dragging just seems like such an inefficient way to go. But I guess until we get some kind of helmet that can provide a more direct link between what is in our head and getting it done on the computer there is probably no way around this busy work.

    When I talk about using a cube (and other primitives) I am no longer thinking about these as parts of more complex object. I am thinking of them as being the object in its entirety. I don't mind spending the time to scale the cube to make tall objects, short objects, narrow objects, wide objects and so forth. But that is about the limit of how much time I want to spend messing around with this stuff. Other than when I am in the code animating these primitives to bring them to life.

    I think the next time I am in a mood to spend any time on this stuff I will mess around with it. See how many distinct behaviors I can come up with for something like a simple cube.
     
  42. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    4,378
    @neginfinity
    I understand were you come from, but frak! those argument trigger my inner brain so hard. I will reveal to you, I'm actually evaluating limitation and putting down spec for a digital puppetry/cinematography tool for my project, I mean that's the task I was talking about, so you will see attempts as soon as I finished it, we will see how many crow I will eat lol. Though I'm well aware of limitation I don't aim to do everything. I have identified some key parameters( feets, hip, torso, head, eyes target, wrist, hand poses, facial "poses") to drive through a hierarchical macro system.

    BUT what I'm proposing here is not even on the same level, Because of the next point ...

    You are simply placing the bar too high for the context of this discussion. This happen even in AAA games (mortal kombat x anyone? or mass effect andromeda) and are you seriously expecting than an untrained programmer to do anything more than wooden animation?



    I'm not arguing for him to implement principle of good animation or do complex animation anyway.

    Well no wonder the discussion is going place lol

    I'm not advocating for state of art coding doing state of the art animation. Simple old tricks that do teh works. We have been applying sin to cube to mock walking for a very long time. That's head bobing, that's swaying weapon, etc ...
    Even Basic facial is not state of the art anymore, Perlin and Ekman gave us the FACS mod and even 12 years kid do it with game maker now.

    I'm no so sure about that, depending of the complexity, you can do many of them in a wooden style ala mortal combat quite easily. A punch is moving the hand and let the arm adjust, a flip is rotating the character, now proper posing is out of reach of beginner so it might be enough. I remember playing sf4 and be livid at ryu's jump being a simple linear rotation of the model.

    I mean I'm trying to think of what basic generic game movement would be so hard for a simple system. It just won't be the most artistic game.

    We aren't talking AAA level here.
     
  43. GarBenjamin

    GarBenjamin

    Joined:
    Dec 26, 2013
    Posts:
    7,442
    I haven't looked into any more graphics creation tools since my last posts here. I think I'll actually put my thoughts into practice.

    Using very basic cube based graphics and either AGK or C# & Irrlicht Lime (probably AGK because then I don't need to spend any time setting up an IrrlichtLime template project and can instead focus on learning the AGK 3D api).

    See how much I can EASILY bring them to life and see if it is truly much faster for me to work purely in code as I think it is. Of course first project will be super simple tiny.

    At least I do feel motivated to do some game dev again. :)
     
    neoshaman likes this.
  44. Siccity

    Siccity

    Joined:
    Dec 7, 2013
    Posts:
    143
    I haven't read the previous 15 pages of comments, but to answer the title; no.
    If anything, Unity makes game dev too easy. This is partly why Unity is infamous for spamming the world with low effort assetflips. Anyone can make a game, but making a good game will always take a lot of creativity, discipline and skill.
     
    TenKHoursDev likes this.
  45. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    4,378
    I haven't read the rest of the post
     
    MD_Reptile likes this.
  46. Murgilod

    Murgilod

    Joined:
    Nov 12, 2013
    Posts:
    5,969
    You haven't read the previous 15 pages of comments so you just decided to necro a year and a half old thread while adding nothing to it?
     
  47. Lurking-Ninja

    Lurking-Ninja

    Joined:
    Jan 20, 2015
    Posts:
    4,570
    Reading the previous 15 pages of comments requires discipline and skill. And sometimes creativity. Forming a good answer certainly does.
     
    neoshaman likes this.
  48. TenKHoursDev

    TenKHoursDev

    Joined:
    Nov 9, 2014
    Posts:
    1,089
    Calm down guys, the guy obviously saw a topic they thought was interesting and wanted to contribute. Necro posts may be a sin, but they're not a crime. :p
     
    Siccity and MD_Reptile like this.
  49. APSchmidt

    APSchmidt

    Joined:
    Aug 8, 2016
    Posts:
    2,494
    Could be fun too. :)

    I sometimes necro threads too; on another forum I roam a message box pops up when you necro a thread, politely asking if you really want to continue or create a new thread. An example to follow, I guess... ;)

    Edit: as for the original question "Does Unity make game dev too difficult?" I'd ask: "How difficult making games would be if Unity didn't exist?"
     
  50. Arowx

    Arowx

    Joined:
    Nov 12, 2009
    Posts:
    7,133
    I tried this with Doom and made the first level in Unity from scratch in about 10 days .



    I was surprised to hit problems with lifts and navmesh AI.

    However the largest issue in this type of game is the use of a level editor and loading saving levels to files, this is something games always used to do, from tile based games to 3d games. Storing a level as a set of reusable elements or assets and then as a single file that loads them and lays them out for that level.

    For instance Wolfenstein 3D has about 6 episodes of 10 levels or 60+ levels in the game.

    The issue here is Unity uses Scenes which have the problem of then the developer having to clone their code and UI between multiple scenes and figure out how to pass data between scenes.

    Why doesn't Unity have a root Scene system where the core game logic can live without lots of DontDestroyOnLoad logic to get around the issue. Or the data hack of storing information between scenes in the PlayerPrefs data store.

    Hopefully multi-scene loading should get around this problem now with root scenes being able to be stored throughout the game session.

    Although having levels as just lightweight files that load in prefabs is difficult when you add occlusion culling, navmeshes and global illumination all static systems that need a static scene for a level.

    I think hiccups like this in the way the game engine works are the first set of problems developers who want to make a game with more than one level encounter in Unity, alongside the lack of navmesh support for lifts/elevators.

    Or imagine lighting 60+ levels with GI in Unity the developers are probably still waiting for the GI to bake...
     
    Last edited: Nov 19, 2018
    GarBenjamin likes this.
Thread Status:
Not open for further replies.