AuthorTopic: GR#007 Yeti Project - Sprite Development  (Read 33626 times)

Offline EvilEye

  • 0010
  • *
  • Posts: 497
  • Karma: +1/-0
  • Game Developer Extroaordinaire
    • View Profile

GR#007 Yeti Project - Sprite Development

on: July 31, 2009, 06:08:17 pm
AKA: Why I don't do this for a living.

For those who don't know there was a dispute between me and a few others on this forum as to how long it would take someone to pixel a particular piece of artwork. They said an hour, I say, not a chance.

Here is the sketch at about 400 x 500 resolution:



The goal was to pixel it at about 120 x 140 resolution.

Here it is shrunk down to 118 x 138 and cleaned a bit:



The pixeled version would have to look better then that. Otherwise, why would you bother?

So I intend to demonstrate exactly why high-quality pixeling takes a long time and there really isn't much you can do about it.

And the first thing we start with is the lines........

In the beginning, there was the sketch:



Now the first thing you should know is this sketch has some flaws, and some of those flaws are postural. The creature ( based off a gorilla ) is slightly off balance and the posture is a bit unrealistic. This is because I didn't use much reference when making it ( because I couldn't find gorilla pictures with the right pose ). The flaws aren't very noticeable at first because the coloring and outlining draws your attention away from most of it. But when you do an outline of it all of the sudden they jump out at you, and you have to decide what to do about them. Leave them there or try and correct them ( potentially a HUGE time waster when doing this at the pixel level ).

Now how I usually start off is I shrink my sketch down and then outline it in red. I use red so I can see what I'm doing.

Also since this is a quality piece I don't just outline willy-nilly. Instead I look at the sketch and try to capture and bring out the strong points, and reduce or fix the weak points.



Total pixel time about 3 minutes so far.

And I just remembered about mirroring the image. This is something you want to do every so often ( and you will hear this in art class also ). Sometimes you will miss flaws when looking at an image only from one angle, so mirroring the image helps your brain see what you missed before. Thats why the images above are reversed. I forgot to re-mirror them :lol:

But it's not all cut-and-dry when outlining. You get a lot of things like this here:



What looked to be a clear line with his bicep is really a mess of many different colored pixels. So where do you put the line? Well, that's a good question. There is room for interpretation here, but you must still keep in mind the overall picture. It's at this stage where we have our first chance to make the picture better.

Here is what I did:



I put a few pixels in key places where I thought I would definitely want a line, and left the rest open. It's not to say I won't fill the gaps in later, but for now I have the key parts mapped out, and later when the picture is more mature, I can connect the loose ends.

After about 12 minutes heres what I have:



This is where I start noticing anatomy problems, especially with the feet, which are wrong but not hard to correct later.

You will also notice the result is starting to look different then the original sketch. This is unavoidable, and it's actually one of the reasons that if something looks good when I shrink and clean it, I don't bother pixeling it.

I am up to about 15 minutes pixeling so far.

Now at this point some of you may be asking "why outline at all?". Why not just throw shapes of color on the canvas and add the outlines later.

Well, because lines are a guide for your work. Without them its easy to go wildly off track. Even though my lines are not colored now, I still have a rough idea of what I want in the future. Thats their purpose, to hold down a flash of inspiration while you work to bring everything else up to speed. Eventually that flash of inspiration and your progress will meet somewhere. Besides, I have found from experience that its much easier to correct a few lines then a blob of color. The less pixels you have to rearrange, the better.

------

Took about 5 minutes to fix the feet. They were bugging the hell out of me.



20 mins.

Maybe I should take this chance to preach about why its important to start with good lines. Good lines will save you time when coloring. Bad lines will make you unsure of where to put color and you will waste lots of time reworking your original shape. I would say good lines are as thin and efficient as possible. Thicker lines are ok when used to emphasize shadow and draw the viewers attention.

---

This is around the best time to make any major changes that you think need to be made. Once we start coloring changes will be much more time intensive and frustrating.

Usually I start seeing improvements I could make just by looking at the lines.

For instance I saw a way I could make the posture a bit more believable and add bulk up the muscles:



This may be hard for other people to see since its so rough, but I made the hand on the left touch the ground and the hand on the right is open ( very rough I know ).

I lost track but I am guessing this took me about 10 minutes of playing around.

So I am up to 30 mins.

---

I decided to edit the rough parts a little more since it probably wasn't easy to see where I was headed with this:



I say was because there is a little problem here.

The arm is now way too short compared with the leg. This means either we have to lengthen the arm or shorten the leg. So I am going to scrap that idea entirely since this post is mainly about the time it takes to pixel something and not about how long I take to screw around with the line art.

So lets shave those 10 minutes right off and pretend that never happened. But if I was doing this normally I probably would have kept the edit and shortened the legs.

So we are back to what we had, and lets mirror it just to make sure we haven't missed any major flaws:



Again, way off balance but we can cover most of that up later.

And were back to 20 minutes.

---

Now is the part where I start tightening everything up and adding "detail markers".



Basically what I did was flesh out the detail and close the lines on the outside. Adding little marks for detail is an important part so later I remember what look I am going for and where / how to place color.

Of course they are just little lines and can mean anything to someone else, but for me when I see them I know exactly what I wanted to remind myself of later.

Note at this point I am no longer tracing the sketch, its all freehand.

That took at least 25 minutes. There is no way to make it faster, sorry. Not if you want quality.

Were up to 45 minutes.

--

Moved the feet and tightened the lining just a notch.



I could go a lot farther with the lines. I should probably better define the face, toes, fur and even thicken some of them. Again, to some people this may seem like overkill. And even though I will lose many of these lines later they are not wasted, they are being used to capture and solidify my ideas. This is an important part of any great work of art, since even the best artist cannot just throw his inspiration at his medium and have it turn out like he wanted. A great artist doesn't know exactly what he wants, he just has a rough idea. It's the process of making the art where the rough idea forms into something solid.

But that was the easy part. Now comes the real pain. If only you knew the sort of hell we're in for now!

I think Bill Murray said it best: http://www.youtube.com/watch?v=O3ZOKDmorj0&feature=related

----

Coloring!



There is my first step, pick a base color and fill. I picked a lighter color because I actually want the fur to be white. In the sketch it's blue-ish. I am going to make the face blue and the fur white, since I think that will look better. Actually now that I think of it that still may be too dark. However its easy to change since were working with a palette. Just don't change it too late.

Took me a minute.

Lets restart the clock from here, since some people seem to think lines don't count. 1 minute.

--

I almost forgot to add the blue for the skin.



You may notice I am ignoring the left foot in the background. That's one of those things where I am just not sure what I want to do, so I will put it off for a while.

And now I need a rough idea of the light source so I start adding some basic highlights.



However, its not just about where the light is hitting. It's also a matter of where I want the observers attention to go. I want the observers main focus to follow that arm down to the hand. In fact the hand should be the part that sticks out the most. Right now it is not, so I have to do something about that later.

The above took me about 15 minutes. So were up to 16 minutes now.

--

Added the some rough shadow:



Actually I got a little carried away and did some detail work, which I shouldn't have done because I will just have to pixel over it later :blind:

It would be very tempting at this stage to go with the current look and finish it. That's what a good artist would do. But a great artist would hold off and wait.

I would say an indication your lines are good is when you have trouble resisting the urge to color and detail them.

Oh, by the way, that took me 30 minutes.

So were already up to 46 minutes.

---

Right now I am mostly dinking around with things, trying to decide how I want to tackle that fur.

I could always do something like this:



But that is not quite what I originally had in mind. In fact I may not be able to do what I originally had in mind. In which case, this would be a good fallback option.

But what I really wanted to do was to imitate that "sheen" look I got with the sketch.

---

This is probably going to piss off the anime enthusiasts out there. Tell you the truth, I hate to do it myself. But I believe in the long term its the right move.

I decided I need thicker lines and some ultra dark shadows:



Probably about time to mirror it too.

Still off balance, but its less noticeable now that the detail is catching your attention more.

Took about 15 minutes.

1 hour, 1 minute. And the picture is about half done. I said the picture is half done, but the time it takes will increase exponentially from here.

---

Are you ready for some pain?

http://i174.photobucket.com/albums/w104/chris_6713/realpain.jpg

Well get ready, because your about to get your first taste of real pain.

Turns out we have 3 trouble spots:



The knuckles, the foot and the face. They are all poor looking and fixing them will cause many a headache. So grab a bottle of asparin and lets get started.

Might as well start with the face.

---

The problem with the face is simply, we're trying to fit this:



Into this:



And it doesn't work very well as you can see.

This is the hard part of pixeling. Because this problem has no right answer, just a whole lot of wrong answers. Even for someone like me who has been pixeling on and off for 15 years, I still have no idea what to do here. Sad isn't it? >:(

Usually what I do is I say to myself "what here do I really want to keep and what could I live without".

Some of this detail must go, because it is hogging space that could be used to better define other detail.

---

Soooo... I want to keep the eyebrows, in fact I want to make them white eyebrows. I want to have a defined upper mouth. I want to keep the eye shadows. The cheekbones. The tongue. The teeth. And I kind of like the original angle, but that may not be possible.

Here is what I got after 6 minutes:



Cleaner, but now it looks like the Grinch. We will have to readjust it a bit so it looks more like an ape.

1 hour 7 minutes.

----

ARGH!!!!!!!!!!!!!!!!!!



Too happy.



Too weird



Yikes!

Well, I wasted a good 20 minutes to go through a bunch of arrangements that didn't work ( most of which I didn't show here ).

Lets go back to the old version and see if I can make that work somehow.

----

Closer:



10 minutes.

That's a little better. So it seems we should stick with improving the old version.

---



Well now we're actually getting somewhere. 16 minutes give or take.

There's another problem though, this is really too detailed for the piece, it should be simpler and look better from a distance.

1 hour and 53 minutes.

---

And.... Voila!



10 minutes.

There it is. At least for now anyway. It's like putting together a puzzle, once you see the picture start to form, its a breeze from there.

Now I know some of you are going to think I just wasted nearly an hour.

But compare:



That's the difference between ok and spectacular. And it's a big difference.

We're probably not totally done with this face yet since its one of the focal points of this piece. We might have some touch-ups when we are nearly finished. Also if you noticed we changed the slant of the head. This might need to be tweaked later because now the pose looks a little lifeless.

2 hours and 3 minutes.

And there is so much yet to come....

---

Well I fixed that face again, because I knew having it straight like that would come back to haunt me later.

So I readjusted it:

New



Old



Why does this matter right now? Because you are constantly ( or should be ) zooming in and out looking at the whole work to get an idea of what you need to do. And when you see that straight face it will throw you're current pixeling off, and that will again throw your later pixeling off. Eventually it avalanches to destroy your whole work.

Sounds anal but it's true. You want to fix major problems sooner then later, because seeing and getting used to something that's wrong will mess you up every time.

My fault for not paying attention so we won't add any minutes for that.

*Think I just reached my limit for today, I'll be back tomorrow*

---

I feel like some pain today howabout you?

Lets do something about those feet.

( please be easy to fix....... )



NOPE! :yell: That's going to be a <female dog>.

A smart thing to do would be to look at some reference first. So I am thumbing through google images.

Seems a gorilla would probably never have his foots thumb like that, it would be spread open for balance. However that would look bad on this piece, so we will meet it halfway I think. I can get away with this because I am not making a gorilla, I am making a yeti ( a bigfoot type creature ).

I find a good way to rework stuff like this is to just erase all the detail and make a basic shape.



If that basic shape looks promising we can move in and add more detail. If not we try another basic shape.

Howabout something like this:



Looks reasonable enough right?

Well, looks can be deceiving, this may not work at all, we just have to take it further to find out.

All this is taking around 7 minutes total btw.

---

But before I got any farther, I hit another snag.

Almost all the lines in the picture pull down toward the back ( as of yet unseen ) foot. This is an important effect that we don't want to ruin.



But having the foot like that will kill it for sure. So we can't do it that way.

I suppose I should have caught this earlier but sometimes you just never know something is a mistake until you do it.

You see in many ways we've boxed ourselves into a corner. We need that foot to spread out so it offsets the bad posture of our original sketch. But at the same time if we do that we harm one of the major draws of our picture, which is the way the whole shape seems to pull down toward that back foot.

We will probably have to compromise with something like this:



You see how the toes gravitate toward the other lines? The outside toe lends a little more balance to the picture, while the other toes try and draw your attention back in.

---

As I said before this is not really about skill, its mostly about luck. Will I happen on the right solution?

It's like playing a slot machine. Minutes are quarters and the jackpot is the right look.

C'mon baby, Evil needs a new foot!



..
..
..
..

Jackpot?!



*going to take a few hours for lunch. I can't believe it has taken this long and I still haven't really gotten anywhere*

---

Well I'm back and it turns out that wasn't the right one. The angle is pointing too far down. So we got some more of this hell.


no

no

*A LOOONG TIME LATER, I won't even bother to post all the attempts*

I think I got it:



Ya, that will work.



Heres the old foot:



----

In case noone else can tell, he's walking on his knuckles. Or at least that is what it is supposed to look like. Then his ( foot ) thumb is propped on the ground like a stand. I hope that's what everyone else is seeing anyway.

I am going to leave that un-detailed for now and finish it up when I have the rest of the piece done.

It took longer, but lets say that whole ordeal took 30 minutes.

We're up to 2 hours and 33 minutes.

----

The good news is that there's only one more problem spot to go. The bad news is the real hard part is up next.

Lets fix those front knuckles. Here is what we have to work with:



This shouldn't be too bad actually. Knuckles are pretty simple. Nothing like toes :yell:

---

Did I say it was going to be easy? What I meant to say is it should be easy. Sometimes these seemingly simple parts can surprise you.



About 5 minutes so far.

---

ARGH!!



No, No, No, No, No*Feeling the pain yet?*

This is taking me way longer then it should, and that usually means I'm hitting burnout. I'll be back tomorrow evening if not sooner.

Wasted a good 20+ minutes right here.

2 hours and 58 minutes.

---

I'm Back. Sorry, had to do a lot of real life stuff today.

I see some people have posted opinions stating I am wasting a lot of time needlessly. Well we'll see about that when it's all finished.

There was something throwing my eye off about this whole knuckle thing. It's that the forearm is too long and too far to the right.



So I am going to move the arm to the left a few pixels and that should help.

---

Or, maybe not:



I think I just may be stuck with that bad look unless I were to rearrange that whole arm, which I am not going to do. Plus some people are jumping on my back now about needlessly wasting time editing things.

We'll say that whole thing just took the original 20 minutes, which means were still at 2 hours and 58 minutes.

---

So lets compare shall we?
VS

Not looking so tough now is he?

Except he has something we don't, a nice shiny coat of fur.

But we're about to change that.

----

“God is a comedian, playing to an audience too afraid to laugh.” - Voltaire

It's here.

This is it. The part I was dreading.

The entire reason I did not want to pixel this piece in the first place.

No doubt many a pixel artist has died trying what I am about to attempt.

Yes, that's right. The fur.

But without this part, we will never beat that resized version.

This is where I pull out all the stops and throw everything I have at this piece and hope something works.

----

It's not that I can't do fur, don't get me wrong. I can do great fur, hair, wood and pretty much any texture. But to beat that resized version I will have to do spectacular fur.

I have a few good ideas, but will they work at that size / angle is the real question.

---

But you don't go at something like this all scatter-shot and just try stuff randomly. We have to have a plan. Then we have to get a backup plan. Then we have to get a backup backup plan.

I sketched a few ideas to see how they would look:

The sleek look:



A yeti with a 1950's hair-grease job. Basically it's what we have in the original sketch. Not my favorite.

The rough look:



That may be impossible, but its kinda cool. Like the wind is going through his hair.

The fluffy look:



That's the look I would like to get, its a little calmer.

If all else fails we do this:



Actually that may not be a bad idea, but it would be too easy if we just did that wouldn't it?

*looks like I gotta continue this tomorrow guys, bed time*

---

I decided to start testing some techniques for feasibility.

Here we have the shiny look:



Actually I got a little farther but a STUPID *&%@ING LOCKUP in photshop forced me to hard boot my comp! I need to buy a new wacom so I can get a more recent version of photoshop. Still running elements 4.0.

Doesn't look too bad though huh?

Welcome to the world of hyper-pixeling folks. This ain't your grandmas pixel bingo. This is hardcore.

That took me about an hour right there, just for that.

When I finally settle on a style it will take me around 3-4 hours of straight pixeling to complete. And then after that there are always touchups.

*got some business today so I will have to come back later tonight*

---

Back a little late here. I have a lot to do this week so I am going to have to slim down a bit on my updates.

I might as well respond here to some misconceptions:

1. I am not saying all pixel art takes this long, just high quality pixel art with dynamic poses like this one.

2. I am not being indecisive with any of this, the medium just does not allow some things and you have to find a workaround. This is a difficult pose to do in pixels and there's going to be hangups.

3. my monitor is on default and I usually work on a bright pink background.


Now, I tried another look for the fur that almost worked... but not quite. Just not high enough resolution I think.



Still interesting. Hows that background? Better?

Oh ya and I expanded the palette a bit.

---

Update. I'm coloring now BTW.



I decided to go with a hybrid between the two test styles. We'll see how it turns out.

Oh ya, I messed with the arm again, sorry I had to do it.

I'll try to post my process later when I get the energy to type tons of stuff again.

---

Another update.



My guess is I'll finish sometime saturday. I'll post my process after I'm done. It's too distracting to do it while I'm trying to pixel.

--

Update. Slowly nibbling at it bit by bit ( or pixel by pixel ).



--

Update. We're getting to that stage where a lot of these improvements will be based on personal style and might look better or worse for some people.



---

Ran into an extreme bout of lazyness over the weekend so I hardly worked on this at all until today.

I decided to just have a mane and leave the fur on the rest of the body undefined. I just like the way the muscles came out too much to ruin it.

Here are a couple of tries at remaking the "mane":





I think I will go with the second one.


----

DONE!



Wow, that only took me about 3 weeks :yell:

Lets see if we beat the resized version:



Yep, no contest, totally blew it away. But at what cost? Took me less then an hour to do the sketch, and maybe 1 minute to shrink it down and clean it up.

But oh well.

Like I said before I work slow. It probably takes me about twice the time it would for someone else to make something.

My estimate is a real pro with some fast pixeling skills might have been able to make this same piece in 5 hours of pure pixeling. That's not counting things like taking a break or anything like that.

---

Not done yet apparently, helm gave me a good edit, so I've got to mess with this some more:



---

I'm really done this time. Not because it's perfect, but because I just can't spend anymore time on it.



Did a portrait for it too:



Not pixeled, but good enough.

Here's a screenshot:



« Last Edit: August 28, 2009, 02:59:36 am by EvilEye »

Offline bengo

  • 0011
  • **
  • Posts: 566
  • Karma: +0/-0
    • View Profile
    • My Art Tumblr

Re: [WIP] The Yeti project

Reply #1 on: July 31, 2009, 11:58:33 pm
Nice lineart, start coloring!

Offline Ryumaru

  • Moderator
  • 0100
  • *
  • Posts: 1670
  • Karma: +0/-0
  • to be animated soonly
    • ChrisPariano
    • View Profile

Re: [WIP] The Yeti project

Reply #2 on: August 01, 2009, 03:35:42 am
All this lineart stuff I don't believe was intended in the whole " 1 hour" thing. You remarked how long it would take to get a fur texture like the one in the resized version and the larger sketch in general. I already showed you an example of what I did in an hour, yeah it was different than your sketch ( purposefully) but it still shows that a reasonably finished sprite with a nice fur texture can be achieved in an hour.
with that said the lineart is obviously good- it is time to color. I would say one of the reasons your original attempt didn't capture the look you were going for is because the correct values were not placed where they needed to be.

Offline EvilEye

  • 0010
  • *
  • Posts: 497
  • Karma: +1/-0
  • Game Developer Extroaordinaire
    • View Profile

Re: [WIP] The Yeti project

Reply #3 on: August 01, 2009, 02:09:05 pm
All this lineart stuff I don't believe was intended in the whole " 1 hour" thing. You remarked how long it would take to get a fur texture like the one in the resized version and the larger sketch in general. I already showed you an example of what I did in an hour, yeah it was different than your sketch ( purposefully) but it still shows that a reasonably finished sprite with a nice fur texture can be achieved in an hour.


Like I told you before, its OK, but it doesn't beat the resized version. I will show you what that will take, and then you will wish you had never asked.

Quote
with that said the lineart is obviously good- it is time to color. I would say one of the reasons your original attempt didn't capture the look you were going for is because the correct values were not placed where they needed to be.

The reason my original attempt didn't capture the look I was going for was because I dipped my toes in the water with the first attempt and said "HELL NO!".

Offline Atnas

  • Moderator
  • 0100
  • *
  • Posts: 1068
  • Karma: +2/-0
  • very daijōbs
    • paintbread
    • paintbread
    • View Profile

Re: [WIP] The Yeti project

Reply #4 on: August 01, 2009, 02:28:49 pm
I was under the impression you would just be finishing up your current pixel wip when I stated it would take an hour to finish it. Not completely starting over. Anything you fix with the line art could have been fixed on top of the actual WIP, much faster because you are working with a level of rendering much closer to the final presentation.

Offline EvilEye

  • 0010
  • *
  • Posts: 497
  • Karma: +1/-0
  • Game Developer Extroaordinaire
    • View Profile

Re: [WIP] The Yeti project

Reply #5 on: August 01, 2009, 06:43:24 pm
I was under the impression you would just be finishing up your current pixel wip when I stated it would take an hour to finish it. Not completely starting over.

I won't even count the line time.

Quote
Anything you fix with the line art could have been fixed on top of the actual WIP, much faster because you are working with a level of rendering much closer to the final presentation.

Experience has taught me otherwise. Try editing colored line art and then try editing just lines. Editing less pixels = faster. Don't worry, well be editing the colored lines plenty :lol:

*edit* I meant to add I started with new lines because the old ones were kind of half-arsed. I just did them to get an idea of how hard it would be. If I am going to do a perfect ( or near perfect ) work here I have to start with good lines.
« Last Edit: August 01, 2009, 07:01:25 pm by EvilEye »

Offline Scribblette

  • 0010
  • *
  • Posts: 185
  • Karma: +0/-0
    • View Profile
    • Livejournal

Re: [WIP] The Yeti project

Reply #6 on: August 02, 2009, 02:54:18 am
Regardless of how horribly long it takes me to get pixel art done too, I'm hoping you'll turn out to be wrong on this, Evileye. It'd be wonderful if the other good artists hopped on here and demonstrated how they'd do what you're trying to do in less time - even if it's a little less 'spectacular'. I think everyone would appreciate any tips on how to save time pixelling. :)
Now reading: Animator's Survival Kit, Drawing On The Right Side Of The Brain, Fun With A Pencil. No time to pixel!
Pixelated Anatomy|Foliage

Offline Souly

  • 0011
  • **
  • Posts: 957
  • Karma: +0/-1
  • Killer of threads.
    • View Profile
    • Punkys Portfolio

Re: [WIP] The Yeti project

Reply #7 on: August 02, 2009, 11:28:03 pm

I was really liking this hand, the arm hanging back just seems odd, I don't know why.

Anyways all is looking good.  :y:

Offline Helm

  • Moderator
  • 0110
  • *
  • Posts: 5159
  • Karma: +0/-0
    • View Profile
    • Asides-Bsides

Re: [WIP] The Yeti project

Reply #8 on: August 03, 2009, 03:55:20 pm
Evileye you take longer than others (or I) would because you're undecided about things and your process is bogged down by this indecision and all the variable ways you're trying to achieve things (and anatomy problems, and a lack of comfort with abstraction). It's not a matter of pixelling being inherently slow(er), it's a matter of focus and a solid pipeline of getting things done. You don't need to relineart the lineart, you don't need to put cel-shading to something that will not eventually use cel-shading. It's enjoyable to watch you go through all this, but it doesn't mean your problems are inherent to the medium or insurmountable. You'll progress much faster if instead of trying to prove a weakness as a rule, you work against the indecisive parts of your process. I would say more and edit for clarification, but I don't have the time to do it.

Offline xhunterko

  • 0010
  • *
  • Posts: 362
  • Karma: +0/-0
    • View Profile

Re: [WIP] The Yeti project

Reply #9 on: August 03, 2009, 04:17:17 pm
I may be only one poor opinion on here, but it's what we're supposed to do right? I have to say, that I liked your shrunk down one better. But then again, it's not pixel art is it? However, I think that you might also be suffering from lack of natural reference. Well, you probably looked at something when you drew the sketch, but, have probably forgot what it is by now. I also know that yeti's are hard to find so this is a second best.



I also have to contrast with souly here and say I do not like either hands. They look messy and obviously not finished up. They seem a bit less structured then the back leg and maybe a bit out of proportion. It's kind of hard to tell what the hand is supposed to be doing, or if it's holding something. But maybe I don't know what I'm looking at here. I'm still a noob after all.