Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - Squiggly_P
Pages: 1 [2] 3 4 ... 7

Pixel Art / Re: [WIP]Trees related
« on: December 24, 2009, 02:20:07 am »
I think what you need is just some solid drawing study.  I know you're drawing a lot, but you're not trying to draw what you see, you're trying to draw the idea.

Trees are a really difficult subject in general, and for someone who still doesn't have a grasp on 3D space, perspective or a decent visual vocabulary, trying to draw trees is going to be a lot more frustrating.  You seem to be looking for some method that will take care of all of your future tree-drawing problems, but there is no hard and fast rule for drawing anything.  There's no "right way".  All the advice in the world won't matter if you can't use it properly.

I don't want to discourage.  I totally support you in your studies and practice.  I just think you would learn a lot faster if you remove as many things as possible and focus on one skill at a time, whether it be perspective, light and shadow, anatomy, positive / negative space and silhouette...  Trees basically force you to know a lot of stuff in order to be able to competently visualize one, unless you go for a really really stylized approach, and even then it's best to know as much as you can.

If you really are totally focused and determined to draw trees, then start studying trees.  SERIOUSLY study them.  Get a blank book or paper you can draw and take notes on, go outside (or onto google images, if it's too cold right now :P) and draw some trees from reference as closely to the reference as possible.  Don't stylize anything.  Draw whole trees, draw what the individual leaves look like, draw closeups of the bark, the seeds, knots, branches and everything.  Take notes on what kind of tree you're drawing, the different parts of the tree, etc, but draw them all specifically.  Don't generalize them.  Draw THAT exact tree.  Not 'a maple tree' or 'a pine tree' but that exact pine tree in your back yard, or that maple tree down the road.  Pay attention to how thick the branches are at the trunk, and how thick they are near the end.  look at how they bend.  Look at the angle they're at, and how many other branches sprout from them.  Then do a bunch more.  You'll end up with a huge amount of tree vocabulary in your head that you can later use to draw and pixel whatever sort of tree you want.

Draw some still life stuff as well.  Doesn't have to be fruit or anything.  Just get some stuff you have lying around and put one or two things in a nice, lit area (a windowsill or table where there's good sunlight coming in is a good place) and draw that.  As was also suggested, getting a book or two would help a lot.  Replies here aren't going to help you nearly as much as a solid book, like the Loomis books or "Drawing on the Right Side of the Brain".  Ask for them for christmas (if you celebrate it) or your birthday.

Anyway, keep working and posting.  I wish I could help more with the tree thing specifically, but I don't really know what sort of advice anyone could give that would make it 'click' without overwhelming you with stuff.  Doing still life and other stuff is my suggestion, but that doesn't mean you can't work on the trees or whatever else you want.  Just start laying a foundation.  Trial and error can only take you so far.

2D & 3D / Re: Sprite interpretations
« on: September 11, 2009, 08:30:10 pm »
I've always loved your stuff, man.  These are great.  I might join in, cause I've been watching some videos lately that I think you'd probably appreciate:

World of Longplays -

There's a guy who seems to be systematically posting these on Youtube as well.

Basically they're full playthroughs of entire games.  There are Amiga and C64 videos as well.  Much nostalgia to be had.  So I might end up going back and re-envisioning some of those older games to work in a modern sort of way.  Actually, I've always wanted to do a remake of Kid Chameleon, so that's probably one I'd do...

You want us to just post our own interpretations in here?

2D & 3D / Re: The good, the bad and the just plain ugly
« on: September 07, 2009, 04:13:38 am »
I agree with allaze-eroler.  You could be getting so much more out of your textures right now.  That last texture you could easilly fit onto a 256 if you mirrored some bits and re-used other bits.

Dunno what tools your using for these, but It helps if you're mirroring the model's geometry to unwrap it before you apply the mirror to the mesh.  I dunno how other apps do this stuff, but with Blender that's the way it works.  I'm positive that other apps have a UV mirroring option, tho.

I like the really painterly style to these.  I wish more games would use that sort of approach to their art rather than the endless parade of normal-mapped hyper-real gray/brown crap that gets shoved out all the time.  I'd love to see more color in games.  So in that regard, keep doing what you're doing.  My only suggestion is to do it more efficiently from now on :P

Pixel Art / Re: crit help deciding which character is better...
« on: August 30, 2009, 02:30:29 am »
If they're going to all be used in the same game, then the shading and colors should be consistent from one to the next.  The far right badguy has a lighter outline than the other characters.  His background hand is also flat shaded while the other bad-guy has lighter shades on that hand.  The guy on the left is standing in a similar pose as the other two, but I can't see his far arm the way I can see them on the bad guys.  I'm gonna guess that the guy on the far right isn't quite finished yet, since he's missing some shading that the other one has as well.

As far as style goes, the human character and the two enemies seem to be done in different styles, almost as thought they were drawn by two different people.  The enemies have really exaggerated proportions and those legs that taper to points.  The human is less exaggerated and overall is less stylized than the enemies.  It might just be personal preference on my part, but the difference in styles is pretty jarring to me.  The proportions look off to me as well - for the record - but the colors look alright.  The shading could use some love, tho.  It looks like the light is coming from just to the left of the 'camera', but there's some conflict in the way the character's shaded.  The arms seem to indicate light from slightly above, but the highlights suggest it's slightly below.  Both the heaviest shaded parts AND the highlighted parts are towards the back of the arm which suggests light coming from both the left and the right of the camera.  The other areas seem to be pretty much in agreement that the light is slightly above and to the left of the camera.

The bird, however, is lit from the right, and the highlights on the character's hair also indicates a light source to the right.  If the bird is going to be on the character like that often, I'd recommend they both get the same light source so they don't clash with each other.

The 'old' character you have in that image looks a LOT better.  It has better proportion/anatomy and perspective and is more interesting to look at...  I don't think either of them really 'fits' the goblins, tho.

Pixel Art / Re: [Wip] Skull monster C+C
« on: August 30, 2009, 12:06:52 am »
I didn't do an edit of the sprite, I did a paintover of your drawing.

I think a lot of the problem stems from the lack of structure going on.  I just picked one light source and did a quick shading pass over a trace of your line art (and I see I screwed up a couple of shadows, but meh...)  In your painting and in your sprite you have some serious pillowshading going on.  The edges of all the shapes get shaded and the center of the areas all get lighter and lighter.

This flattens out the whole thing, but more than that, it makes the actual structure of the object pretty hard to figure out.  I just guessed at a lot of the upper structure, because I had a hard time reconciling the overall form of the head and the way the skull-mask thing attaches to it.  The goal of this was basically to point out the pillowshading and hopefully show you why the sprite was hard to figure out.

I'd suggest getting the basic shape of the sprite down and figure out how to shade it, and then worry about the details like the veins and surface texture and stuff.  The veins add a lot of noise to the shape right now, and they're not really placed in any sort of logical way, so they probably add to the confusion.  You can tell what they are in the larger drawing, but in the sprite they could be veins or scars or a texture or some sort of indication of the sort of material it's supposed to be...

But I'm more a painting/drawing sort of person.  I don't have a lot of pixelling experience, but the good thing about drawing is that you can translate the same basic skills into pixels.  I think the one thing you can pull out of this particular critique is to work on your overall 2D drawing / painting stuff as well as your pixel stuff, because I think you could stand to learn a lot about form and shading.

// EDIT // Added a better version of the paintover.  I forgot the bulge for some reason and it threw me off.  Crit remains the same.
// Edit 2 // slightly cleaner / better shading, but it sorta strays from your concept a little I think, so I linked it instead...

2D & 3D / Re: VIPER
« on: July 12, 2009, 06:13:54 am »
It's funny you ask.  I have been at the cusp of finally getting off my ass and learning the game engine lately.  I still dunno if I'll work on that yet, but I do know that there are some very helpful people on IRC.

#gameblender on Freenode.

Several weeks ago I actually started learning it, and set up a blog.  I tend to do better at learning things by showing people how to do the stuff I just learned how to do (even if no one else is around to teach...).  I have one post on there right now, but it's literally no more than just some extremely basic info on the game engine that it pretty much just common sense.  I had started on the second one, but right now the files for it are rotting in a folder.  Had some other crap come up.  Might jump back on it tho.  There aren't a lot of beginner-level GE tutorials and information.

I just have so much crap that I'm working on, and none of it really seems to get anywhere :P

Your initial layout was a lot better.  Texturing and Unwrapping is something that you really just have to learn with practice and by looking at others' work.  There are a lot of .obj files floating around out there.  You should DL some or check out some game models in obj format and take a look at how they did their layouts.  I think Valve has some tool that will let you look at their source-engine stuff?  Or maybe you can import them into one of the SDK tools somehow.  I've not used that stuff in ages.  Anyway, you'll learn a lot that way.

2D & 3D / Re: VIPER
« on: July 11, 2009, 03:44:08 am »
When using Blender for small UV layouts, there are a few tricks you can do to save space and get the pixels looking nice and square / uniform:

1) If you use a mirror modifier for the model, you can unwrap the model before you apply the modifier.  This will let you unwrap only half the model and mirror the texture to the other half.  The downside to this is that any text from one side will be reversed on the other side.  You can get around this by applying the mirror and then editing the UV layout for the faces you want to edit.  Very useful technique if you want to make a character with an asymetrical face or some other asymetrical aspects, but with other areas mirrored on both sides.  Another downside to this might be the fact that mirroring this way requires that there be a seam running down the middle of your model, which can add to the geometry.  If you have a very square model, it might seem a waste of polygons.  There are some methods to avoid the extra geometry, but they would require more texture space and a bit of extra unwrapping to get rid of the seam.

2) There is an option called "Snap to Pixels" that will snap the verts to the corners of the pixels, allowing you to get perfectly square areas to have perfectly square textures.  Snapping takes some practice, tho, as you can easilly end up with warping if the shape of the face in the UV editor is too far off from it's shape on the 3D view.  For really low-res models, these little warps can draw a lot of attention to themselves.  I'd recomend snapping to pixels to do the layout, getting the basic layout down and then turning that feature off and making adjustments to minimize any warping going on.  Snapping is incredibly useful on some models.  Especially models that are very boxy or that are extremely low-poly.  Try to keep the UV islands as simple as possible, with few polys per island.  Makes it easier to prevent warping and easier to conserve and re-use texture space.

3) When I unwrap a very low-poly model like this, I tend to pin (p key) all the verts after I unwrap it, and then if you need to edit the seams at all, unpin the areas you need to alter in the layout and unwrap it again.  You'll have to scale them down to fit (Blender will unwrap anything to make use of the full UV space, ignoring any bits of the layout that were unchanged), but you'll be able to alter parts of the layout that way without losing all your work in the other areas of the layout.  Especially useful if you have those areas textured already and don't want to have to waste time re-texturing those parts, or re-fitting the islands to them.  Especially useful if you are trying to add some asymetrical details to sections of the model while other sections are mirrored.

2D & 3D / Re: Any one know the specs for psx/ps1 games?
« on: July 04, 2009, 06:39:31 am »
Yeah, specs for the PS1 are hard to really nail down.  I think the performance of that system was largely dependent on the software it was trying to run.  Just looking at screenshots I'd say that the average character had in the ballpark of 250-400 triangles.  It's hard to really know for certain.   A lot of the 'limitations' on that system were probably just the result of the technology and the techniques being so new.  I mean, obviously it can only do so much, but according to some of the numbers I've seen, there are games that came out on the system that sould have been impossible.  I doubt Sony even knew exactly what that sytem could do.  Seems like every 6 months or so a game would come out that was previously not possible on the system.  They even made a big deal out of crap games like "Blasto" because it streamed the levels off the disc on the fly.  It was revolutionary at the time (tho also not that fun...), and the levels did look fairly decent sometimes.  Probably a lot more geometry going on in games that did that.

And I don't know about 16 colors per map, either.  I've been looking at some other games - some of the later FF games and RPG's especially - and some of them seem to have used more that 16 colors per map.  FF9, for instance...  lemme find a texture real quick...

Just checking the Zidane texture, there are well over 400 colors in use.  I'm going to assume that this is for the battle mode models, where they are more detailed.  The resolution is odd as well.  Looks like 128x192.  Dunno how common that was.  It could also be a 128x128 and a 128x64.  I honestly don't know how the system handled something like that, but it couldn't have been an easy feat or more games would have used textures like these.  I'm assuming that there was some some bit of code that square used to shoehorn several of these into the PS1's memory at the same time.  If you cut that into two textures then it looks like each sheet might have an 8bit pallet, but both would have their own pallets... I'm too lazy to check :P, but that's how it looks.  It's definitely more than 16 per sheet, tho.  The backgrounds seem to throw off that limitation as well:,25059/

If I had to guess from that game alone, I'd say that you could use 128x128 textures as a max size with a unique 8bit pallet for each.  That doesn't even sound possible on the system, but apparently FFIX did the impossible.

It might look more 'genuine' to limit to 16 colors or 24 or 32 or whatever, but those limitations seem to have been defeated by the last couple years of the system's life, as there were several other games that used more colorful texture sheets.  The later FF games, Vagrant Story...  I generally just paint and don't worry about it.  You can always reduce to 8bit later and then clean it up a little. 

Vagrant Story.  That dragon thing in the second shot there has to be around 400 or more tris.  The main character probably has a similar count.  Guessing from the screen shots I've seen, I'd guess very DS-like limits.  Like 2-3k on screen at once...  maybe 3 or 4 times that in a scene or level or whatever...  I wonder what the tri-counts were for games like Gran Turismo...

But yeah.  There are papers all over that i've read that say the PSone can push this much or have this size max texture, etc, but then I'll stumble onto some screenshot or something that will defy those numbers.  It appears like you could get away with a lot if you knew how to cheat the system a bit.

General Discussion / Re: Official Off-Topic Thread
« on: May 30, 2009, 05:43:10 am »
I think the way they've lit them is gonna kill a lot of the effect.  Look at the Egypt stage.  It's dim and red and the ligthing is from the bottom.

Now look at the characters...  They're high contrast and lit by white light from above.

See...  I always thought that most 2D games avoided really emphasizing things like light direction and high contrast lighting for this exact reason...  you'd have to design the characters and stages to conform to some specific rules about lighting and light sources, and it would have to be from directly above, or flipped sprites would be lit from opposite directions...  The only stage that would make sense with their ligthing would be a prison yard stage at night or dusk with searchlights following the characters or something like that.  I don't see them fitting into any other environment really well, especially not something lit so differently as that Egypt stage...

I'd like to think most people won't notice, but that seems like it would stand out a lot.

It could be the emulator or it could be the rips.  i dunno how to rip sprites, so I dunno if they have to re-map the colors to a normal pallet or what that whole process entails.  It could be the rips have the incorrect colors.  i dunno.  With the NES, all you have to do is just alter the colors to match the NES's closest colors if the colors are off on the sprites.  With the 16 bit systems, there are so many colors that getting them just right will probably be hard if the rip or screenshot is off.  I dunno.  I think Sonic will be recognizeable regardless of what specific shade of blue he's got.

There was one other thing I forgot to bring up, tho...  I know the older systems had a limit on how bright a color could be.  Their pallets generally stopped before they got close to being pure white or bright bright yellow to prevent screen burn-in.  Maybe the emulators are brightening the colors to compensate?

The scanlines, I do believe, are a way for the games to look more like they did on TV's as well...  You can't turn them on unless you're playing at at least 2x the original resolution.  Then the emulator will just tint every other horizontal line a bit darker.  on a CRT TV you couldn't really see individual pixels very well...  everything sorta blended in ith the pixels surrounding it.  On modern LCD monitors, tho, playing at 2x or 3x the resolution of the original games will result in huge blocky pixels and the scanline options are there to soften that effect a bit.

Pages: 1 [2] 3 4 ... 7