GR#167 - Creature From Hell - Shading

[Zizka]

Your colours were kinda monotone (at least more particularly in v2 which I had saved to work on ); as long as you keep the local colour (as uninfluenced by lighting) around the middle of the ramp you can retain the impression of it being red…without everything being red. I desaturated my lows ‘cos less light = less colour and moved towards yellow in the highs ‘cos that’s a nice fiery look, and so they don’t have to be ultra-saturated in order to contrast (yellow being inherently lighter than red) while adding white would lose colour identity and look drab.

There are a few things I’d to discuss here. I’ll use the 3rd version however since it’s the latest version.

So, by local color you mean the color which isn’t touched by either light or shadow? But everything in influenced by light, so you mean the parts of the monsters which aren’t touched by highlights, right? I’m not sure I understand. Look here:

 
So this zone, would be considered the one with the local colors, right?

retain the impression of being red… without everything being red.

You lost me here.

I’ll try to make sense of what you’re saying.

So by the ramp you mean the variations of the local color I think. In version 3, I used this:

 
The arrow points the “local colour”. At that point of time,  I assumed the local color was this:


But based on what you’re written, that would be incorrect as this is influenced by lighting.

I desaturated my lows ‘cos less light = less colour and moved towards yellow in the highs ‘cos that’s a nice fiery look

By your “lows” you mean the darker colors? Like these ones:

 
So my low-color here would be adequate because it’s fairly desaturated.

By “high” you mean “highlight”. So what you’re saying is that instead of remaining in red, you tended to shift towards yellow in the highlights. So you sort of did what I did with my highlights:

 

I think the issue with detail is as much the bit shapeless/awkward monster design as lack of space for relevant stuff. I dunno how the painted ref. is related to what you’re doing but it leans quite heavily on those omni-rimlights to be understood which I think would be a major pain as an animated sprite, the tentacle-things flatten/unbalances stuff in their current position (also tangent on the farmost) and the yellow wing tips distract from the glowing mouth; are they glowing too?

I agree about the rimlights which means I’m gonna have to change the concept for the monster. I don’t like those tentacle things either, they look off. I need something which won’t be too much of a hassle to animate and I can already see how those rimlights will drive me psycho when I get to the animation of the wings.

Gotta reiterate quickly what’s been said on form already; looks like you’re flood filling black linework rather than shading and I’m sure this has been mentioned before for stuff you've posted.

I do use sections which I flood fill afterwards. It might have been mentioned before but then again I’m usually very meticulous about criticism but I might have missed that. Also, sometimes, it’s just a matter of me not knowing how to do something. For instance, here I think you mean that I should share general areas first and then detail things. I’m still in the process of learning so sometimes what might be obvious for a veteran is not clear for a beginner at all (i.e. my questions above).

You could really do with practising simple volumes/contrasts without lines and sketch stuff out as rough geometrical primitives (boxes, spheres) before starting details or specifics

So what you’re saying is that I should follow more of a painting approach as opposed to a drawing one. In other words, in your opinion, I should try to create shapes and volume without using lines. I’ve read in books that painter draw general sketches before painting everything. What is the issue with drawing lines before? Do they act as crutches and prevent me from learning other fundamental things?
 
So I’ve decided to change the body of the beast.


So I will show you how I would normally do this.

First of all, I would draw a general outline of the thing with the 1 pixel pen in black like so:

 
Afterwards, I would clean up the outline like so:


(Yes, I know it’s not accurate just yet, I’m just sharing the way I work).

Keep in mind that the whole time I have my reference to look at like so:

 
Then normally, I would use the same tool to draw the various, obvious separations like so (only doing one for now), which I would flood fill with what I consider to be the dominant color:


Now this is what you’re saying I’m doing wrong in your opinion, Facet.

Now that I’ve explained the way I’ve done things, how would you tackle this? I’d certainly crave improving the way I create stuff, that’s for sure. 

[Johasu]

This took about 30 seconds.
You could always grab a selection by color of the reference images background ( white in this case) and then paste it onto a colored background to get a rough shape of what you are drawing. It's almost instantaneous, though it removes the aspect of really drawing the shape.  If you resize it and clean up the choppy edges from the resize and the .jpg pixelation on your reference image you get a solid blob of color shaped fairly close to the reference that you can then work on detailing without border lines.
Since you are working from such a close reference, it seems like a much faster way to get started.
Also...
I have been learning that working with black lines can be a hindrance. They immediately shadow the outline of your creation/piece and give it a flat/embossed look right from the start. [Sorta like a sticker with a shadow around the edge]
Working from a shape of color seems to work much better for keeping your mind wrapped around shaping the piece with colors rather than borders.
It's better to fill the inside and then trim the edges to look like the reference than the other way around, I think.

[Ymedron]

Probably drawing by eye and working the form rather than trying to trace the outline would yield better results.
If you wanted to create a sprite out of the picture without studying how to create original ones, it could be simplest to just shrink them down (possibly with some contrast trickery) and trace over them.
Arne's work-in-progress tutorial ( http://androidarts.com/pixtut/pixelart.htm ) made a good point about some designs working better in pixel form than others. It might be a good idea to abbreviate details or make them bolder so they'll still be visible in small sizes.

[Zizka]

You could always grab a selection by color of the reference images background ( white in this case) and then paste it onto a colored background to get a rough shape of what you are drawing. It's almost instantaneous, though it removes the aspect of really drawing the shape.

Thanks for the suggestion but that would be "cheating" in my book. My objective in doing pixel art is to improve as an artist so I'd rather not use any shortcuts.

Same thing here:

If you resize it and clean up the choppy edges from the resize and the .jpg pixelation on your reference image you get a solid blob of color shaped fairly close to the reference that you can then work on detailing without border lines.

I have been learning that working with black lines can be a hindrance. They immediately shadow the outline of your creation/piece and give it a flat/embossed look right from the start. [Sorta like a sticker with a shadow around the edge]
Working from a shape of color seems to work much better for keeping your mind wrapped around shaping the piece with colors rather than borders.
It's better to fill the inside and then trim the edges to look like the reference than the other way around, I think.

This might be an idea however, thanks for the input.

If you wanted to create a sprite out of the picture without studying how to create original ones, it could be simplest to just shrink them down (possibly with some contrast trickery) and trace over them.

AH! Blasphemy !  ( ). No tracing though or resizing either. Then I would just be stealing the original ref' and I don't want to do this.

I want to go at this as I would if I were drawing with a pencil and using a reference otherwise, where's the challenge? The learning?

Thanks for both of your replies however, I think they offer good ideas but I'd rather do it by relying on my own skill as opposed to using tools. I realize I'm making life harder or myself and that it's not necessary but that's the way I'd like to go about creating pixel art.

Your drawing looks neat though 

[Ymedron]

I think I typed my message poorly- I meant to convey that if your aim is to mindlessly copy, tracing is simplest.
However, to learn the most, it's best to study what forms create the creature. Trying to visualize what the form is like.
(for example, the maggot is basically a tube with a flattened belly + a muscular back of the neck. The tail/butt portion is more like a fatty bulge with a tail sticking out, from my understanding.)

It might even be useful to completetly ignore the entire colorscheme and details of the creature until it is shaped how you intended.

My drawing was made freehand, I didn't do any tracing at all. The smaller version is merely shrinked from the first, where I drew a little bit over some parts that had disappeared.

Here are some examples of trying to understand the form, except I usually wouldn't draw it out like this. (the top picture has the worm laying entirely on its belly rather than scrunched up like it is in the picture.)


Edit: On further look, the top picture isn't necessarily even correct (the butt-area is pretty weird. I need to study lighting more as well.

[Johasu]

It is cheating.
But, you are already kinda doing that by framing your image around the reference so closely. Everyone was talking about how you were getting caught up in lines and then filling them with color, and you are templating your image by looking back and forth.  I was just suggesting a method that might help you get into shaping more than filling in outlines.

Either way you go about shaping the image, the most growth you experience will be on the contents within.  Not the black line around the outside.  Your way isn't wrong. It may help you get better at drawing gradually. I think that it is limiting you a bit, as I have noticed that method does on my own efforts when I try it.

Ymedron has a good point about altering the pose or picturing it from another direction to help you visualize the form.

There is nothing wrong with practicing the process on any level.  The best way to improve is to pixel a lot.

[Facet]

Ah ok, it seems like a bunch of people already replied while I thought about what to write and made some tea, so there's some redundancy in the following, hope it helps anyway    

Local colour has couple of connotations perhaps; obviously everything is in fact affected by lighting or you wouldn't be seeing it! So that's poor wording but I mean to say 'native' colour; that which you know the object to be under neutral conditions (moderate white light) or intend for the object to be (like a material colour or texture map in 3d software before being lit). The local colour might well not even be present in a lighting scheme if it's not well lit or the source(s) are colourised (like a photographic colour cast), but it's most likely to be truest away from extremes, and on the lit side.

In a sprite you might want a clear colour read in multiple contexts or at low detail so you don't want to be too far removed from the local, and the reference is not a good model for that. I mention it also because you had some long ramps with a lot of one colour (with some very high saturation that can appear 'naive'), or that added white in some WIPs, and you mentioned wanting to discuss theory earlier, not because I thought you weren't making progress.

Highs, lows, the 'key: yep.

Aside from being a pain to animate a reliance on rim-lights is much the same as a reliance on linework in that it's often making up for a shortfall of the underlying form; it's not that lines are bad practice: they're great and more rapid for many things, I love 'em, it's that for the work I've seen of yours, I think that that process for you at the moment is prioritising the wrong stuff, keeping you fixed in a 2D mindset.

Whether the outline is neat and accurate to the ref is not important when that's all you've got done, what's most important is the mass of the form; if you zoom out and squint at both ref and WIP they should look right; you want to approximate the broadest, simplest stuff first so you aren't dealing with everything at once and can stop at any point and either you or any audience will have a good sense of what's going on and what the finish might be like.

Actually just the process of constantly reproducing other 2d images closely is not helpful because it's not requiring you to think in terms of design (how an impartial audience will see something) or of form and light, and this is what I've seen mentioned in some of your other threads, and why I suggest making the effort to practice by ditching lines altogether and drawing from life.
 

[Zizka]

Again, thanks for the thoughtful replies, I really am blessed.

@Facet:

Ok, I'll ditch the lines and focus on drawing from life if you think this well help me progress. Just one more question.

When you say ''drawing from life'' do you mean that I should pixel photographies in pixel art or that I should draw using a ''real'' medium on paper directly through observation? I'd just like to be clear.

[Cyangmou]

So what you’re saying is that I should follow more of a painting approach as opposed to a drawing one. In other words, in your opinion, I should try to create shapes and volume without using lines. I’ve read in books that painter draw general sketches before painting everything. What is the issue with drawing lines before? Do they act as crutches and prevent me from learning other fundamental things?

In general there is no issue if you start out with lines.
Even for a painting lines are the  fundamental, which might not even be visible in the final artwork any more (since the painting just works with contrasting color-areas)

But with digital art, you don't have to think about that your colors aren't 100% opaque, so you aren't restricted to a very strict working process, as you are with traditional drawing and painting, since you can overpaint stuff hundredth of times and nobody will see the layers of colors in the end.

However, we are talking of pixel art and the difference with pixel-art lines is, that they get relatively thicker the smaller you get and you have to AA them, that they look good.
And lines are also use space on the pixel grid, means if you have a line-heavy style, you have not as much space.
Thick lines tend to clutter things, if not used right - arnes tut seems to illustrate that great
And pixel lineart can get messy easily and be a lot of unecessary work for slight changes

I mean this are various boxes, shaded with different line approaches


First you have to know about how things look in the final artwork and then you have to think about how to achieve it.
 
1 is the line art approach,
4 is the painterly approach
2 uses internal lines, while 3 uses just an outline
2.1 uses highlighted lines 3.1 uses a shaded outline
3.2 uses shaded outline and a highlight

You can start with 1 or with 4, but depending on the look you want to have 4 might be a much faster way for organical stuff.
If you want to end up with 3 and you start with 1 you would have to erase a line later on - if you don't draw them in and start with 4, you don' thave to erase the lines and once you fixed the shapes you could just add a line around them and end up with 3 immediately with less time spent.

People prefer different styles. Some might prefer 4, some maybe prefer 3.2, some completely other variations.

I mean in the end it's just all about the process, how you get to what you imagine.
I wrote down some notes about why using line art with pixels might not be optimal.
http://cyangmou.deviantart.com/art/Pixel-Art-Process-413418385

For architectural stuff or mechanical, general helplines etc. it's maybe better to use lineart. YOu have to decide it on a case-by-case basis.
In the end you anyways have to decide yourself how you solve a problem to reach the outcome you have in your imagination with the most efficient way.

[Zizka]

Wow, another insightful message. I really am lucky.

Truth, I really just want to improve as an artist (through pixel art if possible). If I have to stop PA for a while to focus on real life observation drawing (which is that I think is implied from "drawing from life"), then I will do so.

I think you've added some valid points for and against lineart. For my part, I find that is sometimes use lines as shortcuts to achieve something I could have done with a better understanding of form and/or contrast.

I'd be really curious as to how wolenoctis, facet, ryumaru and HDMD proceeded in getting the result they got. At this point in time I'm more interested in the process than in the final result. Deep down I feel like the way I draw stuff is not the best way so I'm really curious to read about other people's art methodology. In other words, which of the 4 approaches mentioned here the various members use and how they go at it in a step by step fashion.