Let's Draw an Ogre!

[Shane 'ProgZmax' Stevens]

Okay, so as some of you may know, my primary area of expertise is pixel art.  I've never been particularly 'adept' with hand drawn artwork and I avoided getting a Wacom for years because A) They were inhibitively expensive for me and B) I didn't really need to do any kind of hi-res digital rendering.  The latter has changed and now my trusty Intuos 4 is starting to get the shit beat out of it on a daily basis until I get reasonably competent, which leads to the piece below:

[imgzoom]http://i485.photobucket.com/albums/rr218/ProgZmax/ogre_zps7a847470.png[/imgzoom]


Basically, I think I spent about 3, maybe 4 hours on this and wonder if that's too long for something so basic?  The steps I took were:

1.  Establish a basic sketch (this always takes me the longest).
2.  Fill it in with basic values.  I have the hardest time getting the paint bucket to just fill areas I want without dumping on the whole canvas, so advice on how to make it work more like DPaint/Pro Motion flood fill would be great (or an alternative).
3.  Add highlight without any aliasing.
4.  Add shadow without any aliasing.
5.  Use an 80% opaque brush that tapers to blend the highlight/midtone/shadows.
6.  Go back and add detail and cleanup.

I am still learning how to get the most out of paint programs and how to maximize my workflow.  Also, a general critique of the technique would be nice, though take in account that I'm aware it has some of that paint streak look still because this is still very much practicing techniques for me.  I've switched to using Paint Tool SAI as the controls are so much more intuitive than photoshop and the brush movement is smoother and I don't suffer any of the occasional streaks or lag I do in photoshop.  Fortunately, it outputs to .PSD so I can easily import my linework there to add effects and such, so any tips for photoshop are welcome.

[Andail]

I think it's hard to say if you've spent a reasonable amount of time doing that thing, since a person who's really skilled and experienced wouldn't end up with something like this anyway (no offense). But let's say that with a year or so of training, you would probably reach this stage in a matter of minutes.

Some c&c regarding your technique.
I don't think you should ever use a paint bucket. Vary brush size, shape and opacity, and that's all you'll ever need.

Avoid adding highlights and shadows to every individual shape. It deprives the piece of a focus, and creates a lot of small "pillows" everywhere that are just distracting.

I think the colours are too disparate. The face is cyan, the lips are really pink and the eyes hazel. You have used both blue, black, darker cyan and desaturated cyan for shadowing.

I would like to see the sketch you've made for this, as I fear it's not... sketchy... enough. I'll illustrate my points later.

[Andail]

I read my post again and saw that it came off as a bit rude - what I meant about "a matter of minutes" is that I think you will be that quick, considering all the experience you have with drawing, not any Joe Shmoe.

Since I hate to leave comments without leaving a paint-over, I rushed home to create this:

It's basically meant to suggest a better starting position - from here you can define and refine shapes and lines, and add more saturation and highlights to selected parts.
Note how in your picture, eyes and teeth seem almost phosphorescent - without a direct light, even whites will disappear in shadows.

PS:
Looking again at my picture, I realize I have altered too much of the original - rather cartoony - shape and style for it to be a fair paintover. I hope it can still be useful.

[Daniel Thomas]

Shane, what resolution are you working in? I think you should go really high, as long as there isn't a lot of slowdown there isn't really any downsides working really large. Especially for production when you might want to go back and tweak the image.

Basically, I think I spent about 3, maybe 4 hours on this and wonder if that's too long for something so basic?

Hmm, too long for what? I personally wouldn't focus on the hours, but just try and get a good result. Speed comes later.

1.  Establish a basic sketch (this always takes me the longest).

Yes, spend as much time as possible on the sketch. It is really the base of the painting and you focus on only solving drawing errors.

2.  Fill it in with basic values.  I have the hardest time getting the paint bucket to just fill areas I want without dumping on the whole canvas, so advice on how to make it work more like DPaint/Pro Motion flood fill would be great (or an alternative).

Paint Bucket is fine In my opinions, especially for production art the bucket and selecting tool are time saver creating crisp and clean art. Unfortunately I haven't found a way to make it behave like Dpaint or PM. I use lasso tool to select areas and fill them in under the line drawing. But if you have really clean lines that are closed you can probably experiment around with the settings - make sure you check and check contiguous when you want to fill a closed area.


Something else you could consider is to use a more neutral background color. And maybe not to paint "floating heads", but to connect it to a body/neck.

[Anian]

In Photoshop, if you have a separate lineart layer, Ctrl+click on the icon of the layer to select the lineart (marque it).
Then select inverse AND expand the area by at least 1px.
Now (still keeping selection) in the layer below you can fill in areas with the Paint bucket/large brush and it will make a very cool basic flat color stage. And every thing will fit nicely with the lineart.

This I found to be a nice process if making simple/clean illustrations, but it can be used to set up the basis of a painting as well.

On the actual drawing - I like the construction of the face Andail made, it's much more symmetrical in shape, although that's a pretty annoying tangent on the ears  but ProgZ's has a more interesting silhouette, but you really shouldn't cut it out like that, makes the whole thing look flat.
And yes, as DT mentioned, since PS is raster image, you should probably start of big. I like to do work at about 200dpi and 2048x2048px canvases when sketching stuff (it leaves room to enlarge parts and still not brake the 72dpi resolution of monitors), but you should test how big you can make the canvas and resolution and still work smoothly.

[Monsieur OUXX]

The drawing technique is very good. Overall the drawing looks good and pretty and arty and skilled.

BUT the face's shape is not very photo-realistic:
- why is the skin so shiny?
- the cheeks are so round and seem totally separated from the face. Maybe they could be flatter, and joined better to the jaw.
- the eyes could have more depth (if the forehead is a real orc's forehead, the eyes are way deep below)
Overall, I'd say the face could be more angular, and have more ripples and scars and moles, etc.

[Shane 'ProgZmax' Stevens]

Okay, let's see here..

Andail:

Avoid adding highlights and shadows to every individual shape.

I don't really see that as a problem since I didn't apply highlights to everything, just areas that in my mind projected outward enough from the face to catch some light.  I don't find what I did distracting to look at as you mention, though I know it can happen and I am mindful of it.

I think the colours are too disparate. The face is cyan, the lips are really pink and the eyes hazel. You have used both blue, black, darker cyan and desaturated cyan for shadowing.

I actually prefer my colors to the ones you chose to be honest.  I like brighter shades and this is a fantasy piece with no 'realism' inferred or implied, though I agree the lips and teeth weren't ideal.  Also, you don't have to worry about hurting my feelings because you're just offering your opinion and everyone is entitled to theirs.  And I do agree that I should be faster at this sort of thing and it bugs the shit out of me but what can you do but keep working at it?  I've never been particularly good with pencils so I shouldn't be surprised I'm not particularly good with a wacom pen either.

The main thing I find interesting about your edit is the use of one of the blotchy, foam pad type brushes.  I'm still pretty much using a round brush and going between opaque and semi-transparent so I haven't really experimented with the benefits of different brush types as yet.  Are there a specific set of brushes you would recommend for various stages of drawing?

Daniel:

Shane, what resolution are you working in? I think you should go really high, as long as there isn't a lot of slowdown there isn't really any downsides working really large. Especially for production when you might want to go back and tweak the image.

My screen resolution is 1680x1050 and I started off (in this case) with a 1024x768 canvas and then copy/pasted the head to a smaller document to show here.

I use lasso tool to select areas and fill them in under the line drawing.

I'm going to give this one a try, thanks!

Something else you could consider is to use a more neutral background color. And maybe not to paint "floating heads", but to connect it to a body/neck.

Very basic rules I'm aware of, but thanks for the 'heads-up' heh.  As I mentioned earlier I copy/pasted the head from my project to a smaller document just to show off, there was actually a body but it was still a sketch and it seemed pretty silly to include.


Anian:

And yes, as DT mentioned, since PS is raster image, you should probably start of big. I like to do work at about 200dpi and 2048x2048px canvases when sketching stuff (it leaves room to enlarge parts and still not brake the 72dpi resolution of monitors), but you should test how big you can make the canvas and resolution and still work smoothly.

I started out with larger canvases but found that they don't really work for me.  I like to have what I'm working on close enough that I can quickly zoom in and out without a lot of panning and screen manipulation, but if bigger works for you then go for it!  Maybe when I've gotten a better handle on the various modes and tricks I'll give it a try again but for now I'm sticking with what works for me.

In Photoshop, if you have a separate lineart layer, Ctrl+click on the icon of the layer to select the lineart (marque it).
Then select inverse AND expand the area by at least 1px.
Now (still keeping selection) in the layer below you can fill in areas with the Paint bucket/large brush and it will make a very cool basic flat color stage. And every thing will fit nicely with the lineart.

This is what I'm on about with photoshop.  Something like filling in an empty area should be very basic but not in photoshop.  I'm not sure I'll even remember (or want to use) this approach over and over but I appreciate knowing about it.

Monsieur OUXX:

BUT the face's shape is not very photo-realistic:
- why is the skin so shiny?
- the cheeks are so round and seem totally separated from the face. Maybe they could be flatter, and joined better to the jaw.
- the eyes could have more depth (if the forehead is a real orc's forehead, the eyes are way deep below)
Overall, I'd say the face could be more angular, and have more ripples and scars and moles, etc.

How many photos of real ogres have you seen?  As far as the skin and facial features go, that's just the way I wanted it to look at the time and how I visualized it in my head.  I do think it could stand to have more detail but this was an exercise rather than a production piece so I didn't want to put on my perfectionist hat for it .


I'll work on an edit taking in account some of the good advice offered and post it up sometime tonight.  Thanks dudes!

[Renal Shutdown]

Andail's paintover looks more "realistic", but I'm not so sure that's the best way to go on this.  What's your original intent for the image, Progz?  Concept art or T-Shirt design, I'd be all for your choice of colors.  I'm also curious as to your progression depending on the intent.  If it was concept art, for example, I'd expect it was a case of blocking in one solid color then adding highlights.  If it was a specific coloring of a pre-designed sketch, I'd expect an entirely different approach from the beginning.

MOUXX is right about it being non-photorealistic, and moreover, if it's not meant to be very stylised, it lacks detail.  A lot of artists using tablets scruff up their canvas first, before actually painting on it.  A lot of concept artists 'texturise' the image part way, with dumping textures into the areas and then continuing to paint over them after, just to break up the plain white they started on.  They both used to remove the solid blocks of color, and both work really well as a shortcut.  I'd suggest trying both options (http://www.fengzhudesign.com/ has video examples that explain it better than I can, especially the one with the shuttles).

I'd also suggest aiming for 6400x4800+ on any beginning canvas.  The tablet picks up so much detail, I find it's better to go for a much higher res to start with a sketch than trying to upscale a low-res version.  Finding a relatively grungy brush, that's shaped how you'd prefer is a good method, too.  I've spent ages getting a canvas to "feel" right, and when it hasn't, the image has generally turned out as recycle-bin fodder.

[EchosofNezhyt]

Am I the only one that thinks working a super high resolution isnt as good? Like 2000+?

[Anian]

Am I the only one that thinks working a super high resolution isnt as good? Like 2000+?

What do you mean by "isn't as good"?

[Shane 'ProgZmax' Stevens]

Okay, I had another look at it with fresh eyes and nothing ever looks as good as it did yesterday so I made some changes.

1.  I adjusted the contrast and hue slightly.  It's still colorful but I've evened the tones a bit.  I also changed the lip colors entirely as I was not happy with them.
2.  Did some reworking to the chin/jaw/cheek shading as well as bringing the eyebrows together and integrating them into the head more. 
3.  General fixes here and there, cleaned up some messy shading, added some depth below the cheeks to the lips.
4.  Made the ears connect a bit better to the jawline.

[imgzoom]http://i485.photobucket.com/albums/rr218/ProgZmax/ogre2_zpsec4a55ff.png[/imgzoom]

General improvement?  I don't really want to spend too much more time on this drawing since it's practice, I'd rather apply some of the tips to something new and see how that works out.

Renal:

Concept art or T-Shirt design, I'd be all for your choice of colors.  I'm also curious as to your progression depending on the intent.  If it was concept art, for example, I'd expect it was a case of blocking in one solid color then adding highlights.  If it was a specific coloring of a pre-designed sketch, I'd expect an entirely different approach from the beginning.

It's very conceptual.  I made a very basic sketch and then applied solid colors over it and then began with highlights and shadows and just started blending.  It wasn't meant for any specific project, just practice.

I've had a look at that website before and for some reason none of the free tutorials load for me .  I also don't 'paint' like most people I guess so texturing the surface didn't even occur to me as something to do.  If you have some advice on what brushes to use when and for what, please PM them to me (or make a tutorial!) as I'm just going by feel right now and some painting tutorials I've looked at, so if there's faster (or better) ways to achieve nice looking stuff then I definitely want to know.

Thanks!

[EchosofNezhyt]

It looks good but I think the whole form of it is kinda flat.
The mouth to cheeks to chin all have a very flat form compared to the rest.
Lip color is way to colorful looks like lipstick I'd go with a purple like before.
I think if you redo the nostrils ad some more geometry/shading to the lips cheeks that it will look great.

Hey btw can you give me some cc on my spaceman thread

[loominous]

I personally hate working in higher resolutions, don't like the feel of it, I like sensing the pixels, then again my stuff is kinda notorious for being rather unsharp, though that's also to do with the brushes that I use.

So I'd say work in whatever resolution feels best/is needed, the only thing it'll affect is the sharpness, the rest is resolution unrelated (unless you go really low res). I personally prefer drawing/painting in reses around 640x###/800x###, and upsizing at some stage if need be. Higher reses does give you the opportunity to work on minute details, but focusing on those is counter productive in many cases anyway.

Cool to see you try out hi-res. Could you post the sketch?

[EchosofNezhyt]

Ok So I spent a few mins...


I dont really consider myself a painter but I think adding a bit more shading on the lower have kinda makes the mouth look better.
I kinda messed it up and I might be wrong but I think it the shading I added gives the face form? or whatever.

Also I changed the lip colors imo it was to bright and didn't look natural in your latest edit.




Ok messed with it again.

So I changed the lips (This came out bad).
I changed a tusk to give it a bit less flat shading.(left one came out decent imo but right looks flat and they look 2 smooth)
Changed the shape of the tusk so they are pushing on his lips.
Changed up nose and face shape trying to give it a more meaty look and darked eye nose area.
and maybe a few other things idk...




And to my other comment drawing at crazy resolution of 8000+? Doesn't make any sense to me at least not for paintings.

Idk about you guys but with a giant resolution the brushes dont blend well (Maybe do to my not as good quality tablet?) and photoshop seems to get alot laggier. (I got a 3.7ghz quadcore and 12 gigs of ram with gpu stuff enabled.) 

I've used 2k/3k maybe but any larger seems like over kill. /shrug

[Andail]

Ok, given the direction Progz has chosen to take, the picture is improving alright, and Frito's version looks good.

I'm just a bit concerned with the process - I really think that more time should be spent on making a sketch and getting the basic shapes and angles and proportions out there, and build up the image from the bottom, so to speak. When you pixelate in low-res, you add elements to the picture piece by piece, but that's a method that shouldn't be used in hi-res painting.

[dactylopus]

I feel that the artist has a strong desire to see an ogre in lipstick.

It's good art, and I like the improvements you've made.

[miguel]

Regarding critique some things felt wrong, it looks like a mask of an ogre rather than a face. And the mouth was very weird. There's no "meat" between the nose and the upper lip. This ogre has too many teeth as well. I think that this has to do with drawing a 100% front view. A bit of an angle and it would have been much easier.

Well, here's a quick sketch I made, hope it helps:

[Andail]

Miguel, couldn't agree more, with basically everything in your post.
Maybe we should make it a workshop activity, where everyone paints their own monster head, from scratch? Any takers?

[Shane 'ProgZmax' Stevens]

Frito:

The darker lips look better to me but I'm not really 'bothered' by the brighter ones, either.  Bringing the contrast down a bit would be a good compromise.


Loominous:

Cool to see you try out hi-res. Could you post the sketch?

I had a look back at my SAI project file and found out that during the process of edits I merged over the sketch I made ;(.  I probably should have considered that someone might want to see the sketch.  Fear not, I'm going to post something else I've been working on a little more seriously that I've broken up into several layers and I'll be happy to send you the pdf for it if you're interested as I'd especially like to know what you would do differently/the same.

Miguel:

There's no "meat" between the nose and the upper lip. This ogre has too many teeth as well.

Again, this really comes down to personal interpretation or vision of what an ogre does or should look like.  I agree that the upper lip area could stand to protrude more but as for the teeth it's just silly to presume how many/few a fantastical creature should have.  The angle I chose certainly wasn't ideal to maximise depth but I don't quite agree that it looks like a mask, either.

Andail:

When you pixelate in low-res, you add elements to the picture piece by piece, but that's a method that shouldn't be used in hi-res painting.

Why do you think so?  I'm curious because I've seen some photoshop tutorials both on the net and on youtube that do pretty much what I'm doing and I like their results, and I've never been one to obey 'rules' simply because they exist.  What sort of problems do you foresee happening as a result of doing things this way?


I'm ready to move on from this with some of the tips offered in mind, especially since I didn't do a great job of using layers (I'm still not used to or comfortable with breaking things up).  However, I do have another drawing I was working on a week ago which does have things broken up nicely into layers that I will be posting shortly!

I'll leave the thread open for a week or so in case people want to respond or continue to offer advice.  Onward!

[loominous]

What sort of problems do you foresee happening as a result of doing things this way?

One problem with this approach is that it kinda assumes that the colours/values of objects exist independently, in which case you can simply work on them individually, and then in the end everything will work out fine.

The problem is that they don't. All that we can see are reflective surfaces, basically more or less muddy mirrors. And just as you can't paint a mirror without knowing what it's gonna be reflecting, you can't really paint any object without what it's reflecting, that is, if you're going to do it convincingly.

It's quite popular to talk about 'bounce light', but it really is a misnomer, because everything we see (apart from the lightsources themselves) is bounce light. It really is like a hall of mirrors, where the reflections never really end. If you're painting a soccer ball on a grassy field, the ball and the grass will be reflecting the sun, more or less diffusely, but the ball will also reflect the green grass which in turn is reflecting the sky and surrounding, and the grass will reflect the white ball with the sky and surrounding reflecting in the ball, and so forth..

To illustrate, let's say you're painting a pool player, about to make a shot.

If you go with the individual element approach, you might paint the head first. Let's say you start with black and white and am considering what colours to use.



Now, what skin tones should you use? Well, if you go with the individual approach, you'd simply pick some colours which you think seem like nice skin tones.

What you probably wouldn't do is to go with colours like these:



So what on earth are those greens doing in his face? Well his face, being a muddy mirror, it's simply reflecting whatever light there is, and in this case it's mostly reflecting the green table, because the main light source is obscured, so what we see is the lightsource being reflected in the table, which is then reflected by the skin (this is simplified, since the table is also reflecting the surrounding, which is reflecting everything else, including the table, which is then reflected in the face, which is reflected in in the table etc).



Now, adopting this truly when painting would be completely crippling, so what artists utilize are simplified models. In the case of the pool player, the artist might think: "hm, those shadows would pick up quite some greens from to the table", and when choosing the colour for the table they might go: "so the light source is yellow, and the room walls have a reddish colour, so the table would mostly reflect the yellow light, and only really reflect the red walls to any larger degree in the shadows, unless they're at an angle where they mostly reflect the bright portions of the table".

This probably sounds confusing, but you go as far as you like with this.

By the way, this is one reason why "tinted" paintings often look better, it's because in a way you've already made sure all the objects are connected, and seemingly reflecting one another, as they all share a colour.

I should lastly point out that while I've mainly talked about colours, all this of course applies to values as well, which is why all this has to be considered as soon as you leave the outline stage.

My advice would be to start out with a dark background, perhaps not black because it's tricky to sketch in, and then deliberately start adding light reflections to everything (if you start out with a bright background, you're kinda going to opposite direction, reducing unwarranted reflections everywhere, which makes little sense, really).