I went and fell into a pit for the last few months. The pit was writing the pathfinding for npcs in the game project I've mentioned in here. I have been very productive... just not with drawing or painting.
However, I did have an interesting talk with Muzz. He mentioned that all digital color spaces are gamma corrected. Which in plain english means that EDIT: colors tend to blend lighter, incorrectly. Paintings might be skewed toward lighter values, be more in midrange values. This means paintings have a tendency to lighten.
Before the edit I had here that the color picker wasn't presenting accurate info. That's actually wrong. I checked and the color picker in CSP is correctly presenting linear color space. But the blending in CSP will still be in srgb with a lightening tendency.
Anyway, something to look out for and experiment with. Tried to apply that w/ the paintings below.
Also, I need to figure out some scifi helmets and armor for the game... so more of that soon-ish.
When you say 50% is lighter in the corrected color space, if it looks correct on your screen, and it looks correct on my screen as well even though it's not really 50% grey, why do you need to correct for that? I could see an issue if you were going to print out your artwork. But I mean if it looks right it is right... right?
I feel like you just need to hit a bit more the dark and white spot just to create more interest and texture in the character. When you smash the value and leave a lot of the extremity of the value range out you get what i call ''sad value'' which you loose the richness the drama and the opportunity to create strong visual hierarchy. Smashing the value ( or taking working mostly with midtone) you create muted interest you make everything equally important to a degree.
By adding the hotspot(or highlight) you create some warmth and with a bit more dark inside the shadow you separate the figure from the background even more not that important here but sometime if you don't light properly loose the figure to the background when you don't create a ''1,2,3'' read as scott robertson would say. The hotspot also suggest the wetness/greasiness of the skin.
I also tried to see what this approach would do with the girl butt to see if it was a one time thing and i think you should consider what being said seriously in that one the form was getting lost in the background.
(11-01-2024, 12:36 AM)JosephCow Wrote: Wow that last portrait looks beautiful!
When you say 50% is lighter in the corrected color space, if it looks correct on your screen, and it looks correct on my screen as well even though it's not really 50% grey, why do you need to correct for that? I could see an issue if you were going to print out your artwork. But I mean if it looks right it is right... right?
Re: the portrait, thanks!
And I had a conversation with someone else about the color space thing, and it's interesting to me that other artists have no issue with this. Or don't care, either way.
In the other conversation I had, the other person said something like, "Why does it matter? If something is too light, paint it darker, and if it's too dark, paint it lighter. Knowing this about digital color spaces changes nothing about how I would work."
Well, in theory that makes sense, but here are the issues I have with the color space.
EDIT: There used to be incorrect info here! Actually the UI is correctly displaying the linear color space in the color picker. Rather it is the blending which will tend to be happening in the srgb color space, and that will lighten colors in a way that is nonlinear and somewhat deceptive. The actual change in workflow should be to post process with adjustment layers, and mostly paint as normal. The idea of not neglecting darker colors is still valid, but that's not as related to this srgb color space stuff.
I'll quote Muzz who explains it this way: "8bit srgb has far less fidelity in dark ranges, making blending with dark colors glitchy and kinda broken. It sacrifices the darks in favor of having an expanded midrange."
(11-01-2024, 05:38 AM)darktiste Wrote: I feel like you just need to hit a bit more the dark and white spot just to create more interest and texture in the character. When you smash the value and leave a lot of the extremity of the value range out you get what i call ''sad value'' which you loose the richness the drama and the opportunity to create strong visual hierarchy. Smashing the value ( or taking working mostly with midtone) you create muted interest you make everything equally important to a degree.
By adding the hotspot(or highlight) you create some warmth and with a bit more dark inside the shadow you separate the figure from the background even more not that important here but sometime if you don't light properly loose the figure to the background when you don't create a ''1,2,3'' read as scott robertson would say. The hotspot also suggest the wetness/greasiness of the skin.
I also tried to see what this approach would do with the girl butt to see if it was a one time thing and i think you should consider what being said seriously in that one the form was getting lost in the background.
I think you're right. The diffuse light value got too high in that portrait, and the highlights especially could've been stronger.
The butt painting I think is a better experiment with controlling the value range. It looks good to me, at least. Of course, all this stuff is a matter of preference to some degree...
I am not that aware of all the conversion theory from greyscale to color sadly so take it with a grain of salt and yes monitor and printing is an other level of expertise. My advise is try turning the drawing to color and see and find out for yourself when an when not to use different value range.
Oh yeah, I think I've heard that before in regards to blending of colors. It typically doesn't do it how you would expect traditionally. Not sure I really understand it, maybe I should look into that more. Though I think as long as you are aware of what you're doing and how it is looking it should be alright.
Love seeing updates from you, you have a great line quality in your works that I really enjoy. The Metroid sketch is fantastic, as are the recent value studies you completed. Keep up the great work!
JosephCow: I think the TLDR is simple with the color space. Colors will mix lighter than they should.
cgmythology: Thanks!
Got the first helmet done for the game. I don't have a ton of 3D modeling XP yet, so I spent a lot of time on it, not getting results as nice as I would've liked. Really this was the first 'hard surface' thing I've modeled, and even now I know a few ways it could be a lot better if I made it from scratch again right now.
I wanted to do most of the texturing in CSP but the 'quick edit' function doesn't like applying small details. Or there's some trick to it that I'm not aware of. But anyway, the plan is to make more helmets/armor, so I'll figure out a better workflow as time goes on.
The human figure comes from the Human Generator Blender addon. I retopo'd the body into a bodysuit kinda deal so it's wayyy less verts everywhere except for the head and hands. Big time saver there.
And after that, some concepts for the body armor, and random doodles.
I'm leaning toward some kinda triangular shape/s for the torso armor. And generally I don't want the armor to follow the contours of the body too closely. It's a bit safe and dull to do that.
Keep it low poly you can then apply what you learn to more complex project later on. Getting defeated by complex subject matter before learning to ''drive a new software'' is just not what i advise.
I recommend playing with the matcap option and cavity option to get a better read during preview. Cavity kinda highlight edge and matcap kinda change the overall quality of the object like a quick material you don't need to assign example metallic matte or glossy(not always useful when you have many material but it more use when your planning your object where you want the material to be pretty close to what you imagine. You can also always adjust the matcap if you want to preview a new object and revert back to the older matcap and so on as you switch. Personally i don't play with them that much for me texturing is just something that require a good understand otherwise you are always tweaking parameter just to get there by mistake. I think blocking a flat color is enough you can always use the flat color as mask later on to separate object. Personally i think it best to drawing the texture instead in a beta test instead of skipping and seeing what stick.Like test the reflectivity the bumpiness of the material the color the transparency. The trap with the software is never getting satisfied being able to commit to something in a drawing form i feel it help have a clearer vision once inside blender.
Excellent updates! Everything looks great - I'm particularly impressed by the armor design in your sketch/linework, very well done and professional looking, great job!
Thanks for the advice Darktiste. I've found good use w/ the cavity settings especially with texturing. And thanks cgmythology.
I've mostly continued w/ gamedev stuff, and I forgot about this sketchbook or even uploading stuff to tumblr for a bit. But anyway, I've got the armor done. At least the first iteration of it. Then I moved on to weapons and still have quite a bit more to do with them.
I did eventually figure out the trick with Blender's Quick Edit. Just leave it at 512x512 resolution. It doesn't make sense that it performs better than using a higher res, but for some reason the higher res I try to go with it, the worst the results. Not sure what's up with that.
I've also discovered that using a very soft pencil brush in CSP gets me much better results. Harsh dark lines have never cooperated with me. I redrew one of the faces below. You'll see what I mean.
On the topic of lines... I actually made a discovery shortly after that last post. I had been using the stabilizer setting incorrectly in CSP for years, literally since I started using it. All because I probably read the tooltips for those buttons once and then just never bothered with it again. Which is kinda funny. But anyway, it's worth making sure all the buttons/settings in your program do what you actually think they do.
On the HardRound Discord I'm doing a month long challenge with someone. The prompt is simple, just do the strongest drawing/painting we are capable of.
My plan is to take Greg Manchess's posts on Muddy Colors as a guide. Now, there's way too much in all those posts for me to try to follow it all. I'll try to remember to point out when I'm following what he said.
I had a few different ideas for this painting. Or different paintings.
A painting of some characters from a book series I really enjoyed, The First Law. The characters are sort of Frazetta-like soldiers and barbarians. I imagine it as a snowy, wooded scene, with the characters emerging from the trees, and they are just spotting some enemies. Each character has a different reaction to this discovery.
Some sort of witch painting. If you go back in my sketchbook you'll find an incomplete witch painting. I always wanted to go back and do a better one since then. In the new image, I want to get across that the witch is good, doing something beneficial or wholesome. Or at least fun/silly. It could be reviving dead soldiers on a battlefield, or the sillier option, turning cats into people and/or people into cats. There's a little more to this second idea but I'll leave it there for now. Still thinking.
Having an actual good idea for the painting is from Manchess. Why did I need to read that in some guy's blog post? I dunno. I think I'd gotten so used to drawing/painting from ref simply for the purpose of study/laziness I forgot I could do my own painting.
The First Law image is right out, just because I would feel the need to research the characters, armor, and weapons. If I do the witch painting I have more freedom. So I'll go with that and hopefully come back to the other one later.
The next thing to do is to try to thumbnail things properly (Manchess), which I've usually skipped. Once I've found an idea w/ the witches that I really like.
Not too much other new stuff, because I went back to studying Blender hard surface for a bit so I can make better weapon models in my game project. Leyendecker study, and a random few others.