Knight's sketchbook

[dimensional-knight]

I want to know, though, how do you go about drawing/measuring?

I don't. *prepares to dodge incoming flying objects*

This is a good question. I can't vouch for the answer since I don't fully understand it yet either but a thing is certain: It's heavily perceptual.

I'm not really outlining anything, I'm representing super angled planes and shadows. Everything is done in the service of volume, and the initial faint shading I sometimes do could be described as a sonar ping. It's a way to discover how far and in which directions an area stretches from the original point A where the pencil was. This is one of an umbrella of surface-seeking methods: Casting wavy lines into the nothingness trying to find the silhouette and angle of something is another one; so is sketching small bundles of volumes like cheeks. The funny thing is that I never set out to find the surface, I'm either holding the place for a detail so I can judge the entire object dimensions or evaluating the perspective or whatever, yet looking beyond the immediate goals all of these approaches are ultimately about surface.

I know if it's accurate enough if the resulting object has the "weight" I expect to see, not quite angles in the silhouette but approximate dimensions and mass feel and center of gravity.

I instinctively further check it using what I know about the subject. Faces are relatively easy: I know many things about them. I know what possible glabellas this forehead inclination and eye depth and that nose type can create, and where eyebrows peak and turn, the ways chins connect to mouths and more. When a drawing doesn't have room (or has too much room) for something I'll know immediately a mistake was made.

Yeah, I'm not drawing what I see. I pretty much only draw what I see when I can't understand what I'm seeing, and in most cases it's possible to at least understand what a surface is doing—if it's a depression or protrusion—even you don't know what the hell you're looking at.

Two interesting data points here: I'm used to drawing from imagination. That's pretty much all you do as an illustrator and concept artist. I'm also extremely weak at conjuring mental images. That's why I started to draw, I can only see the things in my mind once I create them, otherwise they're just words and unsubstantial close-ups. Plain line art drawings are pretty but confusing. I can't project volume into a line drawing and every distance I measure by eye is either by unfocusing and judging the negative spaces between blobs or by tracing the distance with my eyes.

Sooo, in short this is handy for tight deadlines and imagination stuff. It has a few pros:

- Solid-looking objects
- Easier to light up dramatically as you have all planes info needed
- Fewer passes when you know your subject: Your "sketch" is part of the final shading
- Easier to rotate stuff, maybe? I don't know how hard it is from the POV of someone used to a flattened 2D approach. It's not easy but tends to look structurally sound
- Easier to stylize and enhance mood, you just branch out from a same set of base assumptions about the subject, playing up or down stuff at will

Cons:

- Very distortion-prone. We all got biases (eg I pull things to top left and tend shy away from edges of objects). Those little mistakes build up as you progress and without abstract/distant points of reference checks you either compensate by doing things like slightly rotating the subject so the shape still holds or do more passes to keep it from falling apart
- Flimsier accuracy checks for things outside your visual library
- Doesn't cultivate eye accuracy in general, not for angles and distances nor for absolute values


It's no wonder I struggle with 1:1 reproductions. In any case I'm raiding other sketchbooks for tips on books covering developing eye accuracy etc.

---

Minor update!

Two studies angling for position accuracy. Digital paintings since the goal is stressful enough without me fumbling with the medium. I mixed and matched my usual approaches with conscious long-range checks (when I remember to perform these...) and while I don't know if I'm internalizing them yet it's an interesting change of pace.

Don't mind the hues, values not hues were the secondary goal:




The reinterpretation of light to fluff up the atmosphere was on purpose, but the change in angle and rotation and even a bit of the FOV are most definitively not. 




No idea if the time lapse reveals anything useful but here it is: