File: IMG_6415.jpg (92 KB, 813x749)

92 KB JPG

Managed to score in the 70s (previously I had been in the 40s) after trying repeatedly, but I think that’s just because I started to memorize the game’s built-in patterns. It always shows you the nearest face and it only does so many angles. Not sure if my sense of 3D is actually getting better or not

>>8029778
>Specifically I want to find the line that’s perpendicular to the given plane

There are infinitely many lines that are perpendicular to that plane, and though they are all parallel when considered "in perspective", on the surface of the page they are not.

I don't see how you can evaluate the accuracy of your placement when you haven't drawn straight lines. Your poor rating result from the curvature or the failure to meet at a single vertex rather than your intended angular placement. (Assuming that the screenshot is your work.)

The challenge seems to be more of a gimmick than a useful exercise. Two similar angular errors on the surface of the page could represent very different error quantities in perspective, depending on the orientation of the cube.

Also, I haven't really thought through the math, but the singular face reveals two vanishing points. The placement of the third likely varies with viewing distance or some other factor. That is, I don't think a single face limits the result to a singular cube. Many cubes could share that same face.

>>8030034
>there are infinitely many
Yes I’m aware of that. Just need to find one. Ideally one of the edges of the box
>you haven’t drawn straight lines
OP pic is just a screenshot of the game’s homepage, I didn’t draw that, sorry for the confusion
>more of a gimmick
If drawing boxes has any meaning at all then there has to be some way of objectively evaluating your boxes, what do you suggest?
The guy who made this game gives the parameters that define his boxes fwiw, fov and edge length and etc. It’s not just lines in a void with no context

>>8029778
Do you know what vanishing points are?

File: IMG_6418.jpg (31 KB, 913x180)

31 KB JPG

>>8030034
>>8030059
The parameters I mentioned.
The boxes all look “normal” to me, like what you would see if you opened up Blender and looked at the default box at a reasonable distance, I don’t think there’s anything weird or unfair about the game

>>8030065
Yes.
When you have one plane and you’re trying to find another line that’s perpendicular to it, you don’t know what vanishing point those perpendiculars will go to. That’s the whole problem

>>8030068
You can deduce all the vanishing points from just one side of an object, the cube will draw itself if you just follow the vanishing lines. You can tell this is in 3 point because none of the sides are parallel, if this is a randomizer i guess the chances of it not being 3 point are very low.

>>8029778
have you considered that you can just go "close enough" and move on to the next artpiece? im starting to think you people would be better off doing aim trainer or one of those chimp memory test games instead.

>>8030059
>If drawing boxes has any meaning at all then there has to be some way of objectively evaluating your boxes, what do you suggest?
You make a very good point.

When I learned perspective, I did several constructed drawings (with a ruler) of basic forms from different viewpoints. Then, as you might expect, I moved onto buildings, groups of buildings, images that relied quite a bit on orthogonal elements. And then on to using carefully constructed surfaces to define the limits of more organic forms. Time-consuming guidelines, you might say.

Sure, I gained some experience with how to draw elements in various rotations, but the most useful lessons were about where constructions worked and didn't work. Constructed perspective elements too far away from the center of view start to look unusually sharp and angular. Also, it's easy to produce elements that look unnecessarily fat when constructing-- A simple window has many tightly-packed parallel lines. Unless you're very careful and drawing something large, it's tricky to get those lines to intersect others properly. And so you might tend to draw fatter elements that are more fault-tolerant.

The other lessons relate to the cost vs benefit of construction. What elements need a careful measured basis and which elements work best simply filled in by hand. What lines can be left out entirely? When does time spent on freehand detail contribute more to the final image than more perspective layout?

So perspective for me was about where it works well and when careful construction (ruler) is worth the effort. I never did spam boxes.

I could likely take one of those provided faces and draw several convincing cubes with it. Some would appear that I'm standing close, some further, some with different tilts. But I can't (without some calculation) spit out the exact intended cube on the basis that I'm 1.00 meters away at a particular numerical focal length. I don't see that as very useful.

>>8030126
(comment too long)

But if the exercise is helpful to you, I'm happy.

File: 1772502621480598.png (104 KB, 646x658)

104 KB PNG

is this good

>>8030151
no

File: 1770251992750392.png (72 KB, 614x606)

72 KB PNG

>>8030153
heh, *takes off training weights*. how do you like me now?

>>8029778
youd do better to study architectural drafting perspective from a simple beginners point with a basic book first, i think (though you wont use most of it, i imagine, you will get immersed in it for a bit, which would help).

so can you "measure" it, the implied lines? Technically yes i think so... but you have to have key data points to actually lock that info down, which i m not sure are actually included in a representation of single 4 sided plane. So I guess in this example, No.

For example, all orthographic drawing needs TWO views to determine a point precisely in space (to get x,y, and z). Here we have one view and no scale (and are thus missing Z, and yes, im considering a perspective drawings as the proportional and scaling product of an orthogonal drafting system).

Here you have the vertical vanishing point and one horizontal vanishing point. they intersect at the location of the eye (where VVeritacal is 90 degress and where VH1 is 0 degrees or flat). these V points will help determine the lines for the bottom and right most edges.

you need VH2, the missing vanishing point, AND you need the means to locate the forshortened points of intersection for the lines. so what determines the length of the missing lines? what deternines the location along the horizon for VH2?

the shorthand way to estimate its location is the length of the plane you see. the longer the plane, the further out the vanishing point along the horizon. the shorter, the closer.

said another way, by understanding generic perspective geometry as a plan...the vanishing points are found by putting a right angle, sides parallel to the object being viewed - with the center at the location of the viewer, and extending the lines til they hit the picture plane. so you can imagine rotating a right angle around a point, seeing the vanishing points move, and the perspective view changing... try visualizing that.

>>8030255
its not measuring... because how far is the viewer from the picture plan? if you had that info, you could lock the angle of the lines in space, as youd know the precise angle in plan between the Vh1 point and the viewer, youd still need a last bit of info from the plan to determine the precise location of the bottom right corner.

so, having a 'feel' also comes from the experiece of the conditios that determine the shapes in the first place.

>>8030257
>But I can't (without some calculation) spit out the exact intended cube on the basis that I'm 1.00 meters away at a particular numerical focal length. I don't see that as very useful.

lol thats likely what youd need to complete the visualization i described above.!

Ill leave th eutility question aside for the moment, though my off the cuff thought would be, just for visualizing environments.

the space that matters for most drawing is people in space, proportion, and composition.

>>8030255
This is carefully explained.

>>8030059
there's not infinitely many really, it's not going to throw you distorted cubes to figure out. there's really only 1 possible cube given a face and the assumption it's a perfect square.

>>8029778
if you seen VPs i would understand this but im not going to do math just do draw fucking cube...

>>8030360
well, technically, if you pick two extremes for locating the second vanishing point, there are technically infinite possible points on the horizon between them... you need technique and experience to cull the set

>>8030376
not math

>>8030081
I want to stop being shit. I post my /beg/ scribbles and people say “learn fundies”. Well I’m learning them.
In order to draw a Loomis head you need to be able to find the perpendicular of two arbitrary axes in space. Establish the vertical axis running through the sphere, draw a line from the center of the sphere to a point on its surface. There are your two axes (much like the one plane you’re given in the box drawing game, which also establishes two axes). Now in order to draw the brow line correctly, you need to find the line that’s perpendicular to both lines you already drew. Drawing a box and drawing a Loomis head is basically the same thing. All I’m asking is if people have tips that make estimating this critical line easier, or if it’s really just “feel it”. I know that I’m not going to literally measure and do math, so the process has to involve estimation to some degree.

>>8030425
or you could just draw a face

>>8030427
I’m not going to be shit anymore. If nothing else, I’m going to be able to say that I at least know how to draw a fucking box

>>8030431
boxes are complicated
working through the tone and highlights and shadows of a face perhaps are less so, or in a different way.
but yeah, go for it

>>8030431
fr tho
just go draw from life
fuck this shit

>>8030435
Real life things are harder to draw than boxes. Boxes are the easiest volume to draw, that's why it's a good point to start your practice with, every other object will be subject to the same rules of perspective.

>>8030447
no. you can be a master of boxes and not know shit about the complexity of an eye or eyelids in perspective.

thats why Bargues gives up two pages to basic eyes. but even copying well doesnt mean you get space and boxes wont help.


life drawing will.

>>8030451
Life drawing is only possible once you learn to break down complex volumes into simpler ones. Boxes and other simple forms are related to life drawing, otherwise they wouldn't be a thing. The goal of box exercises isn't to draw boxes but to draw everything else. If you go draw from real life directly without understanding how to deconstruct things you will only trace silhouettes, and that isn't what drawing is about.

>>8030478
no.
life drawing is always possible.
you just draw from life.

>>8030478
also. box drawing is literally silhouttes.

>>8030481
>>8030485
Enjoy your tracing, always drawing characters in the same pose as your references and not being able to rotate things.

>>8030425
Ahh, I see your motive now, and it is a good one.

But this game focuses on an arbitrary and particular set of parameters. You're not drawing "the" accurate box. You're drawing the one accurate box that occurs when assuming a specific inflexible viewpoint. You know this already! It's unnecessarily pushing you towards a singular answer out of many possible convincing results.

You could try this experiment: Draw a few copies of the two axes you describe, the vertical line, and the radius. Try to make them similar. Then draw the eye line at a variety of angles, and complete those faces. Some will appear a bit exaggerated or strange. But you'll find that a great range of angles for the eye line "work".

I don't see that chasing a very specific option out of those many reasonable possibilities will firm up your intuition. You'd be better to explore what different points of view do to those lines. What happens if I step back? Forward? Crouch.

But that's for later. Sounds like where you are now is here: Draw your axes and eyeline, and then try to draw a face that is consistent with those choices.

Yeah this game would be way more useful if they used real objects for us to draw a bounding box around. Random shit like fruit. a house, a car, a head, a rib cage etc. Fun, though.

https://sketchfab.com/3d-models/loomis-head-01-3-5eb43d5ad264483e844684abe5205371
This shit would be literally perfect and would solve all my problems if it weren’t for the fact that it’s in orthographic perspective

>>8030425
getting better requires that you understand perspective so you can guesstimate where the lines should probably go. knowing the formal theory gives you a basis for figuring out what feels right without doing all the busy work.
https://files.catbox.moe/fd09wb.mp4
obviously the cube website isn't going to give you enough time to do that, but just understanding where the vanishing points probably are gives you most of what you need, then it's just a matter of cutting off the depth somewhere that feels about right. not too long or too short.

>>8030508
never traced in my life, pal.
nice zinger i guess tho