>>7779365
Looks cool, I hope you draw a pumpkin knight or something similar

mogs me

>>7779365
you are ready

>>7779365
So how do you actually go about selecting the colors/palette and doing the layering?

>>7779756
So how do you actually go about selecting the colors/palette and doing the layering?

By not being a faggot ass bitch and picking the right colors. Even a child can pick a color what are you doing with your life?

>>7779777
based phonechad

>>7779365
very solid work, and good picking of color assuming you didn't directly color pick. only real critique is to avoid how you did that background, which is unfinished
bokeh in painting is divisive, and i personally don't like using it when im doing backgrounds even if it's present in a photo study, but if you decide to include it, it's probably good to switch to airbrush to soften the edges. your compositional subject is the pumpkin, but the fence has sharp enough edges in the thumbnail that it brings it a lot closer than in the reference. the fence lacks details, which is good considering its a background object, but the repeating vertical pattern along with its sharp edges brings more attention to it than it should have.
tldr:
1. don't be afraid to omit or replace things from photo reference, especially backgrounds.
2. find an airbrush/soft edge brush you like using
3. bokeh is a purposeful photography effect for photography compositions, and if you decide to omit it, you should change or modify the background that the bokeh is softening because the background will feel too sharp and closer than in the reference.

i'd comment on minor feature placement inaccuracies between study and ref, but the study is executed well enough that they don't matter. besides, it's a jack-o-lantern so symmetry and feature placement aren't as important as it would be on a real face, so "who cares"

>>7779852
I did color pick, which i feel like i should prolly tone down in future studies, but i wanted to get as accurate as i could

background was supposed to be very simple, wasnt the main focus and i generally like the idea of simplifying detail, but still a fair criticism

on the topic of feature placements, im in two minds about it:

A. if i try following the reference too closely im gonna drive myself mad, so i more or less went for the same "effect" of the refernece as close as i could. if that makes sense

B. the one thing i think i really messed up with the features is that i realized the pumpkin looks less round than the photo. the features just feel a bit too 'flat'

tyy for your feedback though

>>7780032
>I did color pick

You're DISGUSTING. I vouched for you, and you tell us this?! Don't you have any shame? Some people were just not raised right.

>>7780058
I NEED to kill myself now

File: soldierpractice.png (3.21 MB, 2028x1400)

3.21 MB PNG

wip second practice thing, no color picking this time

>>7780032
Next time
>lower the opacity to about 20% for your ref
>in a new layer do a light lineart pass, doesn't need to be detailed, just map the relevant bits (pumpkin, eye and mouth cutout, the fence lines, etc).
>new layer underneath, map the colors your perceive as flats first, if you can't do it, either watch a basic video on color temperature and value (try some Sinix tutorial on value) or make your best guess
>doesn't need to be the exact hue/saturation, just make sure the values are close and the relationship between these hues follows the logic of your reference
>create a new layer on top of the flats, pick the bucket tool and fill it will black, then make it a "color" blending mode layer
>it's not exact but it'll give you the ballpark of your values and how far off you are from the ref, you only make it visible to compare
>keep adjusting your colors and values until it looks close enough to the ref
Don't color pick, it makes your study virtually useless aside from being able to fill a coloring book digitally. Normally you wouldn't go with colors from the get go, a simple way to learn values is by using the color blending mode method for your ref (if colored, you can also just get a ref in grayscale), then pick 3-5 grayscale values (white to black) and then force your eyes to group values as best as you can, even if some are left out, it forces you to separate them visually, squinting helps here, it's also a good idea to use a hard round brush for flats, no blending to keep things clear, blending only makes it look "prettier", but if the value grouping doesn't work, nothing will fix it. Starting with color will generally throw you off a bit at the start so you can avoid it or learn about it as you go with your studies/tuts.

>>7780559
Already started on a new study but ill keep your advice in mind. youll be glad to know im going 100% without coilor picking for this one