Martin M
Estonia
Used to draw in high school. Picking it up again now at the age of 35. Started taking courses online September 2023. Super excited!
Albums
Activity Feed
Mounir B
•
19d
added comment inProject - Dynamic Shapes
Asked for help
Is there something wrong with the seals file ? I can't download it
I have the message below
Martin M
18d
It downloaded for fine for me.
Try to log out and log in again and then navigate to this lesson again.
Daniel
•
21d
Asked for help
I chose a garage since I figured I could put a lot of boxes inside it. I also tried to do some cylinder shapes to figure out how they would work with perspective. Open for feedback!
I don't see any convergence issues. Good job!
I think some lines could be darker and some lighter.
It's good to see the x ray vision for this exercise but some of the "hidden edge" lines are darker than some of the "visible edge" lines.
For example the rear edge of the lowest shelf should be visible between the two boxes. A darker line there for the visible part of edge would really make the self feel 3D.
Styrbjörn Andersson
•
24d
Asked for help
Here is my attempt at the dynamic seals/sealions. I tried to focus less on accurate proportions and more on getting some kind of flow.
You have understood 1 point perspective!
There are just a few lines that you likely drew based on other lines near it. Like the lines for the shelves are all almost parallel to each other, while they should be converging to the vanishing point.
I think you understand the theory and just rushed a bit with those few lines.
Just one of the boxes seems to converge to a separate vanishing point.
Otherwise good job!
Martin M
•
24d
Asked for help
I tried to spend less time on setup and more on the lessons content this time. So I grabbed an older grayscale project I had and started to lay in the flats.
In hindsight I should have spent more time with the flats I think.
Something that helped me with painting the shadows layer: TURN OFF PRESSURE SENSITIVITY!
You want to paint in the shadow areas with consistent opacity and leave gradients only for the soft edges.
I kept painting over and over to fill in the shadow area evenly as my pressure sensitivity required me to use many strokes until an area was filled. This meant that it was difficult to get a clean hard edge and difficult to get a consistently smooth gradient.
My painting became A LOT LESS MUDDY after turning off pressure sensitivity and keeping it at 100% and focusing on drawing the correct shadow edge with as few strokes as possible.
I'm not happy with the result but I feel like I am finding the sources of my frustration and tackling them one by one.
Martin M
•
25d
Small detail about Nightshade: it's purpose is to "poison" the datasets that are used to train new models. So using it on a trained model might not yield any noticeable results as the model has already "learned" enough about art.
But if enough artwork online would be covered with nightshade then this will make it harder to train new models and advance the AI. At least in theory.
It is an arms race and one side has A LOT more funding and brain power invested. So I share your feeling that Glaze and Nightshade will at best only slow down their work a bit, but will not provide a guaranteed defense.
My personal hope is that the whole AI wave is still just new and novel.
Each 6 months a new version gets released and it looks better than before.
Kind of like computer game graphics in the 90s and 2000s.
Everyone just wants more eye candy and the whole industry is geared towards pushing for new heights of realistic 3D renders.
By now realism has been achieved enough times that it no longer is the must have. Unique art style or vibe of the game is arguably more influential.
And this is something I feel like the AI has a hard time with.
Consistently working on a new and unique direction.
Ai just gravitates towards the average. Large Language Models will lean towards the average of what previous texts have said before and image generation will lean towards what it has seen before.
This doesn't mean that art jobs will be safe. A lot of demand for single images or advertisements will be moving away from the "human" market.
There will be work for experienced artists taking on larger and more ambitious projects.
But smaller and low pressure jobs and tasks will likely go to the lowest bidder: the Ai
•
24d
Thank you for sharing your thoughts, @Martin M!
"Ai just gravitates towards the average [...] towards what it has seen before." - This bit, in particular, called my attention! It makes a lot of sense with what we're already seeing: average art jobs that require "same old" solutions look like to be in the most danger of replacement, while a smaller percentage of somewhat "truly innovative" projects will keep requiring human brain force. It's still a very tough change, though, since the majority of art jobs in a pre-AI era were exactly those more ordinary projects, and it's a sad and very real social problem that there will be a much smaller room for the "everyday artists" regarding jobs, especially for those who were still trying to break into the industry.
On the other hand, I think things can gain some perspective if we consider art-making as a broader human practice and don't restrict "being an artist" to industry jobs or commercial profitability. In this scenario, perhaps it's easier to see AI for what it really is: a technical artifice, rather than some sort of "evil menace"? Perhaps this moment of change in the art industry could be an opportunity for people (and artists?) to start looking more at art not only as a job, but as a sort of human existencial practice than can connect people to one another, something we do because we do, not (just) for money, and something that can make our particular lives and stories resonate with everyone's shared humanity (and I guess this a place AI might never reach). Is this too romantic or naively optimistic? Just some loose thoughts... which, of course, don't reveal the tragedy of real world artists having to find other ways to pay their bills, and also don't exclude the urgent need for updated legal systems and practices worldwide.
But, yeah, regarding the "digital protection" concern, I think your words sum it up pretty well.
Thanks again!
Martin M
•
26d
I have spent 3 evenings on this. If the core of the assignment is to learn layers and different associated tools then great - that only took like 40 min of practice.
But if the goal is to produce a good grayscale painting study then there's over 10hours of messing with brushes and I gave up before getting to the details like the vail and beads.
I think I would benefit from doing some material studies with digital medium. Like fur, clothes, skin and how to paint it efficiently. My main struggle is with brushes as I can't land correct opacity and edge. I keep going over the area with many strokes to get it darker but that makes the edge muddier. I guess this is because I haven't figured out the correct brush to use. Overall feel frustrating trying to follow along and not getting the same efficient mark making as in the demo videos. At least this time there was the layer window visible but I would have really needed to see what kind of brushes were selected or modified. I know it is a newbie question "What brush are you using?" but I really struggle to make the same marks on my canvas as I see in the demo.
Ariana Peralta
•
1mo
Asked for help
Hi everyone, I'm not able to download the practice files. Every time I try, I get an error message. Is there anyone who can help me, please? :) Thank you!
What kind of an error message do you get?
The assignment files are in a zip format. So after downloading you will need to first unzip the container before you can use the images.
I don't understand what inverse is doing
Selection tools allow you to select an area. There are many selection tools: rectangular, circle or in this case the Magic Wand tool.
The Magic Wand tool tries to select an area that you clicked on. In this video he selected an area outside of the character. The selected area is indicated by the black and white border line that appeared around the character.
If he were to use the paint bucket tool to fill the area it would have painted everything around the character to that color.
But he wanted to create a silhouette of the character so he needed the "inverse area" of the background.
Thus he used inverse to invert the selection area.
It's hard to notice because nothing changes visually as the black and white border of the selected area remains the same, just different side is selected.
Now using the paint bucket tool fills the characters area, not the background.
You might ask: But why didn't he immediately just click on the character with the magic wand tool to get a selection of it?
The magic wand tool only selects similar pixels into one selection. So it would work if the character was all painted in one color. But he already painted in different values thus the Magic Wand tool would not select the entire character, only the single value area. That's why digital artists use this hack of selecting the background and inverting the selection to get the entire area of the character.
