We’re in a blizzard here in Massachusetts. Fortunately the power is still on, but we ain’t going anywhere for a looooong time.

Settling in.

I decided to try painting another portrait (my own) based on a photograph.

Here’s the inspo pic from last fall:

Here my painting:

Portraits are hard! This makes me want to take another class with my teacher from last fall. She was good at portraits, but alas, I have registered for a different (cheaper) class, closer to home, where we will focus on “cute baby farm animals.”

ChatGPT gave me some solid feedback on the portrait though:

Lightly mapping planes (forehead plane, side planes of nose, cheek planes) before committing to color would increase dimensional accuracy.

The biggest difference between the painting and photo is contrast and value structure.

The photo has stronger: Shadow under the brow ridge Shadow on the right side of the face (viewer’s right) Cast shadow under the nose Subtle shadow under the lower lip

In the painting:

Midtones dominate. Shadows aren’t quite dark enough, especially around the eyes and under the cheekbones. This reduces form and depth.

Fix: Push darks 15–20% deeper in:

Eye sockets Side of nose Under cheekbone (viewer’s right side) Neck shadow under chin

That alone would dramatically increase realism.

Biggest Growth Area:

Value structure and form modeling

If you deepen shadows and increase plane awareness, this would jump from a good portrait to a very strong one.

OK, I can take that criticism (nicer than some human art teachers I’ve known) and I agree with it.

Related post:

ChatGPT as Art Teacher