HOME | DD

Description
Fusion of Figure, Type, and Forms. Combining forms and typography from Illustrator to drive image generation using Control Nets in Stable Diffusion.

gymdreams8.com/blog/typography…

You call them chav, I call them scally. Shaved head, tracksuit, sneakers, lean, and always with a look that doesn’t seem to care. I decided to use “Chav” here because I wanted a keep the word short enough (something 4-5 characters) so that the letterforms will be in proportion with the size of the images. For example, the letters in these examples are similar in size to the head of the figures, and as such gives proper balance to the composition.

I made these using the same technique as the previous two sets — combining a design that I made inside Adobe Illustrator and feeding that into control nets to drive the composition, color, and style. Stable Diffusion breaks up the figure into fragments at each of the lines, thus creating natural divisions of forms that are energetic, realist, and abstract at the same time.

As I wrote previously, combining type with AI image generation — while being able to keep the type so legibly and clearly — is a huge breakthrough for me. Previous attempts purely from using text prompts alone will render legible type, but the spellings are always off, making it less useful for typographic designs.

There are only two things that matter to me in design — typography and the grid. They are the foundations to everything that I do. It likely would be meaningless to most of you, but ask any graphic designer, and they would be able to tell you why this series has such profound meaning to me, in more ways than you could ever imagine.

24 images on gymdreams8.com

Tech: Text prompts with Stable Diffusion: 30 steps, DPM++ SDE Karras, CFG 5, 768x768, Airfuck’s Wild Mix v1, Denoising 0.5. ADetailer 1 mediapipe_face_full, ADetailer 2 hand_yolov8n.pt. CN0 Canny. CN1 Shuffle. Hires 1.5x (1152x1152), 15 steps, 4x_NMKD-Siax_200k. Post: Topaz Gigapixel HQ 2x (2304x2304), Adobe Lightroom post-processing.

Elsewhere

- www.instagram.com/p/Cvu_lQySiB…
- twitter.com/GymDreams8/status/…

Description
Fusion of Figure, Type, and Forms. Combining forms and typography from Illustrator to drive image generation using Control Nets in Stable Diffusion.

gymdreams8.com/blog/typography…

You call them chav, I call them scally. Shaved head, tracksuit, sneakers, lean, and always with a look that doesn’t seem to care. I decided to use “Chav” here because I wanted a keep the word short enough (something 4-5 characters) so that the letterforms will be in proportion with the size of the images. For example, the letters in these examples are similar in size to the head of the figures, and as such gives proper balance to the composition.

I made these using the same technique as the previous two sets — combining a design that I made inside Adobe Illustrator and feeding that into control nets to drive the composition, color, and style. Stable Diffusion breaks up the figure into fragments at each of the lines, thus creating natural divisions of forms that are energetic, realist, and abstract at the same time.

As I wrote previously, combining type with AI image generation — while being able to keep the type so legibly and clearly — is a huge breakthrough for me. Previous attempts purely from using text prompts alone will render legible type, but the spellings are always off, making it less useful for typographic designs.

There are only two things that matter to me in design — typography and the grid. They are the foundations to everything that I do. It likely would be meaningless to most of you, but ask any graphic designer, and they would be able to tell you why this series has such profound meaning to me, in more ways than you could ever imagine.

24 images on gymdreams8.com

Tech: Text prompts with Stable Diffusion: 30 steps, DPM++ SDE Karras, CFG 5, 768x768, Airfuck’s Wild Mix v1, Denoising 0.5. ADetailer 1 mediapipe_face_full, ADetailer 2 hand_yolov8n.pt. CN0 Canny. CN1 Shuffle. Hires 1.5x (1152x1152), 15 steps, 4x_NMKD-Siax_200k. Post: Topaz Gigapixel HQ 2x (2304x2304), Adobe Lightroom post-processing.

Elsewhere

- www.instagram.com/p/Cvu_lQySiB…
- twitter.com/GymDreams8/status/…