Compose styles
Rather than blending styles toward a middle, this gives them roles. The first one carries the structure, the next comes in as the accent, and anything after that shows up in a single detail. Where they fight each other, it says so before you start.
How to divide the work
- Composition
- Follow DeepDream: Preserve the source's large composition and amplify related features at several scales to create recursion into depth
- Type
- Set in DeepDream's manner (Add type after processing, keeping it short, bold and simple enough not to fight the detail), and let Generative Art's lettering (Letters handled as coordinates and repetitions with the face itself fixed) appear only where you want the eye to catch.
- Material
- Ground in DeepDream's material (Choose layers in a trained vision model and iteratively increase their activation through gradient ascent); bring in exactly one thing from Generative Art (Lines assumed at a plotter's constant weight, tone made by overlap density).
- Colour
- Build on #21134A, #2FD5C4, #F0C84B and admit one accent from #fbfaea, #7a4ab8, #2c2e18.
Where they fight
- DeepDream and Generative Art both belong to Digital Art, so the difference is small. Push the roles further apart than feels comfortable.
- DeepDream and Generative Art are techniques rather than looks. Decide which one runs last in the process, because that is the one that will show.
Caution
- DeepDream It is not a swirl filter or pasted dog faces. Preserve which layer and feature were amplified and how they bind to the source.
- Generative Art A random range set too wide widens the gap between good and bad results and leaves a pile of output instead of a work.
Image prompt
Produce four original images that take their structure from DeepDream (technique, 2015–) and their accent from Generative Art (technique, 1965-). ## What to make Make the kind of thing DeepDream exists for: exhibits and critical images that make machine misrecognition immediately visible, or music video and cover imagery that keeps a source landscape beneath a multi-scale hallucination. Choose one subject and hold it across all four. ## Structure comes from DeepDream - Eyes and animal faces repeated across the frame - Swirling texture multiplying along source contours - Multi-scale repetition that reveals new figures when viewed closer - Saturated blue, yellow and violet rolling continuously Composition: Preserve the source's large composition and amplify related features at several scales to create recursion into depth. Type and lettering: Add type after processing, keeping it short, bold and simple enough not to fight the detail. ## Accent comes from Generative Art, used sparingly - One simple element repeated to fill the field - Variants of a single rule shown side by side - Repetition drifting in angle and position - A density and precision no hand could reach Let one material quality come from it: Lines assumed at a plotter's constant weight, tone made by overlap density. ## Colour and feeling Build the ground on #F0C84B, carry the structure in #2FD5C4 and #21134A, and let a single accent come from #7a4ab8. These are read off real works, so a direction rather than a fixed palette. Feeling to carry: technique, futurity, exhilaration, calm, play. ## Where they fight - DeepDream and Generative Art both belong to Digital Art, so the difference is small. Push the roles further apart than feels comfortable. - DeepDream and Generative Art are techniques rather than looks. Decide which one runs last in the process, because that is the one that will show. ## What goes wrong - DeepDream: It is not a swirl filter or pasted dog faces. Preserve which layer and feature were amplified and how they bind to the source. - Generative Art: A random range set too wide widens the gap between good and bad results and leaves a pile of output instead of a work. ## Across the four Keep the roles fixed in all four. Vary the framing, the scale and the arrangement so the set shows what this pairing can do rather than four versions of one picture. ## Constraints - Do not imitate, reference or name any specific artist, studio, brand, character or copyrighted work. - Do not reproduce the composition of an existing work. - Produce original compositions in the general period languages described above.
Related entries
- DeepDream 2015– / Technique / Digital Art
Google's visualization technique repeatedly amplifies features learned by an image-recognition network, making eyes, animals and towers emerge excessively from clouds, rocks and leaves until misrecognition becomes a recursive hallucinatory surface.
- Generative Art 1965- / Technique / Digital Art
A way of working in which the maker writes only the rule and the randomness, leaving the drawing to the machine. As in Nees's plotter work, the result is shown as several variants born of one rule.
Share this composition:https://indexstyle.org/compose?ids=deepdream+generative-art