mastodon.gamedev.place is one of the many independent Mastodon servers you can use to participate in the fediverse.
Mastodon server focused on game development and related topics.

Server stats:

5.1K
active users

#actiniaria

3 posts2 participants0 posts today

More #Actiniaria progress: Added a soft global constraint to create more of a (macro)organism feel w/ a sense of belonging and fuzzy boundary. Also slowly updating the behavior/relationship matrix between the six different types now, to create varying temporary alliances...

(Note: Sadly Firefox doesn't respect the Rec2020 color profile in the video, please download or use Chrome or Safari for viewing...)

Found some time last night to implement multi-behaviors and I'm very excited about where this is going... 😍 The video shows 6 different types interacting with (and avoiding) each other. Next step is to vary the behavior matrix over time, causing changing alliances and breakup behaviors...

See #Actiniaria for more context...

(Note: Sadly Firefox doesn't respect the Rec2020 color profile in the video, please download or use Chrome or Safari for viewing...)

Replied in thread

@akkartik @ashley @datarama I'm not fully sure which magic constants you're referring to, but the ones used in this example[1] come from fine tuning the desired behavior and tweaking the balance of the main three different force contributions (separation, alignment, cohesion). Sure, you can give them all equal weighting, but you can achieve so many more varied & interesting behaviors by simply playing with these force multipliers (and throwing other constraints/forces into the mix). In the thi.ng/boids implementation these forces are customizable per boid (e.g. to create groups of different species/behaviors) and can also be modulated spatially (based on current position and a possibly underlying data layer/map). That way I can dynamically force different behaviors based on where an agent is in the simulation space/time. The combination of all these provides incredible freedom and allows me use boids in situations you wouldn't normally associate with the cliché flocking behaviors...

ICYMI (related), check out my more recent boids explorations here:
mastodon.thi.ng/@toxi/tagged/A

[1] github.com/thi-ng/umbrella/blo

thi.ng/boidsn-dimensional boids simulation with modular behavior system

Similar to my #LinearMemory piece from last year (check the hashtag for references), these current #Actiniaria explorations make me fully appreciate all the research, efforts and supporting hardware behind the DCI-P3 & Rec.2020 color spaces. The intensity (and subtlety) of some colors & combinations are just stunning (to me), and I don't mean this in any ableist way (just enjoying whilst I still can!)...

Also still amazed that something like this runs at a smooth 60fps @ 2160x2160 UHD resolution on a mobile device with a just Snapdragon 2 chipset... As part of my advisory role @ day job, over the past few months I've been tasked with optimizing artworks for almost a dozen generative/algorithmic artists, often achieving 2-4x framerate improvements... Very much feeling a longer blog post re: code craft vs code art coming up, hoping to connect them more...

"To understand a program you must become both the machine and the program." — Alan Perlis

(Note: Sadly Firefox does not seem to respect the Rec2020 color profile in the video, please download or use Chrome or Safari for viewing...)

Work in progress...

Finally finding some time to continue working on my earlier thi.ng/boids experiments, now also using WebGL instancing and floating point render textures for super-smooth blending/trails to create aesthetics reminiscent of sea anemones...

(Update: I've decided on the beautiful-sounding and fitting #Actiniaria as project title...)

(Note: Sadly Firefox does not seem to respect the Rec2020 color profile in the video, please use Chrome or Safari for viewing...)