demo: placeholder image generator #23

Open
automata opened this Issue May 23, 2014 · 6 comments

Comments

Projects
None yet
3 participants
@automata
Member

automata commented May 23, 2014

First study for a placeholder image generator using bezier curves only. Some samples:

captura de tela 2014-05-23 as 18 32 47

@bergie

This comment has been minimized.

Show comment
Hide comment
@bergie

bergie Jun 6, 2014

Member

@automata looks very nice! We need to be able to store stuff like this as high-res images

Member

bergie commented Jun 6, 2014

@automata looks very nice! We need to be able to store stuff like this as high-res images

@automata

This comment has been minimized.

Show comment
Hide comment
@automata

automata Jun 8, 2014

Member

Thanks @bergie! I guess we can export the canvas as a blob on the runtime.

Now we can use a string as a seed to random numbers, creating the same image for each string (using noflo-seedrandom).

Some examples using bezier curves emerging from figure center:

generative01 generative02 generative03 generative04 generative05

Those ones are more interesting as "paintings", covering the whole canvas without a reference point:

generative06 generative07 generative08

Which approach is more interesting for placeholder images @bergie, @forresto?

Member

automata commented Jun 8, 2014

Thanks @bergie! I guess we can export the canvas as a blob on the runtime.

Now we can use a string as a seed to random numbers, creating the same image for each string (using noflo-seedrandom).

Some examples using bezier curves emerging from figure center:

generative01 generative02 generative03 generative04 generative05

Those ones are more interesting as "paintings", covering the whole canvas without a reference point:

generative06 generative07 generative08

Which approach is more interesting for placeholder images @bergie, @forresto?

@automata automata closed this Jun 8, 2014

@automata automata reopened this Jun 8, 2014

@forresto

This comment has been minimized.

Show comment
Hide comment
@forresto

forresto Jun 9, 2014

Member

The full-screen ones would fit better, I think.

We should experiment with inputting a primary color, and doing transforms
on that, or an array of colors.

Maybe a transform-color module would be good, but make-color outputs
strings, which makes that more annoying. Otherwise could just take hue +
180 etc.

I guess
css string -> HSL(A) object
component would be helpful.

Member

forresto commented Jun 9, 2014

The full-screen ones would fit better, I think.

We should experiment with inputting a primary color, and doing transforms
on that, or an array of colors.

Maybe a transform-color module would be good, but make-color outputs
strings, which makes that more annoying. Otherwise could just take hue +
180 etc.

I guess
css string -> HSL(A) object
component would be helpful.

@automata

This comment has been minimized.

Show comment
Hide comment
@automata

automata Jun 9, 2014

Member

👍

Another points to consider:

  • theme based (e.g. following proper color palettes and shapes)
  • use existing images (e.g. nature) to synthesize a more abstract shape using Delaunay/Voronoi tessellation
Member

automata commented Jun 9, 2014

👍

Another points to consider:

  • theme based (e.g. following proper color palettes and shapes)
  • use existing images (e.g. nature) to synthesize a more abstract shape using Delaunay/Voronoi tessellation
@automata

This comment has been minimized.

Show comment
Hide comment
@automata

automata Jun 17, 2014

Member

More studies (in higher dimension, using node-canvas).

like4
like3
like2
like1

Member

automata commented Jun 17, 2014

More studies (in higher dimension, using node-canvas).

like4
like3
like2
like1

@automata

This comment has been minimized.

Show comment
Hide comment
Member

automata commented Jul 1, 2014

@automata automata referenced this issue Jul 4, 2014

Closed

transform-color #29

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment