Jump to content
eetu

Unofficial Challenge: Boxified images

Recommended Posts

Every Houdnik loves boxes, right?

 

Let's try and create pretty pictures along the lines of these beautiful renders from Lee Griggs:

 

post-2678-0-24164300-1408016394_thumb.jp

post-2678-0-25659200-1408016820_thumb.jp

 

You can see many more at his web page.

 

The idea might look quite trivial to the more seasoned users, but there is much room for twiddling in getting things to be visually pleasing. Or - create your own twist on this theme.

For the beginner it should be a good lesson in Copy SOP or instancing.

 

Attaching your hip files is encouraged, but not mandatory.

  • Like 2

Share this post


Link to post
Share on other sites

first attempt ...

 

is the adaptive sizeing, the brighter the bigger,  in the original renderings achived without overlapping ? how would you go for this ?... (i know, it's a challange here :)

 

EDIT: Some more (Spam:)) ... I also used copystamping + bevel this time. The first couple of images are setup with polyextrusion.  Inspiring work Tom! Maybe you should bring up your BulletSop stuff in here?

 

 

cheers

post-4387-0-72066000-1408036821_thumb.jp

post-4387-0-28889800-1408036838_thumb.jp

post-4387-0-19469800-1408036852_thumb.jp

post-4387-0-58355300-1408117676_thumb.jp

post-4387-0-46090900-1408904941_thumb.jp

post-4387-0-73188700-1408904967_thumb.jp

Edited by hatrick
  • Like 4

Share this post


Link to post
Share on other sites

I gave it a quick go, done with packed primitives and Copy SOP. Each box is beveled and has 150 polygons; in the image with the smallest boxes there is around ~80 million polygons - which Mantra handles with no problem at all thanks to packed primitives. 

I just did some variations on the same thing, and tried out Jims disney brdf for the shader. I'm dissappointed if you don't recognise the underlying image :<

 

I'll share my hip later if you want, at this early stage of the challenge I think it is a bit of cheating. Although it is pretty straightforward already, and fun!

I'm not supersatisfied with the renders, but that's it for today. Did a quick slapcomp as well.

 

post-9276-0-03452400-1408143242_thumb.jppost-9276-0-26234000-1408143254_thumb.jppost-9276-0-33979800-1408143266_thumb.jppost-9276-0-36881000-1408143278_thumb.jppost-9276-0-46378200-1408143289_thumb.jp

  • Like 5

Share this post


Link to post
Share on other sites

I gave it a quick go, done with packed primitives and Copy SOP. Each box is beveled and has 150 polygons; in the image with the smallest boxes there is around ~80 million polygons - which Mantra handles with no problem at all thanks to packed primitives. 

I just did some variations on the same thing, and tried out Jims disney brdf for the shader. I'm dissappointed if you don't recognise the underlying image :<

 

I'll share my hip later if you want, at this early stage of the challenge I think it is a bit of cheating. Although it is pretty straightforward already, and fun!

I'm not supersatisfied with the renders, but that's it for today. Did a quick slapcomp as well.

 

attachicon.gifboxified_dv.0001.jpgattachicon.gifboxified_dv.0002.jpgattachicon.gifboxified_dv.0003.jpgattachicon.gifboxified_dv.0004.jpgattachicon.gifboxified_dv.0005.jpg

 

I would love to see a hip file for this as I'm struggling to get quick/efficient renders out of Mantra for my version. I'm using instances right now because I couldn't figure out how to color the packed primitives. I'm sure I'm missing something basic here.

Edited by JonathanGranskg

Share this post


Link to post
Share on other sites

I would love to see a hip file for this as I'm struggling to get quick/efficient renders out of Mantra for my version. I'm using instances right now because I couldn't figure out how to color the packed primitives. I'm sure I'm missing something basic here.

I like to control everything on the points before copying. So I use the vector attribute "scale" to control the size of the boxes, and then Cd for color as usual. Then I copy the boxes on to those with Copy SOP and "pack before copying". They should get the right size, and I make sure each packed primitive gets the Cd attribute copied over as well (point or prim doesnt matter). Normally the render cant see this Cd attribute, but if you in the shader use a Render State VOP and type "packed:Cd" you can access it. So plug that into color and you are good to go :)
  • Like 1

Share this post


Link to post
Share on other sites

here my boxified image. the colors are bit off...but some nice complex structucers going on :)

simple copy sop work based on image...on top of it I traced(trace sop) the image. I try to getting some nice cuves to control the flow  of the boxes (orientation - attrib transfer the normals to the template points).

 

highres_version

https://dl.dropboxusercontent.com/u/34142573/boxified_world_01.jpg

post-10198-0-06621800-1408374143_thumb.j

Edited by tfreitag
  • Like 2

Share this post


Link to post
Share on other sites

...sorry for spam. but here is another colorful try with pretty high resolution. this time i tweaked the shader. the pure render comes out bit flat, so i did some post work onit.

its mantra-pbr with little DOF...and some more DOF in post.

 

10k-version :)

https://dl.dropboxusercontent.com/u/34142573/final_colorbomb_render_b08_CC%20(0-00-00-00).jpg

 

post-10198-0-23824000-1408532469_thumb.j

  • Like 6

Share this post


Link to post
Share on other sites

hehe...now i am challenging with my self :blink:

but i felt in love with Lee Griggs pics...the boxy things are magic :)

put some more stamps in copysop network. now i am controlling with colorintensity a blend sop ( blending between box and sphere). and  to avoid regular grid pattern i remesh the imageplane with controlling the targetedgelength. of course its really simple(but chaotic) setup, but if somebody is interested in i would upload the hipfile.

highres-link:

https://dl.dropboxusercontent.com/u/34142573/cc%20(0-00-00-00)_rockforam.jpg

post-10198-0-64739400-1408905378_thumb.j

Edited by tfreitag
  • Like 4

Share this post


Link to post
Share on other sites

Great stuff... Yes i am indeed interested in your boxmagic.(seen the updates in my first post?) Not shure how you ve setup the bevel. Is there some crease adjustment for the edges ?

 

the attribute - blend thing is really cool ... let's bring this further!

post-4387-0-33356100-1408913049_thumb.jp

post-4387-0-82232400-1408913079_thumb.jp

Edited by hatrick

Share this post


Link to post
Share on other sites

Great images, everyone!

 

This is also reminiscent of the pinboard effect from the first X-Men (2000) movie except with color. The effect in the movie was done in Houdini.

 

2008-x-men-3d-display.jpg

 

 

It also bears some resemblance to the greeble challenge as well.

 

I guess the next stage of the challenge is to see if we can do some interesting looking animations.

Share this post


Link to post
Share on other sites

Hello people!

 

This is my test, hope you like. Now I am working on improving it!

 

Looking forward to see other result from you!

 

post-10736-0-97672500-1408920865_thumb.p

Share this post


Link to post
Share on other sites

now I tried to rendering a tons of polys...without much success regarding ram-issues.

..first i tried packed primitve (checkbox copysop)....the bgeofile is about 15% smaller then unpacked geo.. when render it ( about 80millions polys) it fillup my 32gb ram and swapping..and swapping

second try...polysoup...the bgeo becomes bigger (ig guess its not uncommon in this case)..its not an option

 

I guess packprim efficiency (copy sop) depence on how the copy sop network is buildup (?) My setup i using a blend-shape...so its make every copy unique which is hard to pack(?)

 

hipfile is attached...maybe somebody can give me advices how to managed lots of polys.

 

thanks!

 

layer_boxes_v05.rar

  • Like 1

Share this post


Link to post
Share on other sites

Thanks for the Challenge Eetu! Even though i cant afford much more time on this challenge, I definitely learned alot trying to complete it. As for method, im really excited to see what route everyone took, as I dont feel completely satisfied with mine. Im wondering how much if this assignment can be resolved solely in vops. I do have a few questions in mind but i think ill hold on to them till the challenge is over. I dont want to give too much away. Ill be more than glad to share my Hip file on the thread later. But for now, Good Luck and copy on.

 

post-8804-0-93234900-1408991945_thumb.jp

Share this post


Link to post
Share on other sites

I guess packprim efficiency (copy sop) depence on how the copy sop network is buildup (?) My setup i using a blend-shape...so its make every copy unique which is hard to pack(?)

 

Sorry, I haven't had a chance to look at your file but if you have unique copies, then there's not much chance for efficiency because those shapes cannot be shared. The easiest way if you just have colored cuboids all of the same size/shape is to instance a polysoup box onto points that have varying heights/orientations.

 

See these pages in the docs for more info:

http://www.sidefx.com/docs/houdini13.0/copy/varying (on how to vary shader parameters across instances)

http://www.sidefx.com/docs/houdini13.0/render/renderingmanypolys (more info, you've probably already seen this)

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×