color added by distance, by frame

[Ernest Burden III]

I've posted about this before, but thought I'd update the subject with examples from a new project.

 

I like to use secondary color in renderings to suggest space and framing. 'Primary' would be the actual color of the object. I add two types.

 

First is framing, where the corners of the frame are darkened and go from warm to cool. This effect is subtle and is constant on all frames, outside of slight animation of a bit of noise in the gradient. In C4D I add a circular shape in 'screen' space and it layers over whatever tone/color the shader would have.

 

Second, more strongly applied in this project, is 3D gradation of color, darkness and temperature. This is applied as a 3D sphere in 'camera' space so the effect moves with the camera. I have the foreground go darker and cooler, then a fade through warmer colors until the farther objects get a little lightened and with decreased contrast. The eye is drawn from dark to light and from cool to warm. So you can use an area of heightened tone/contrast to define a center of attention, regardless of where in the image frame it is.

 

To control the effects, which appear in almost shaders, I had to make a 'master' shader and reference it into all the others. Yes, that's a lot of trouble to go through. But that's how I'm working right now. You could avoid all the shader work and use a Zbuffer and do the coloration changes in post. I may try that next time I have a free afternoon. Which will be...?

 

I have also added controls to the master shader so I can adjust the distances and amounts of the effects, which means thay can be animated. In the project I'm doing I do that--animate the sense of depth from shot-to-shot since it's a big scene but is sometimes seen from close-up. What works for a wide shot does not when close, and vise-versa.

 

 

These are frames from an animated sequence that illustrate the changes of coloration on objects:

 

note the change in the front trees:

 

 

the gate material is whitish, but as you get close it darkens, cools:

 

[Ernest Burden III]

As you move into the space the foreground trees darken, framing the building while allowing the back trees to be more pale and not smother the building:

 

 

the color of the trees is forced to change rapidly (maybe too strong!) to draw the viewer through a complicated image:

 

 

And who's afraid of a little color to enliven the shadow side of buildings?

 

[Ernest Burden III]

strong foreground leads you to the subject building even when its not close to you:

 

 

the effect is more subtle on wide shots:

 

 

but still there to keep a close object from grabbing your attention and instead simply framing:

 

 

this roof should be a boring, simple shape since it is evenly lit, but can be toned to make the 3D quality stronger:

 

 

 

The assignment was originally to show the massing only, so my client could get hired to do the project. We got that with an earlier version, then he started adding surface detail. The strong colors and loose technique were perfect for the massing study phase, and we held onto them for this design-development stage. Later it will probably be treated more realistically. Or not.

[Crazy Homeless Guy]

In a few of the images I am noticing a line, and I can't determine if it is a cut off point of the shader as the distance from the camera increases, or if it is a shadow. I attached a marked image.

 

At first I shied away form the color of the set of images above, but the more I look at them, the more beautiful the colors become. I think the initial image struck me as being muddy. I might have been starring at the front trees, looking at what you were talking about, instead of letting my eye flow naturally through the space.

 

I am actually guilty of having an under painting file in my library that you created and posted in a thread a handful of year ago. It is actually called, "ernest burden underpainting - set linearburn.tif." I overlay this file on a lot of images, and erase areas of it, until it works with the image. Sometimes it is absolutely horrid, and other times it turns out quite nice. I know I should actually be painting the thing custom, but when I tried, I didn't get results that I considered exceptable. I think the outline in this thread better explains the purpose of the technique, which I think I forgot over the years.

 

Is it possible to share a portion of the animation so we can see the process in action?

Edited May 2, 2009 by Crazy Homeless Guy

[Chad Warner]

That's funny--I have the same file and like to bring it out every once in awhile.

[William Alexander]

Sounds like a good application for expresso in C4D.

 

Are you blocking shots buy similar masking treatments or changing on the fly during the animation? The latter is hellava lot of tweeking and rendering in my mind to get it all 'perceptually' right-timing of focus and emotion type stuff.

 

I'd be interesting, if you ever get an afternoon off, to see how doing the effect/s in post might get more iterations of tweeks and test renders than a linear workflow in the 3D appication. I'd lean toward post compositing in a heart beat, but that certianly does not make it for sure the better option.

 

Cheers

[Ernest Burden III]

That's funny--I have the same file and like to bring it out every once in awhile.

 

Sometimes it is absolutely horrid...

 

Glad to be of service.

 

I'll answer the rest of the points tomorrow, just couldn't resist quoting the 'horrid' thing.

[Ernest Burden III]

Sounds like a good application for expresso in C4D.

 

Are you blocking shots buy similar masking treatments or changing on the fly during the animation? The latter is hellava lot of tweeking and rendering in my mind to get it all 'perceptually' right-timing of focus and emotion type stuff.

 

Expresso--hum. I don't know the scripting at all, so I would have to learn it first. And the rules are visual, hard to put down as numbers. However, if you mean to put in a set of controls based on the position of marker objects (non-rendering) then yes, that is something I've thought of and would like to try. As it is I end up doing a lot test renders to get the distances looking right. I set keyframes at the start and end of shots so the colors and such change over that time (if necessary). If they change too much then it will quickly become a distraction.

 

I'd be interesting, if you ever get an afternoon off, to see how doing the effect/s in post might get more iterations of tweeks and test renders than a linear workflow in the 3D appication.

 

Doing the post route would be like this--render with your regular shaders, no secondary color effects. You would render a Zbuffer depth pass in 16bit or 32 bit (since its grayscale) and use that in the post app. Photoshop CS3 and up can handle animation frames, though I've never done it and don't even know how to yet. A better choice would be After Effects.

 

You would copy the depth channel at least once and use it as an overlay. The first one would be set to multiply or linear burn, probably. You would create a custom gradient with the color fades you want to use and then re-map the grayscale layer with that gradient. In PS the command is 'gradient map'. I would bet there is a version in AE.

 

Now, you could have a second copy of the map which would be set to 'lighten' or 'screen' or similar and use it to kill the darkest tones in the back space, maybe re-map the tones also to warm or cool that region.

 

Now create a selection from the depthmap (best saved in an alpha channel in your rendered output) and adjust the contrast up or down for fore/aft/middle as you like.

 

Since I have a different set of controls for the trees vs. the architecture, I would also need an object buffer pass so I could select the trees separately to apply these effects.

 

Finally, one would need to know how to vary all these effects over time in the post app, just like I have to do in the rendering app now.

 

Post may not be easier. While you maintain control up 'till the end, I'm not sure that's always desirable. It's great to tell a client the project is locked once rendering or else it'll void their warranty.

[Ernest Burden III]

In a few of the images I am noticing a line, and I can't determine if it is a cut off point of the shader as the distance from the camera increases, or if it is a shadow.

 

It's neither. It's a low-rent tree technique I use sometimes, seen too close. Rather than put in massive amounts of tree objects, I will paint a displacement map, or just use a GoogleEarth satellite image (what I did here) to make mounds where trees and groups of trees are. I use that to made a lumpy mesh, then map it with a foliage shader. Up close it can look horrid. From a distance it can be effective in putting in background woods. It renders faster and is much faster to create that individual models, it will shadow, etc. For low-fee projects like this one, it's a good technique. But when you see these tree-like blobs closely, or against a building, they aren't that great.

 

Is it possible to share a portion of the animation so we can see the process in action?

 

I should have thought of that. Yes, I could post a sample sequence. You've seen the site, I just need to avoid anything that would give away it's location. If any of you manage to guess, please keep it to yourself. I don't want to piss off a client.

[Ernest Burden III]

Here are a few shots from my animation. I shrunk it down a bit to 640x360, in H.264:

 

http://www.oreally.com/temp2/RDsample-H264-Q42.mov

 

By the way, I'm not thrilled with how the QT output looks vs. avi or mpeg. It is washed-out and less saturated. I put in a gamma change before doing the QT files, but I shouldn't have to. This reminds me that I've seen posts about how poorly Premiere write QT files, and that I should get QT-Pro to write them. I should try that, though it means doing a full uncompressed dump first (gigs or extra files). However, if the results are better, it may be worth the extra steps.

[Inxa]

Looks good. But I felt there was some issue of loss of quality. But that you have mentioned.

 

Anyways, I felt the stills showed it more better than the animation. I didn't notice it that instantly till I saw it the 3rd time I think. :S

 

Trying to understand the workflow process as well

 

Good to see your posts

[William Alexander]

Expresso - node based gui. Use position of 'an' object or null in relation to camera to drive "levels" or opacity or even the RGB values of the effect layer/s. Cineversity I'm sure has some really good tutorals... Look for the one's by Dr Sassi if he's released any there. Seen some stupid crazy cool stuff driven by expresso. Sure you've made that connection and possibilities in your work flow from your comment.

 

Post work - After Effects - ahhh yup stick with a linear workflow any and every day. Not the best solution for "visual effects" type application in a production environment. Gotta work with what you have-know.

 

I see what your doing and talking about here. Just seeing the ability to render out a good solid scene fast with masks-alphas, z-depths, passes to use in a visual effets approach to do so many things. Being able to vary saturation on the fly of none 'subject' items, vary your overlays' depth and areas by comp'n masks and using levels to adjust the z depths all controlling the effects. Mind numbing complex when playing with shaders inside of shaders inside of shaders for animation. Nodal compositers are awsome for that kind of work, which AE is not, kind of bites.

 

Your advantage-difference in a linear workflow ia knowing exactly (or much more efficiently than many) what you want and 'experience' to "name that tune" in 3 notes. Eliminates the excessive tweaking and test renders to get it right. That is admired

 

Great to see you sharing - being around again

[chiquito]

Hi, very interesting concept. I use 3d max, could you think on any tips or plug in to obtain this effect?

thanks

[Ernest Burden III]

Hi, very interesting concept. I use 3d max, could you think on any tips or plug in to obtain this effect?

 

I don't use Max. Look for shader effects that address screen space, meaning the frame as rendered, for the vignetting I mentioned first.

 

Then look for ways to apply a gradient in 3D space centered on your camera position. That's how I do it in Cinema4D. In C4D there is a list of choices for 3D gradients that includes 'world' and 'camera'

 

Any gradient you apply would be layered over any other color, texture or shader effect, usually in 'multiply' mode. In C4D this is as simple as setting up a 'layer shader' and loading in what you want and setting the mode just like Photoshop.

 

I went further by having my coloration overlays controlled by a central device, in C4D called 'user data'. I'm sure Max has something similar.

 

 

William--you write even more complex sentences than I do. I'm hoping I got everything you were saying.

[chiquito]

Thank you very much in spending time responding, ill try to come up with an alternative and see how can this be applied to max, will share if something comes up

Thanxs

[J.C]

With your npr style I would think this would look better respecting the natural colour perspective (inverted)? I feel constantly trapped in a shadow through that sequence. By the way I'm stealing your idea with the sphere (or maybe a light with coloured falloff?) but with inverted temperature, thanks for the tip.

[Ernest Burden III]

I'm stealing your idea with the sphere (or maybe a light with coloured falloff?) but with inverted temperature, thanks for the tip.

 

It's not stealing if I offered it up, which I did. The dark foreground is a 'classical' technique. If you want to make something new out of it then do! I would love to see what you come up with. Typically we say that the eye of a viewer is drawn towards areas of light within a rendering, and also warmer colors. But that is not a law so other ways could work, too.

 

The more important aspect of the technique I'm using here is that there is change within the picture. The visual space is dynamic. You can re-invent the look a dozen different ways and still have a dynamic image.

[J.C]

Understood! Thanks. As you say, I can see a lot of posibilities with your technique, is there a reason why this wouldn't work with a light with a coloured falloff? Maybe as a separate pass and overlaying or dodging it etc. in post as you play with different hues? Not being lazy I'll test as soon as I have the time just wondering if you tried it.

[Ernest Burden III]

is there a reason why this wouldn't work with a light with a coloured falloff?

 

It does work with a colored falloff light, just not as well. My current method can actually darken a material because it can be applied in multiply mode. A light can add color and lightness, but not darken. Well, in C4D it can with a negative value, but then it doesn't lighten. And vray/C4D does not support colored falloffs, so that stopped me doing that.

[Tommy L]

NICE!

But I think you're crazy doing it in the render engine. I'd easy hook this up in post. More control and quicker results (for testing).

 

As for the QT pro thing you mentioned....QTpro is like $20 (right? or am I thinking of the wrong thing). If it gives you a better output than the QT from Premiere than do it. But QTpro output is still a little washy...I wish we could just use uncompressed avi's for everything and that was that.

 

I render out an uncomressed avi for the client. Then they say "Hey, this thing is huge". And then I do QTpro compression. Then they say "Hey, this thing is washy".

Then I say "Hey, your choice pal, huge or washy".

[Ernest Burden III]

NICE!

But I think you're crazy doing it in the render engine. I'd easy hook this up in post. More control and quicker results (for testing).

 

I might be. However, I am producing an output that looks pretty complete on its own. Testing in the 3D app produces a close-to-final result. I'm not used to that!

 

Also, with C4D noise I can use a subtle animation to the color effects, something Photoshop can't do.

 

As for the QT pro thing you mentioned....QTpro is like $20 (right? or am I thinking of the wrong thing). If it gives you a better output than the QT from Premiere than do it. But QTpro output is still a little washy...

 

Yeah, QT Pro is about US$30. If it produced the same washed-out look, why bother? I can do that right in Premiere. But what I actually do is force a gamma correction from '100' to '77' in Premiere, so the QT output is about the same as avi. I just shouldn't have to. I had heard that QT-Pro produced better results, never tried it though.

[Scott Erstad]

Thanks for sharing your technique.

 

Regarding Premiere output, this is what I suggest. Output an uncompressed AVI from Premiere. Use "Super" to compress your output and save as .mov with appropriate parameters. In QT you can saturate the color, slow or speed up a touch and there you go.

 

http://www.erightsoft.com/SUPER.html

Thanks to Jeff Mottle for turning me on to this nice little free app.

 

Scott