IPO's and Composite Nodes progress???

Hi all
Firstly, forgive me if this has been answered recently, I did a search but found nothing useful.

Does anyone know what the progress is, if any, on the ipo curves for composite (and possibly material) nodes?

There was alot of talk about this being in the release after 2.42, but since that release I haven’t seen much mention of it, and not found it in any builds I’ve played with.

Has the idea been shelved for the time being?


I too saw promise of a DOF node, but did not find any, so wrote that section of the manual. Someone has updated the Composite node list, but there are matte nodes and some others; none that were added had anything to do with IPOs. I think the time node is the start.

I was wondering about the time node in materials to achieve animated materials. I guess the release logs is the place to start looking; I understand that there is a freeze on features for the Christmas release, so if it isnt’ in there, it has been shelved.

2.43 is due out before Christmas?
Man these people work fast!

Shame about no IPO’s. The time node is not anything like what the IPO’s would offer. Node IPO’s would make blender a truly great compositing app, something the Open Source community desperately needs…

BTW, DOF can be easily acheived using a depth mask and a blur or defocus node. Defocus is ideal, but more complex than a simple blur, and can be constructed with the available nodes.


fast and free.

rewrote DOF to use depth and blur, just as you said. wish I had listened to you before writing it the first time (wrong). Well, that’s too harsh; it wasnt wrong as it was the best the forum could come up with. Better now.

ah what did you rewrite RogerWickes, a new node???

those tricks are cool when the blurred background is never next to a an object which is sharp in focus. otherwise the blur of the background will mix together with the edges of the objects in focus and thus produce those typical and bad z-depth based fake dof blur effects.

or do you have a a trick to fix that? in my experience nothing goes over real dof and also blurred dof which is faster of course is also the least realistic because everything out of focus is not perfect blurred.

take a finger and un-focus - you will see first a grain pattern in the blurred areas plus you will see that you see something similar to banding. still trying to figure out how to make that in blender :wink:


ah what did you rewrite RogerWickes, a new node???
No, i could never actually write a node code! Well, I could try, but it would just horribly corrupt your computer and format your hard drive. I just write about the great code these wizards spit out, seemingly effortlessly. I wrote/rewrote the DOF section of the wiki and surrounding sections.

actually, i do. See the section I just posted this a.m., because I was trying to apply the concept of blur background to flat photographs, since part of our audience is 2D compositors. See http://mediawiki.blender.org/index.php/Manual/Compositing_Blur_the_Background#Blurring_a_Photograph.27s_Background
I had to work thru a dozen rigs and techniques before posting this one that solves all the problems, esp. as the edges as you suggested. Altho i think Cubic works better than flat.

Now that’s the easy stuff. What’s REALLY cool, if you saw the movie Pi or even some others, is to make a clip where you increase the background blur over time AND scale up the mask and photo at the same time. Really cool as she (my daughter in this case) comes right out at you. No need to decrease the offset that way. The effect is almost frightening, with no artifacts. very nice. I love this stuff.

I agree. I’m not a specular physicist, but I actually don’t think blur is what you really see, at least in the two dimensional sense, when you look at things out of focus. Maybe in three dimensions, with the net effect projected onto your eyeball is a depth of field thing. Maybe that’s why I like cubic better, because it’s ‘denser’ where the change is greater, to mimic the out-of-focus photons adding with the foreground. If that is the true effect, then it would make sense that there would be some interference patterns that emerge, which you percieve as banding. And then of course you have the haze / halo issue…it’s really pretty amazing, you know?

Hi all,
Ok, my solution mentioned earlier was a VERY quick fix, but the results in your wiki page RogerWikes are pretty cool.

cekuhnen, focal blur is a very complex thing, and you are right when you say that a simple depth mapped blur doesn’t do it justice. One of the most recognisable attributes of real DOF is that bright objects are more blurred than dim objects. ie: the sources of light actually get bigger and/or brighter when you de-focus. Maybe this could be simulated by placing a limuninace channel of the image inline and using it to scale the blur effect.

As for the edges, maybe if you used a curves node to isolate the depth of the focal point and slightly scaled back the rest of the z-buffer that would help.

I’m on your side though, real DOF rocks by far. Why can’t it be integrated into the renderer? That would be the best solution… (if it’s already there, apologies to the dev team!). Anyhow, a simple de-focus node would solve all the problems in this respect.

And I still want some Node IPO’s at some point! :slight_smile:

i saw some cool work using a blur as a background to the original alphaovered itself, creating a glow. In writing the tut on image processing, I saw a combo of a filter setting and a mix that gave the highlight glow of which you speak. I think someone could take the time and use like a filter node and a blur node with some mist/smoke/cloud texture and a bunch of mixing to get a verrrrrry close to real thing going.