I created a Video from a room with my camera. In this the is a cabinet. I want to animate a small helicopter flying from behind the cabinet through the room. After tracking the video with my tracking software (http://www.digilab.uni-hannover.de/docs/manual.html) I built a dummycube and brought it in place so it hides the helicopter. I found out that I can use the “Env” button for a material to make it render somewhat transparent so I can see the background. My Problem is as follows:
The cabinet has some glass elements I want the helicopter to be reflected in. When I use the “Env” button, the object hides the helicopter correctly but doesn’t reflect it.
I used Autodesk Maya before and there is a Material called ‘use Background’, allowing me to do this. Any Object with this Material is rendered fully transparent but one can setup specularity and reflections.
Is there something similar in Blender or is the way to achieve this a complete different?
I hope you don’t take this wrong but it seems to me that there are a lot of Blender basics that you need to cover before undertaking a project like this. To answer your question about the material teh answer is no. With your current knowledge of Blender you’re probably going to need someone to lead you by the hand all the way through a project like this. Compositing is all about masking, combining, the artist, and the tools available. Blender has it’s share of tools that make compositing easier than many other apps but it has many more that make it much more difficult to use than other apps also. One of the good things about this is that it forces you to become a better compositor because you have to understand how many of the tools work in order to put them to use. I don’t see this lasting too much longer though because many of the developers are tired of taking the long way around and are building new nodes that DON’T require such knowledge to make things easier on all of us. The bad thing about this is that the folks who start jumping in on the new tools are going to struggle even harder when there isn’t one click node to help them out of a jamb because they won’t understand the true nature of nodes compositing.
As far as the cabinet reflections go you’re in for a very difficult ride what with a bunch of match moving by hand, unless you own match moving software such as RealViz ar compositing software with motion tracking and corner pinning such as Adobe After Effects. I can highly recomend AE for the job because it makes light work of a task like this.
One thing I can promise: you will have a very difficult time pulling off a composite like this with Blender alone (if you have a camera move when the reflections are needed) because no one has really developed functionality of this type yet.
I know that this is a bit difficult. And you’re right: I’m quite new to Blender. This is not a big project, it is a thing I started just for learning.
Anyway. I don’t took it wrong.
As my english ist not the best, I’ll ty to clarify the situation. I already tracked the video with the software I mentioned above (follow the link and you’ll see, its a software like icarus or realviz but its free). So I don’t have to care about matchmoving. I already built my helicopter and animated it with keyframe animation. After testrendering the animation it looks quite ok, the only thing missing is the reflection on the cabinet. So maybe you (or someone else) may point me into the right direction. As I don’t own AfterEffects this is not an option but there may be another way to finnish this.
Posting an image would help. The problem will come in if you have the camera moving but need to pin the reflections to the windows at the same time. You wouldn’t want the reflections wondering over parts of the cabinet that are made of wood. They also need to maintain proper perspective which is where corner pinning comes in. Could you possibly upload a video of what you’ve got so far? That would be more helpful than anything else.
I knew what the link was: Voodoo…lots of us use it but that’s camera tracking software, not motion tracking software. Camera tracking is basically the inverse of motion tracking.
Yes, for sure.
I’m just rendering a new sequence, as soon as it is finnished, I’ll upload it to my webspace. May take 30 Minutes or so until it is complete. Sorry for this delay, I deleted the old sequence accidently.
Without seeing the details of your scene setup (relationship of helicopter to cabinet, nature of reflecting surfaces) this is just an off-the-cuff suggestion – would it be possible to create in Blender surfaces that match those that should reflect in the final scene, give them reflective properties, render some reflection passes, matte (mask) them as necessary, then composite them over the video footage? I’ve done something similar with raytraced shadows that could not be rendered directly with the rest of the scene and it worked very well.
It may even be possible to use either a separate Scene in the same .blend as the helicopter, or perhaps Render Layers, to set up the reflections pass, which could then be mixed in the Compositor without the need for rendering out separate passes.
Again, a lot depends on the nature of your reflective surfaces. If they are something highly specular and mirror-like, this would be more difficult because you’d have to match the material look of the video footage very closely or the “special effect” would be very visible. But for reflections in glass or other transparent/translucent materials, a fairly simple Mix or subtle Add might work very well in combination with a matte (perhaps using a Luminance Key node – type of matte may also be dependent on the details of your scene).
That why I wanted to see video. It may be very simple to pull off, then again it may not. I really don’t think it will be as easy as just slapping some reflective planes into a camera tracked scene though. There are just too many variables that can blow it to pieces.
yeah, but KISS is still a very viable philosophy ;). I always look for spit and bailing-wire solutions before I start getting more elaborate, and it often pays off. It’s all smoke and mirrors anyway, and sometimes you can even skip the smoke.
True. From his description I think it’ll come down to perspective corner pinning though. That’s probably going to be the simplest way out if the shot is anything like I imagine. If that’s the case I can probably do it in AE and send a mask back to multiply against the reflections.
If there are enough perspective and depth cues in the video, it may be that constructing the reflection planes to match would solve the “corner pinning” by having the perspective built into the reflection pass. Essentially he could reconstruct just those surfaces in a 3D match move. All just spec 'til we see some example images, though.
Please don’t chop me about low quality or bad modelling and animation: As mentioned above this is a small project I started to learn. I never intended to publish this.
The original video was created with a digital camera wich is also able to record videos (mpg1, poor quality at 320x240).
I’d go for a match of the reflecting planes, there are lots of cues to build a 3D match. And there’s so much blown-out highlight on the glass that you’d only see a few frames in the smaller dark areas, and they’d be real subtle.
One note: the geometry of the scene tends to make me think you wouldn’t see reflections in the glass unless the 'copter was at the extreme left of the frame or very close to the cabinet, so it may be a moot point, anyway. That’s another reason to put a model of the cabinet in with the 'copter (it need not be a complex model), to establish the proper incidence/reflection relationships to keep the reality as tight as possible.
Actually I think you could do both with a very simple model, just a cubic form that matches the basic proportions of the cabinet. Call this the “stand-in.” It could have the entire front face made reflective, then you could use texture masking to “model” the separate panes, limiting the reflectivity to only the areas matching the glass surfaces. By placing the texture mask on the stand-in as a visible texture at first, you could check the accuracy of the “fit” between it and the cabinet on video.
By using a stand-in that reflects your 'copter model as the cabinet would, you’ll establish exactly how the reflections should appear in the shot, based on the copter’s position relative to the cabinet. It may be that only a few frames will actually show reflections, in which case, you may be able to do a frame-by-frame composite in a 2D-app like Photoshop. This would give you greatest control over the final look. But if there are many frames, then you’ll have the elements to generate the reflections and any mattes you may need to mask and composite it in Blender.
keep the reflections from the video – you just want to superimpose/add/mix your fake reflections over what’s already on the vid.
Thank you for this accurate explanation
Ill do it as you suggested, but this may need some time, so please don’t be angry when I don’t answer immediatly with the results.
keep the reflections from the video – you just want to superimpose/add/mix your fake reflections over what’s already on the vid.
That’s what I meant. I expressed this wrong, sorry for this.
As long as you won’t be angry if the idea turns out to be worthless Or too much trouble.
In any case, that’s how I’d first approach the problem. I’m sure some After Effects-type solutions are just as valid, but this at least keeps it within Blender.
Or too much trouble.