Hi Vern,
NB: My mistake on Depth of Field - you wouldn't be doing DoF mods with a matched background, though you might have other effects to worry about.
Okay, this answer is going to get long - you might want to make some coffee first.
It would be nice if there were an animated example of this situation.
You do, you have the bouncing ball scene posted above.
It comes from an actual working example I used to use for training compositors with Animto 1.8 and 3.x. The exercise tests both the compositor and software, and introduces newbies to one critical element - the director!
Directors are one rogue part of the animation process, those un-namable people who delight in messing everything up because they change thier minds, frequently and repeatedly.
As a result, in the real commercial world, you have to make sure everthing is kept together, KISS. Character A in one folder/layer, Character B in another. That way, you can reposition, resize, change opacities or whatever the little whim these fickle monsters have called for. Backgrounds and anything attached to it are in a single folder. Overlays are treated in the same way - you can be sure the director will request a change to the out of focus settings (and you will have to change it back again when he sees it on screen).
In the bouncing ball test, the first stage is to make the ball bounce on the spot, squashing on impact, and the shadow looking realisic (its size and opacity have to be changed too) and in sync with the ball. Then you tell them the director has changed his mind and wants the ball to bounce through shot. The compositors have to change the east-west positions and timings to make it work as per new storyboard. Then you tell them the director from hell wants the ball to bounce front-left to rear-right. That means they have to get into scaling (the secret is to make the scale centre at the point of the shadow).
But he's not satisfied: Our bloody-minded leader (who couldn't direct traffic, let alone films), thinks its going to be great and funny if the ball falls off the table top and plummets to digital oblivion: This is the stage the bounclinng ball scene above is now at. It means matting the shadow without affecting the ball. In Animo, matting the shadow is easy - the ball is more tricky - it has to have its matte reversed/turned on just before it diappears behind the table surface.
Unfortunately, Moho failed the test when I tried matting it ... you cannot use the same matte on two objects. Neither can you use the background layer as matte on the shadow because its outside the layer: Put the matte inside the layer and it wont work because the layer is moving. You could match-move it if it was a linear move but put the ball onto a curved track and you will never match-move the matte with the background.
Okay, so you could dupe the background and import it over the top. As the director of this test, I'll mess you up - I want the angle of the background to be changed. Now you have to remake/change two layers. If the matte was external to the layer, you could make the change safe in the knowledge it will self-matte and no other changes are needed.
Okay - lets look at the final level of this animation assignment. The director wants the ball to bounce off a new vertical wall, then drop off the table. As Moho is now, even using a cloned layer to do this is very problematic ... I have never tried it and to be honest, I can't see any point in wasting time "fudging it". It requires a different approach.
Being fair, this is a very carefully designed assignment whose aim is to demonstrate all the principal composite tools. When you first start, it looks childishly simple. Then the tutor then introduces the director changes and the smiles are wiped off thier faces: Unless trained, people normally build composites without anticipating modifying it later. If you set every scene up so it can be changed later, your collegues will appreciate it. Usually, by the end of a trianing session, the students fully understand why you need to keep everything organised, elements kept together when ever possible, and always expecting the fickle-minded director to request changes. It's just good working practice.
So what about mattes in practice, for real in the field ... what is the director going to throw at you, and always at the last minute?.
For one, he will tell you its digital and therefore its easy to change. So change it he will. Colours, effects, camera moves, camera shakes get put in, then taken out again, etc. You do not want to rebuild something every time. So it can be better to add an empty layer with camerashake or modified move so you can revert to the old move if he changes his mind and reverts back - believe me they do and have, and want that exact move he liked before to changed it.
In a 10 minute short called
Drums of Noto Hanto (for S4C, UK), the Animo team had to complex-matte every scene - every fill and line on every background and character was a matte replacement. Here is an example:
Everything you see in this image is texture matted in, everything - every fill, colour and line: Not one single line of the original animators drawing remains. If a scene had just one background and one character, the matte count would be around 60 mattes. On the most complicated scene (Sc 12 if anyone is interested), I put in over 300 mattes, in a scene with three separate levels (levels not layers - BG, midde ground and overlay groups). The scene is a vary speed pan tracking a character from A to C pegs (16 inches). When you get to this level of complexity, you quickly understand why everything HAS to be grouped in one folder/layer, with just one control point/timeline. In addition to the matting, most exterieor scenes had 10+ multi-plained airbrushed "mist" layers, with the characters standing in the middle of the ground-hugging mist (aka music video dry ice fog).
Not all films are that complex.
But in the commercial world, the art director will ask for something which looks different. It means the composite has to be KISS. Self-matting layers is just part of that. Cloning isn't. Photoshop-ing a background is practical but matteing the object is better - its faster and the matte can be applied in multiple ways, ways in which a alpha channel cut-out can't. Softedge mattes is one (flame glows intrude on an object in front - a cutout will give a hard edge).
Moho has some great tools in it and I love the program. It can give you a variety of differnt looks and feel to a project: The 3D aspects and pan-arounds are really good - it made my producer sit up and slap people on the back, "a job well done".
But there are elements of Moho which, if being honest, are clunky. This test demonstrated three. The ease-in/ease-out maths didn't work correctly. The work-area dispaly failed to provide WYSIWYG. And the test exceeded the masking capabilities.
I hope people will not see this observation as being overly critical. It's in fact about taking a good product and making it better, one which the commercial studios will leap at in the same way they did flash - and the studios do dislike flash becuase its not intended to be a TV animation tool.
My point is making the matting external so any layer can self-matte another layer, will make the product very much stronger. It will allow aniamtors and technicians to really push the envelope.
And that will be good for sales.
Rhoel.
Please download the scene amd play with the bouncing ball scene - it's both fun and demonstrates just how flexable an all-in-one layer system works. The x-y axis has just two keys .. the scale axis has just two keys. With those four keyframes, you can put the ball anywhere, fast.
As a test, change the ball positions from rear-left to front-right ... even without understanding how the scene is built, it should take you less than a minute to get it right. In a multi-tasking studio, any compositor should be able to modify a scene without hours of figuring out how something is built: Time is money, and time should be on the creative bit, not swearing at some deeply hidden modification. Having to modify multiple clones of match-moving layers really makes me swear as its un-neccessary, and clunky.
Modifing this scene ins't.
