[Steve Bentley]

got it – let me have a look

[Steve Bentley]

Thanks for taking the time you have already. Lots of cinema and mograph sure but my wife calls me the dumbest smart person she knows, so I never assume I know it all, so flap away. I’ve learned the coolest stuff from some of the fresh out school interns we have here. Adam just pointed out the undo view that I never new about – I just think about all those years of importing cameras from older versions of the project (sigh). A wise man once told me to never say “I know”. You never know what gem you were about to be told; and there won’t be a second offering.

I’m better at particles than mograph and I’m able to do most of what I’m trying to do now with that system and since it all seems to be threaded together now in the same engine I thought… And mograph is easier for the client to point to one of the iterations and say ” I don’t like that one”, and you can fix it.

Initially there wasn’t going to be this time offset so the architecture I had set out was sound. But now with this new wrinkle I might be stuck (gulp) hand animating. I thought computers were supposed to eliminate repetitive tasks? I guess whoever postulated that never encountered post production.

Cheers

[Steve Bentley]

Just above the text box is a line of icons. The last one is a disk. Click that and it will walk you through the process. At the end of that process, near the top right will be little box with an http link in it. Just copy that link and paste in the post and it will show up in the forum.

[Steve Bentley]

But are you still editing with Mp4’s?

[Steve Bentley]

No worries on the ?s, thats what the forum is for.
Let me work that backwards.

First a few basics – it doesn’t matter how far the projector is from the screen – ever. If you can get a lens to make the image the size you want from the distance you are at, you could project from the moon and (ignoring air pollution) a 100″ image would be just as bright as from a projector mounted 10ft from the screen. Brightness is only only only ever tied to image size and the punch (lumens) of the projector.

Brightness is also effected by screen material or in your case object material and the angle that material is on relative to the projector. More on this later.

The most important color in a projected image is black (actually this is true of any image) The deepest black you will ever have with a projector is the color of the surface in the lighting conditions you will be projecting in (with the projector off). We always say if we could project on black it would revolutionize the industry. No black means no contrast and no contrast means a wishy washy image. Theaters are dark because without any light even a white screen is technically black. A light colored projection surface in a high ambient light space needs a high bright projector to increase the difference between the brightest white the projector can do and the color of the wall with no pixels lit there (in other words, black). The brain interprets that difference, if the delta is enough, as the lack of light being a sort of black.

You also can’t over come focus on different planes. If you have some columns, depending on your throw distance, whats behind the columns will not be in focus, if what’s on the columns is. Audience distance also affects this- the farther they are from the image the more leeway you have.
And remember that, as in our columns example, there will be a shadow where the column blocks the image from getting to whats behind it. You can over come this with multiple projectors. We’ve done jobs where we have projectors that only fill in the shadow areas and nothing more.

POV. Where the projector is doesn’t matter for the illusion, but you do have to position it to get the best angle on all your surface geometry. And there are some gotchas, but more on that in a sec.
The pov for the audience is key. There is a sweet spot (unfortunately not very big) where all the illusions will work, and as you drift out from that spot those illusions get less and less convincing, and finally they just plain don’t line up. This is one of my pet peeves about the marketing of projection mapping (kind of like the hype of Tupac projected on semi transparent film situated on black and have it be called a hologram – it would have looked the same- probably better- if they had just used a screen) The illusion only works from certain angles. Anyway. That POV spot is where you sit when you do your line up and object trace and where you work to move things around to make the illusion work. And thats the only place it truly works. Now you can give the audience a few of those spots by designating some real world geometry (RWG) for other POV points and doing lines ups for those areas while sitting in those spots. So what the audience sees are some areas that are neat but don’t really work 3D wise and then there is one area for them that is perfect. And each cluster in the audience can have this experience if you do enough POV spots. And contrary to many posts I’ve seen you can do this all with one projector. Its just a 2D image coming from the projector after all. You can also do just simple lighting and geom morphs that only use up and down for example, and keep perspective tricks to a minimum and this increases the POV viewing area.

There’s a famous youtube video, famous not because it’s the biggest gaff I’ve seen in proj mapping but because it heralded the Colts use of the technology (pistons maybe? – not a sports guy) The half time show on the BBall court was marketed as projection mapping like it was the second coming. The POV spot was on the other side of the arena from where the video camera was, so all the shadows are upside down, the perspective is the most wrong it could be and yet it was heralded – argh.
And just to get your juices going have you seen this? One of the most creative uses of projection mapping – sadly (and again annoying for the above reasons) it can only be seen this way by the roving camera – a person in the space would not see the dimensional effects correctly.
https://www.youtube.com/watch?v=lX6JcybgDFo

Here are the gotchas for projection. If you have a RWG plane that is angling away from camera (even a little) the projected image stretches (but you can over come that in your comp with a warp and make it look right), but you can’t over come the fact that there are fewer pixels per sq inch at the back of the plane vs the front of the plane, so resolution falls as does the amount of light – add to that the angle of reflected light and that shape will be darker than the rest of your “canvas” and then darker still as it goes back. So you may have to compensate for that with feathered adjustment layers in your comp. (usually you have darken everything else because the projector can’t do more bright than bright.)
There are going to be spots that are so extreme angle-wise that the image is smeared beyond flexibility. There is no reason areas of your composition can’t be black – make these areas a bold design choice (ya that’s it) and make them black.
Different surfaces will reflect light back in different amounts. Again you can compensate to a degree. Same goes for colored RWG but often your hands will be tied. You can project a green Tron graphic on a red and black checked lumberjack shirt till you are Tron-blue in the face and you really won’t see much.

As for method. It really depends on your RWG. But lets say you have a classic building facade with doric columns etc surrounding a portico. The natural inclination is that you would get into cinema and build columns and then project you imagery on those in C4D using cinema’s projection mapping and then output that for you projectors, but now you are warping your imagery around cylinder shapes twice, once distorted by cinema and once by the RWG, doubling the amount of distortion. Most of what we find we are doing is undoing the distortion that the RWG lends to the projected image so that we can impart our own distortions and depth. And the coolest things in 3D mapping are showing the audience what they think is a flat surface and then letting the dimension leak in.

Usually the process comes down to a flat mesh – with a before and after positioning of the vertexes. One for the planes that you can make content for and the “after” mesh brought into line with the outlines of your RWG. Then exporting a UVW map of that warped mesh to AE, and with ReMap, with the UVW applied, it will warp your content to fit the warped mesh without ever going back to cinema. (BTW you can’t do this in AE direct even though you can make a mesh with a number of tools inside AE – the math AE uses to match moved vertices is not quad based so you get curves inside the polys imparted to the footage)
That mesh warping vertex by vertex is as painful as it sounds so I built an xpresso rig that maps the 3D geometry we build, to match the RWG, onto a flat mesh of vertexes, but there is no reason you couldn’t do that by hand just moving vertexes on a flat mesh to match a 2D view of your RWG. Remember, from the projectors pov there is no 3D. It’s assuming a 2D plane to project onto and something that recedes into the distance the projector sees as it’s edges just getting closer together. Hence the mesh with vertexes just getting closer together to squeeze the image together to fit that receding geometry.

Buuuuuut. If you have mad mapper, let me save you hours of grief. Use Mad Mapper. It can match RWG in a heartbeat and do your warp correction for you. You can send any2D content you want into it and that package will look after what we were talking about cinema doing. Then if you want to you can use cinema to make 3D objects that can give the RWG depth that it didn’t have to begin with. You can approach it (in cinema) as though it is a flat canvas – so you can just film a flat “screen” with objects extending outward from it and MM will warp that for you.

As for outputting AE – you can designate one of your vid card outputs as the presentation screen in the prefs and that can go to your projector – its a live borderless view of the comp window..
What i don’t know is if you can do the same but feed this output to mad mapper because mad mapper will feed the projectors live.

There are lots of other tools you could use for this was well. Cool Lux, Green Hippo, Dataton, but non can touch Mad Mappers price and the former are trying to be all things to all people where as Mad Mapper is tailor made for this sort of thing.

I wish I had map mapper when I was growing up.

Some contents or functionalities here are not available due to your cookie preferences!

This happens because the functionality/content marked as “Google Youtube” uses cookies that you choosed to keep disabled. In order to view this content or use this functionality, please enable cookies: click here to open your cookie preferences.

[Steve Bentley]

Hi Ben, Can you post the project? It could be so many things.
The escaping particles sounds like they are sneaking out “between frames”. When the particle arrives at the boundary it’s in a discreet spot but its position has to be calculated before the boundary check is calculated. If its traveling at a set speed and the next frame puts it just outside the edge detect distance (but technically still inside the volume), or the mesh of the boundary (being lots of flat planes) isn’t fine enough to get good math resolution, or you’ve chosen a simpler sampling method to get either the particle boundary or the volume boundary, then that particle can appear to cinema to be just outside (like 0.00001 outside) and cinema can only shrug its shoulders and say “oh well one got away”.
So depending on your emitter type and particle system you can do a number of things: sub frame sampling, turn up the boundary detection distance, pick a different boundary detect type, give the visible particle an invisible parent that has simpler geom and it’s what bumps into other edge first, make an invisible false boundary object that actually is used in the detect and it’s of a simpler shape, or render at twice the frame rate (in essence your are sub frame sampling here) and then just time warp your final in your comp software.

Also check that all the normals of your boundary volume object are pointing the right way – inside and out- one reversed normal is the same as a pinhole in a bucket (depending on detection routine chosen) .

[Steve Bentley]

What codec are you rendering to?

[Steve Bentley]

Do you have any footage in the comp like mp4’s or H264 or anything that is compressed? These file types have what are called “b” frames (and other things) that aren’t discreet frames and AE only works with discreet frames so AE can hang as it tries to process a virtual frame.

[Steve Bentley]

And if the expressions become a multitude, it can take forever to render. AE is horrible at stepping through layers looking at each attribute. I don’t know what it is about adobe but simple things like scrolling down a list of layers (Pshop too) is so slow I think I’m on my 80 mhz Mac Quadra again.
There’s a great script that helps with grouping layers and setting selection sets, but when you trigger it you can watch in real time as it steps through and turns on layers, and it aint quick – its running code for goodness sake! Whats taking so long? Even running your mouse down through the solo buttons, you can beat AE to the bottom of a 50 layer comp and still have time to take a sip of coffee before it’s caught up.

[Steve Bentley]

Sorry to here that. Is it perhaps a layered file and premiere is only grabbing the base layer and that is either black or has nothing in it?
Is it in a color space that premiere can’t do or isn’t set up for (16 or 32 bit)?
Is it a tiff perhaps that has a mac encoding or vs a PC (same goes for Targas).
Does it have a compression on it that premeire or your computer cant’t do?