[Steve Bentley]

Thanks Dan,
That did the trick. I guess I was overthinking it – assuming I would have to reintroduce the inner object to the new “world” of the outer differently sized comp, or, compToSurface the point on the layer that was in 3d and scaling and moving in depth. I’ll post the fix so others can benefit. I actually thought I had tried this first but my layer getting the position info may have been parented or not in 3D at the time I tried it.

As always, you’ve made my day brighter.
Cheers

[Steve Bentley]

You could also make your mask in Illustrator – it is the king of vectors. Bring your image in and trace it with the pen tool. Then you can either save just the path as an AI file and import that into AE, or you can copy the path and past it into a mask that already has one keyframe.
If you autotrace you may be getting more vertexes than you can see. (especially in the fjords of Norway!)

It’s been while since i’ve been in CS5 but I’m not sure how well that version AE handled Pshop’s smart objects. You may not be getting the vector you think you are from the paths you export from Pshop. But it handled Illustrator just fine.

[Steve Bentley]

I take it you don’t want to parent the shapes to the camera (or the null) – this would keep them the same size as you pull back but they would also travel with the camera so they wouldn’t move relative to the camera. (although you could animate them)

In the expressions the only difference should be that a null will have an anchor point and the camera wouldn’t.
You didn’t include any expressions but in things like a=nullLayer.toWorld(nullLayer.anchorPoint); you have to reference the anchor point of the null. But in your original camera code you probably had a=camera.toWorld([0,0,0]); Since cameras (and lights) don’t have anchorpoints you just use their origins.

I imagine using vectors would be the easiest. You could normalize the distance the camera is going to travel (or the distance to the shapes) and then you can use that as a scalar function on the shapes (or at least a multiplier) because that vector would go from zero to one. That way if all the shapes are at different scales or sizes they would all scale the same amount to keep up with the camera’s distance.

I don’t think there is a “width” transform where you could tell the shapes to stay at a constant width relative to the 2D viewport.

[Steve Bentley]

I went searching for camera footage in H264 and couldn’t find any that were compressed as low as what you uploaded (even my gopro has a higher bit rate that this) Are you sure you have camera footage and not someone’s recompress?
Take a look at the footage in the link below – its H264 (albeit at 10 times the bit rate of yours) and it keys with one click. Its clean enough for a good key, keeping the whips of hair at the edges.
While you do have issues with wrinkles in the screen there has to be a cleaner record of this footage somewhere. And that would help you immensely.
12287_171003d0172k.mp4.zip

[Steve Bentley]

Oh I get you. You have to put a marker on the outer layer and then “update” and all the markers update onto that layer from the inner layer.
Without that initial outer marker that update option does not show up in the right click menu. I was right clicking on the markers on the inner layer.
Brilliant. Old dog meets new trick.
Thanks C.

[Steve Bentley]

Here’s a lengthy response to your question where you quoted my answers:

I think all the issues are due to the codec and the camera.

Some cameras just don’t produce a clean image. The camera makers can hide a number of sins, including a blocky image, by adding noise and this may have been done on that camera. There’s a reason a RED camera is $40k and Arri’s go for more than $100k (and that’s before lenses and doodads!) and prosumer cameras are only in the $1500 t0 $5k range.
Shoot with the best camera you can (we’ll even rent a special camera just on the green screen days while the rest of the production is shot on something else less costly)

When we do green screen we are so picky that we ask for a digital feed out of the camera, live on set- that feed is usually uncompressed, but even if it isn’t, its always miles better than what goes to the tape. Very few cameras record uncompressed to tape. 4K files are so massive that you wouldn’t get too much footage on a single tape – you would be pining for the good old says when film reels were only ten or twenty minutes long.
That feed goes right to our capture card which is capturing in an “uncompressed” lossless codec in 10 bits. And that goes right to our hard drive array. This way we can also do a test comp on the set to make sure the key is pullable. There is just no comparison between any version of H264 and this kind of file. H264 was made as a final delivery codec and not one for doing effects with. So much of the information you need for effects is long gone on a H264 compress even when its done way up at 40mb/s (twice blue ray). Even Apple’s ProRes is way better than H264 for effects work. Do a a little research on what 4:2:2 or 4:4:4 means (and all the other combinations) with regard to color space in codecs and cameras. Its shocking how little color information is needed to convince the human eye they are seeing reality. Keying software is not that easily fooled.

But you can’t always get to the set or they don’t want to “waste the time” (more on this later). And you don’t always want to rent a deck to capture the footage with. So if you can have someone digitize the footage you should ask for a high color space format. Like EXR or ACES or RAW (if you can handle this one) or Cineon. And you will get stills, not movies. Sometimes you can get the camera’s proprietary format, like on a RED. What ever it is, get as close to uncompressed as you can with the largest color space (8 bit is just so last year!). But make sure you have have a codec that can read that format. (in this instance codec doesn’t mean compression; more so its compatible and readable by AE). We like EXR – AE, Photoshop and most 3D software reads it out of the box, its 16 bit or 32 bit, can hold an alpha and multipass rendering.

The DPX format (and I’ll be branded a heretic here by some for saying this) seems to have a different flavor depending on the camera. The big cams that use it (like Arri) use it well but some of the prosumer cams use it too and have mangled it. It can be excellent but it can also be terrible. Ask for a set of test frames. Watch what happens when you play it – view zoomed in at 800%.

As for the ideal color for the screen it really depends on the camera, subject matter and what screen material you can get. If you get a pipe frame for your screen, you can stretch it across the frame to get most of the wrinkles out. There’s no point ironing the whole thing because only parts of it are in the shot or its so far back its out of focus. We take a steamer to set for pesky wrinkles that are showing. Under the strain of the frame those wrinkles take one look at the steamer and straighten up right quick. The frame also means you don’t have to tack the screen up – just attach the frame to two C stands.

If you light the screen with green gelled lights (barn door those lights so none of that falls on the talent) you can pick your gels to tweak the color of the screen. You have to balance these lights with the ones that light the talent which also fall on the screen. We light the screen first so we can start with that evenly lit. The minimum we shoot for is 20 feet between screen and talent and another 20 feet to the camera. A short lens used in green screen screams “Effects shot!” and never looks real. A good distance from your screen also means the bounced spill light from it will be minimized on your talent’s edges – which is the all important keying area.

We get the camera info before hand and then look at the color sensitivity of the CCD (or film stock if you are going that way). Look at where the green color arc peaks or where the green arc is the farthest from the red arc and choose a color for the screen or the gels that gets you closest to that frequency of light. There are color meters out there that can measure this when you look at the screen through them.

But all of that is probably moot on a shoestring budget. That image I sent you is a very nice color of green as a rule of thumb – kind of dayglow lime green. The one you used, while functional, is a little too forest green or emerald green (its got too much blue in it). The idea here is to pick a green that is not “contaminated” by either of the other two primary colors – add some red and the screen is more yellow green, add too much blue and you get your color. You want that screen only showing up in the green record of the file. You can still key with your color, or any color for that matter (ironically the best screen color is red, just not for people) but the closer you can get to the magic color the easier the trickier stuff is – fine hair, veils, glass, water, reflective stuff.

A 20 foot by 20 foot screen (called a “20by” written “20x” in the industry if you are renting) is pretty much the smallest you want. By the time you get the screen back twenty feet from the talent (a minimum you should shoot for) a 20by just barely covers a person head to toe. Renters also have standard frames that fit this size. From there they go 20’x40′ and 40×40. And they do go bigger for outdoor landscape masking screens. You can also paint a cyc in a studio. Green screen paint is crazy expensive (because its hitting the magic color) but you can get Home Depot to match the color you want – matte always. Take in a bit of green screen gaffer tape for them to match.

Lighting is important – try and get a smooth and even a lighting pattern across the screen – use a spot meter to measure side to side and up and down and in the center. Again you can key without this step but often you have to pull multiple keys for different sections of the screen.

Codecs – H264 is a miracle and has single handedly made the internet a viable delivery mechanism for movies (in 1991 I said this was something that could never happen because of film resolution and internet bandwidth – but then I’m in good company: Richard Edlund said the same about digitizing film into a computer the year before, and Bill Gates said 640kb was all you would ever need for RAM – the moral? Never say “never”, or try your hand at futurecasting). But H264 is not a good effects format for your fx source. Sure its lightweight to send around over the internet but you want every drop of data you can get from the shoot and you don’t want the codec fudging the truth just to get the file size down.
Blackmagic Design has an excellent codec that bridges the gap between a lightweight compressed (but noisy) codec like H254 and the near perfect but massive records of a RAW or EXR. And as luck would have it they also make capture cards. You don’t need the capture card to work with BMD’s format but you do need one to play the format real-time. They also make an excellent and reasonably priced cinema camera (whether they have any in stock is quite another matter)

I know this all sounds like a lot to hoist onto a director because the effects guys are already the thorn in his/her side slowing down principle photography. But getting the details right will make your composites go so much smoother and you will have more latitude when something was done incorrectly, or there is too much motion blur in the footage to pull a key, or the tracking markers aren’t showing, or or, or….
Keep in mind too directors are terrified of the effects side of things. So while you are a time delay distraction with all your requirements, they know that it’s in their best interest to give you what you need to work with or their post budget will soar.

[Steve Bentley]

in AE you might be able to track a few pegs to get a rotation speed out of it. I say might because the center of the circle (the rotation point) is outside frame all the time, so not sure if that would work.

[Steve Bentley]

Hey Matt,
It looks like the forum didn’t update properly so I missed your long post with my comments in quotes – let me answer those shortly.

[Steve Bentley]

use option (mac) or alt (PC) in combo with the page up or page down button (think: “down” the time line to figure out which one)

[Steve Bentley]

Thanks Cassius,
But I don’t have that option in 2018 in the right click menu, neither is there anything of that sort in the settings for the marker, or in the time line between markers I tried right clicking on.

I just tried making a newComp and setting markers and then bringing newComp into the outerComp in question – the nested markers show! (as they should) but the original precomped layer with nested markers is still not updating.

Taking that one step further, if on the newComp layer where the nested markers are showing, I dive down and place a new marker inside it, that new marker does not show in the outerComp but the original markers do (made before placing newComp in the outerComp)

So just for fun I went back to the original precomp layer I made the markers in (lets call that markerComp), I deleted markerComp from my main comp (comp1), and then dragged markerComp back into comp1 from the footage list – Tada! markers on the markerComp layer! So there is a way around it, but if you have already animated things on the markerComp when you think, “you know what? markers would be helpful inside this one”, its a pain to transfer any animation you have already done on to a new copy of that markerComp you drag in.

But it gets worse – if you change the position of the marker that do show – the outer reference to that marker does not shift in time to match its inner instance. Even on the layers that you brought in after the markers had been placed.

I don’t know for certain but AE may have always worked like this – but I think I would have noticed by now (I’ve been using AE since before it was AE – remember COSA?)

I just can’t believe there is no way to update the marker display in outer comps. I’ll dig out an older version and give it a whirl.
Ah, to be part of the Unpaid Beta program! Livin’ the dream!
I got better treatment when I was part of the Beta program!