[Gareth Randall]

There could be any number of ways, but it would really help to see what you’re describing so that we can make better suggestions. Can you post a screen grab?

[Gareth Randall]

How long is a piece of string?

Seriously, there’s no one answer to that. It all depends on what you want it to look like. But you might want to start here.

[Gareth Randall]

It was the Ripple Training “Motion Fast Forward” video tutorials that got me up and running quickly.

I’ve also got Mark Spencer’s book, which is an excellent supplement, but I think there’s nothing to beat a video tutorial to get the concepts over quickly. A written description can be misinterpreted or misunderstood, while a video shows you exactly what you need to do with no ambiguity.

Like the man said – a (moving) picture (with spoken commentary) is worth a thousand (printed) words 😉

[Gareth Randall]

Hi Jeff,

It’s a two-stage process, and there are two possible variations of the second stage.

First step is to motion-track your insert video on top of the background, so that it matches the movement of the TV screen. Then you *either* use animated masks to hide the parts of the video that need to be covered up by the shoulder and hand, *or* you duplicate your background, put it on top of your motion-tracked insert video, and use animated masks to comp just the hand and shoulder back over the top of the insert video.

The second variation of stage 2 is probably the simplest, because it’s easier to see where you need to draw your masks. As far as animating the masks goes, you just draw the first one, turn on keyframe recording, step forward a frame, reposition the mask control points as necessary, step forward again, and repeat until bored 😉

[Gareth Randall]

Hi Nima,

[Nima Masood] “- At 1:00 the colored squares going through the screen(it’s only 1 second). THat looks really cool!”

It’s also really easy 🙂 All they are doing is using a shape (sort of an arrowhead/triangle/diamond kind of thing) as a mask to reveal a copy of the clip that has a different colour grade applied. Just after the shape appears in the bottom left corner, the main video cuts to B&W.

BTW, regarding the effect at 0:56, I don’t think any tracking was involved there. It’s such a short effect and the whole feel is so grungy and lo-fi that I imagine they just did a rough mask around Tupac to composite a “clean” copy of him over the background.

[Gareth Randall]

Hi Andrew,

If you check out this post you’ll see some stuff that should help you with what you want.

[Gareth Randall]

[James Frazier] “Will this method keep the shape/clip from following the camera as I move to another screen?”

Yes, it will. The key is to ensure that the clip and shape are in the exact same position in Z-space as the TV.

There aren’t really any alternative ways of doing what you want to do – if you want to put a video clip into a TV screen, you just have to mask the clip to the shape of the screen. An approach that you might find useful, though, is one that I picked up from Mark’s excellent Mastering Motion’s Camera tutorial package; you build your TV, complete with clip, as a single 2D group, and *then* position the group in 3D space. That way, the TV, clip and mask will all move together and you won’t have any parallax issues between the TV and clip.

You can then just duplicate the group as many times as necessary and reposition the copies in 3D space, simply swapping out the clips for each one.

[Gareth Randall]

Here’s the thing, though; even though you *could* do it by actually putting the phone or whatever on a turntable (or rotating the camera around it), you just wouldn’t bother going to all that trouble. It’s just so much easier, cheaper and more flexible these days to do it in 3D.

Either the client provides the 3D model of the product, or you buy it from somewhere like The 3D Studio, and away you go.

[Gareth Randall]

Looks fine. Maybe I’d want a bit of differentiation between each panel with a colour tint or something, but that’s just a taste thing.

[Gareth Randall]

The way you describe it – 4 copies of the clip, each with a mask applied – would be the most obvious, straightforward and flexible way of doing it in just about any compositing system. If you want to avoid using masks, though, then you could use the Crop parameter on each instance of the clip to trim its width. Masks would be by far the best way of working, though.