PDA

View Full Version : Applying reflectivity to a surface in VSE



reilly86
04-06-2011, 04:25 AM
Hi all,

Is there any way to make a surface reflective in the Virtual Set Editor? I know some of the stock sets have reflective surfaces (tabletops etc.) but I cannot see any way of applying this to my own objects. Is this possible using the software or by any other means?

Cheers in advance,
Reilly

joseburgos
04-06-2011, 08:19 AM
As of right now, the only way is to have two layers of the same Input., A or B. Then one is positioned as the main and the other layer is positioned as the reflective source. The reflective need to manually be moved and angled to match the background layer (say the floor in a room) and then the transparency can be lowered to make it look like a reflection.
For a floor reflection may not be to hard but to make the top of a desk, for example, reflective, it is a bit more work. Because you can't matte/mask the Input layer, you would need to position, scale and rotate the reflective layer to match the parallax of the desktop. Then load on top of that a image of the same background but with the desktop matted out. This will then leave only the reflective layer on the desktop only.
Hope I made sense but if not, feel free to ask more questions.

Take care,

kltv
04-06-2011, 09:49 AM
You can mask or matte the input layer if you are importing a Photoshop file. If you want to do a reflection that only falls on the desk, make another "Input A" layer and fill it with the region of the desktop. When you import the PSD file, the input video will be confined to that region only. You can even use a gradient if you want a "fall off" to the reflection. Then you do what Jose suggested, rotate, position and scale the video on the the desktop and tweak the opacity to taste.

Kris

joseburgos
04-06-2011, 09:53 AM
That is what I tried to articulate, you matte/mask the image layers below and above to constrain the reflective Input A or B layer.
What I may have confused the issue was saying that we can't matte/mask the Input layer therefore we need to do it this way.
Thanks Kris and hope this all makes sense to Reilly :)

reilly86
04-06-2011, 10:55 AM
Yep, that's done the trick. Didn't think of doing it that way!

Many thanks, guys. :thumbsup:

kltv
04-06-2011, 11:09 AM
That is what I tried to articulate, you matte/mask the image layers below and above to constrain the reflective Input A or B layer.
What I may have confused the issue was saying that we can't matte/mask the Input layer therefore we need to do it this way.
Thanks Kris and hope this all makes sense to Reilly :)

Hi Jose,

What I'm saying is that you can directly mask the "Input A" layer itself. You don't have to put something back on top to block it. Whatever shape or transparency you create in Photoshop for the "Input A" layer will be applied to that layer. So if you make an "Input A" layer, use the polygon selection tool to make a selection the shape of the desk, then fill it with a color, when you import that PSD file into the VSE, only the region from the polygon fill will show input A video. Does that make sense? Basically, the alpha channel of the "Input A" and "Input B" layers are respected and applied to the video layers. You have to apply any layer masks first, but you are making it harder on yourself than you need to if you are putting stuff back on top just to mask video... :)

Kris

joseburgos
04-06-2011, 11:17 AM
Hemm...
I will need to re-test that as when I did this, it made a keyhole but I guess that would work now that I think about it and exactly what you are explaining. The problem is, since you are faking the parallax for the reflection, you can't see the rotation and scaling as well.
Anyway I will look it over again, the matte/mask in the Photoshop layer as a way to help train clients on how to make simple LiveSets.

Thank you Kris

John Perkins
04-08-2011, 10:47 PM
There are three ways to make a mappable area in the PSD.

1. Leave the input layer completely blank - the mapping area is full screen
2. Fill the input layer with any solid color - the mapping area is full screen
3. Like #2, but leave transparent/semi-transparent areas.

This will allow you to do oddly shaped areas with soft edges without having to put another image layer over it.

When VSE loads the PSD, it will do a best guess of how to position the video inside the used area. It usually won't be perfect, but it will be in the ballpark ready to be adjusted.

You do have to be careful though, if you attempt to make a blank fullscreen layer and accidentally leave a single stray pixel, you'll literally get a single pixel map.