Darby Edelen
Forum Replies Created
-
[Oliver Grice] “like tell me realistic targets of what you would like me to do”
That depends a lot. What would you like to do?
Start with a focused exercise. Have you ever wondered “how would I do that?” while watching a movie or TV? Perfect! Now try to do it 🙂
Alternatively, think of content or a technique that interests you. Outerspace? (How would I make a sun? a star field? a nebula?). Compositing and Effects? (How could I make someone teleport? burst into flames? sprout wings and fly?). I’d recommend starting small but don’t limit yourself creatively.
This forum as well as other resources will be available if you need feedback or get stuck.
Darby Edelen
-
There are a few solutions. John covered a good one, but if you have lots of text layers it can grow unwieldy.
Another option would be to pre-compose your text layers, create your gradient layer above them in this pre-comp and enable its “Preserve Transparency” switch (in between the blend mode and track matte drop downs on the layer).
Yet another option would be to pre-compose the text, then make your gradient in the main composition, apply the ‘Set Matte’ effect to the gradient layer and point it to the text pre-comp. This option would work well if, for example, you needed your text pre-comp to have collapse transformations enabled as preserve transparency won’t work in that case.
Darby Edelen
-
[Adam Trachtenberg] “He can just place all the cloner groups in a null, load the texture pattern jpeg in a material, and then place the material on the null. Set the projection to flat mapping and use the texture axis tool to rotate and scale it into position.”
Personal preference I suppose. Either way the positioning of the projection is a bit of a PITA but I prefer the uniform coloring that the Shader effector provides as well as the additional flexibility of allowing for movement along any axis if the need should arise.
Darby Edelen
-
You should be able to use a Shader effector to drive a material with the Mograph > Color shader applied to a cloner object set to Grid Array (this is a different set up than what you’re currently using).
This will give each of the tiles a solid color based on the average color value sampled by the Shader effector. The texture you provided with your project would have to be altered in order to work for this purpose (there are white sections in the JPEG that may cause the sampling to go a bit nutty).
Darby Edelen
-
[Mike Imam] ”
Any idea how to prevent these modes from affecting the main timeline (i.e. I only want them to apply to the layers in my pre-comp)?”It sounds like you have the “Collapse Transformations” switch enabled on the pre-comp in your main comp. You’ll have to disable that if you don’t want those layers punching holes in your main composition.
Alternatively, if you need collapse transformations enabled, find another solution other than Silhouette Luma to create alpha from luminance. I like the “Set Matte” effect. You won’t be able to use an adjustment layer to apply it, however, as this will still affect the main comp due to collapse transformations. Instead, target the application of Set Matte specifically to those layers you need to matte.
Darby Edelen
-
Indeed, this is not a limitation of AE but a limitation of the codec you’ve chosen in the output module.
Darby Edelen
-
More triage might be helpful in this case.
Are you trying to match a 3D camera move to footage, that is to say: is this a camera tracking question? It doesn’t sound like it from your description of animating the camera, but if it is then it sounds like the track is bad.
Do the 3D layers appear to be sliding relative to other 3D layers? If that’s the case then I highly recommend placing layers relative to one another by creating Parent/Child relationships.
When you make a layer a child of another layer it’s coordinates are relative to the parent layer’s transformations. This makes it very easy to, for example, place footage of a person (child layer) onto a floor (parent layer).
Another problem you might be facing is perspective present in the footage items you are using. Most footage of people, for example, is not shot so that their feet are “flat” on the ground. If you move your virtual camera too far from the actual perspective of the recorded image, then the perspective mismatch will most likely be evident.
If none of this helps then I’d suggest posting a few pictures or a render of the problem.
Darby Edelen
-
Well, I checked in CS5 and CS6 and the smallest resolution available in both of those is 4×4.
What version are you using?
Darby Edelen
-
Most procedural textures allow you to define the coordinate system that is used to generate the texture. The ‘Noise’ shader defaults to ‘Texture’ Space.
You can switch this to ‘World’ but based on your description I think you might like the results with ‘Object’ better. If you use ‘World’ then the noise will not move or scale with your object. If you use ‘Object’ then the noise will move with the object but not scale.
Darby Edelen
-
I’d recommend refreshing yourself on Keylight by giving the documentation a read through. Lots of good information in there. If you google search for “keylight documentation” there should be a PDF result from The Foundry.
You may not need to mask separate sections of this image in order to get a good key. However, you definitely will need to key it at least twice, maybe even three times. With your particular image I saw much better results using a lower Screen Balance value (around 50).
While adjusting these properties I recommend viewing the ‘Status’ which will show you all fully transparent (black), fully opaque (white), and semi-transparent (gray) areas.
Key once for the edges, preserve softness (but don’t add any) and don’t worry about “holes” in the foreground subject. This instance of Keylight won’t need much adjustment in the Screen Matte section.
Key again for the core of the foreground subjects, this is where you’ll use Clip White to make sure the foreground is solid. Since this instance is only for the core of your foreground subject, use Screen Shrink to erode the matte by a pixel or two.
The combination of these two mattes should be good enough for your main subjects. I would usually just lay the edge key layer on top of the core key layer unless more complex work is required.
If you need a better shadow key then you could do that separately as well. Most likely your edge key will preserve the shadows pretty well, so you could simply duplicate that and make some adjustments to preserve more shadow. If you’d like to avoid color spill in the shadows I’d recommend just using Generate > Fill with a black color
Also you’ll likely need to use the Despill Bias to remove blue spill from your subjects. I like to click to open the color picker and then use the eye dropper from there to select a foreground color (try a point in skin of your subject) for this operation.
Darby Edelen