Darby Edelen
Forum Replies Created
-
Think of the masks as alpha mattes with blend modes instead.
When you have a mask set to ‘Add’ and Invert it you are essentially telling AE “fill my alpha with white except inside this mask.” As soon as you add another mask set to ‘Add’ and Invert it you will be filling every part of the alpha with white except where those two masks overlap (because that is the only area that is not being filled with white).
The ‘Subtract’ mode is preferable because it will always black out the alpha inside each mask.
The reason to add a layer sized ‘Add’ mask before using ‘Subtract’ masks is that there’s a long standing unexpected behavior in AE when you reduce the mask opacity. Instead of the subtracted area becoming less subtracted, the area outside of the subtracted area becomes more subtracted. Unless, of course, you have a layer sized ‘Add’ mask above your ‘Subtract’ masks 🙂
Darby Edelen
-
Don’t invert the masks. Instead set them to the Subtract mode.
If you want to be super diligent then you should first apply a layer sized rectangular mask to the layer set to Add, then create your Subtractive masks.
Darby Edelen
-
[Paulo Jan] “-Set given pixels to a certain color, based on their brightness level (as in the image above).”
You could use one of the keyers (Color Range? Extract?) and Channel > Solid Composite to do that.
[Paulo Jan] “-Take a snapshot of the current comp and perform calculations on it as a bitmap, and then save it as a JPG.”
This part has me a little confused. If you want to perform calculations on the image inside of AE then I would expect that is most likely scriptable depending on what the goal is. What is the goal?
Darby Edelen
-
It’s possible to link some values and textures through expresso but it’s not as user friendly as the hypershade. You might like the free cmNodes plugin:
Darby Edelen
-
[Scott Johnson] “The above forum contains information on the video in question that I am trying to edit. I think what I am saying is, why should a green screen or green background have to be near perfect for it to be used easily, quickly, and effectively. If you check out the video I am trying to edit, I don’t see a non-rotoscope way of doing this easily at all. But I figured if you can simply extract all pixels that utilize the green channel at a particular level, then that (along with feathering) might be an effective tactic and you follow it up with deletion of transparent pixels.”
Chroma keying relies on the differences between the Red, Green and Blue components to generate an alpha. When keying Green the ideal screen is one where the Green component is brightest and there is a large gap between it and the Red/Blue components in the screen.
The grass in your image has more Red than Green in it. That’s immediately problematic. The changes in luminance across the trees in the background also correspond to a smaller gap between the Green and R/B components.
This is not a situation in which chroma keying will be particularly effective. You can create transparency from this image using keyers, but it likely won’t look good.
Darby Edelen
-
Not possible.
I’d suggest using a single Slider control rounded to integers instead.
Darby Edelen
-
Blend modes might get you there.
If you’re working at 32bpc you could use multiply with an exposure effect on the ‘color’ footage to control intensity. Otherwise try soft light or overlay if you’re working in a lower bit depth. If there are details in the color footage you don’t want to transfer then use a blur as well.
Darby Edelen
-
The clip black property of keylight will do this if you increase it to near 100.
I don’t think this is really what you want to do. I can promise you that you’ll end up with poor edges and holes in your subject. It’s a given that some of your subject is going to be partially transparent when initially keyed.
Darby Edelen
-
I’d put all of the Photoshop files into one composition, trim them each to be 1 frame long and then use Animation > Keyframe Assistant > Sequence Layers… with no overlap. This will place each layer after the previous. The composition’s duration will need to be long enough to contain all of the PSD files in sequence (number of layers * 1 frame per layer = number of frames of duration for the composition).
Then take put that composition into a new composition and apply time-remapping to it. Apply an Effect > Expression Controls > Slider Control to the layer and then apply this expression to the time-remapping on the layer:
n = Math.floor(effect("Slider Control")("Slider"));
framesToTime(n);Darby Edelen
-
Do you want the layer to jump a certain distance when a loop completes or move a distance over the life of the loop?
Darby Edelen