lisztfr9
5th November 2012, 14:42
If you need a tutorial, write it.
The question is, for example, how actually see your overlay mask, since the mask itself is not visible. Only it's effects are.
>-----< mask
>-----< layer
>-----< background
The dark parts of the mask will erase the underneath parts of the layer, the white parts will preserve them ; In order to see the layer mask, one should turn the layer into a piece of solid white color, which will receive the impression of the mask thoroughly. But this isn't enough, since the layer itself is merged in a way (blend, subtract, etc) with the background clip, so this clip should also be turned to a solid color.
This, because it's a difficult task to actually tweak your layers as you want, because of the lots of parameters.
Imagine you want just to reduce a white (not moving) spot. You want to draw a square around, and make a copy from that for using as layer. Ok, now you have many possibilities, setting levels and contrast on the layer mask, on the layer, and set a blending mode for the layer. Trial and error is not a good idea.
First thing, discard the layer mask. Focus on the layer, reduce it's gamma level to make the white spot acceptable (in blend mode, overlay). Then the borders of the spot will look to dark, therefor we need the layer mask. Now the difficult part is to set the luma range for the layer mask in order to make the layer just act as needed.
I have no method for that. Obviously playing with levels won't do the trick, since the whole range of luma is affected in a linear way, or almost. So one should try Tweak (contrast) on YUV12 before entering in the mask_clip stage, because it's RGB and Tweak is YUV. Or trying YLevels since it's not linear.. or mt_lut with range processing... All empirical methods. Therefor it's good to be able to see the mask.
L
The question is, for example, how actually see your overlay mask, since the mask itself is not visible. Only it's effects are.
>-----< mask
>-----< layer
>-----< background
The dark parts of the mask will erase the underneath parts of the layer, the white parts will preserve them ; In order to see the layer mask, one should turn the layer into a piece of solid white color, which will receive the impression of the mask thoroughly. But this isn't enough, since the layer itself is merged in a way (blend, subtract, etc) with the background clip, so this clip should also be turned to a solid color.
This, because it's a difficult task to actually tweak your layers as you want, because of the lots of parameters.
Imagine you want just to reduce a white (not moving) spot. You want to draw a square around, and make a copy from that for using as layer. Ok, now you have many possibilities, setting levels and contrast on the layer mask, on the layer, and set a blending mode for the layer. Trial and error is not a good idea.
First thing, discard the layer mask. Focus on the layer, reduce it's gamma level to make the white spot acceptable (in blend mode, overlay). Then the borders of the spot will look to dark, therefor we need the layer mask. Now the difficult part is to set the luma range for the layer mask in order to make the layer just act as needed.
I have no method for that. Obviously playing with levels won't do the trick, since the whole range of luma is affected in a linear way, or almost. So one should try Tweak (contrast) on YUV12 before entering in the mask_clip stage, because it's RGB and Tweak is YUV. Or trying YLevels since it's not linear.. or mt_lut with range processing... All empirical methods. Therefor it's good to be able to see the mask.
L