Parametro Metodo Miscela

From Synfig Studio :: Documentation
Jump to: navigation, search
m
m
Line 3: Line 3:
 
{{Category|Parameters}}
 
{{Category|Parameters}}
 
<!-- Page info end -->
 
<!-- Page info end -->
The various compositing methods available for Layers in Synfig. (''What compositing/blending is.'')
+
Presentazione dei vari metodi di composizione disponibili per i Livelli in Synfig. (''Cos'è composizion/miscela.'')
  
 
In the following descriptions, 'A' refers to the color on the layer with the blend method setting, and 'B' refers to the color on the layers beneath it.  Note that in almost all layers, the alpha channel of the colors will have a scaling effect on the blending.  The 'amount' parameter will also have a scaling effect.  In most descriptions these 2 details have been glossed over.
 
In the following descriptions, 'A' refers to the color on the layer with the blend method setting, and 'B' refers to the color on the layers beneath it.  Note that in almost all layers, the alpha channel of the colors will have a scaling effect on the blending.  The 'amount' parameter will also have a scaling effect.  In most descriptions these 2 details have been glossed over.

Revision as of 17:00, 9 February 2012

Languages Language: 

English • français • italiano • română • русский

Warning!

This page contains outdated information.

The release of Synfig Studio 0.64.0 introduced new terminology and this translated page needs to be updated according to original English text.

You can help updating this page - see instructions here. Thank you!

warning end


Presentazione dei vari metodi di composizione disponibili per i Livelli in Synfig. (Cos'è composizion/miscela.)

In the following descriptions, 'A' refers to the color on the layer with the blend method setting, and 'B' refers to the color on the layers beneath it. Note that in almost all layers, the alpha channel of the colors will have a scaling effect on the blending. The 'amount' parameter will also have a scaling effect. In most descriptions these 2 details have been glossed over.

In the examples that follow a gradient ranging from black to white to transparent (A) is blended on the layer above an inline canvas of Tux on a transparent background (B).

A) straight.png & B) synfigtux.png

The currently available blend methods are:

Composite

This blend method is the default option, it simply displays the content of the layer. This blend mode is similar the layer blend mode Normal Mode often found in 2D programs.

straight.png + synfigtux.png = composite.png

Straight

This blend method looks similar to the previous one, except that the objects under a straight-object will be invisible. So if there is a more or less transparent object on the Straight mode layer, the objects on the layers underneath won't show through it.

More precisely, the resulting color is "(A-B)*amount + B". So if amount is 1 the result is A and if amount is 0 the result is B. In particular, if amount is 1 and A is a very transparent color, the resulting color will also be A; despite the fact that A is very transparent, none of B's color is used.

straight.png + synfigtux.png = straight.png

Onto

If a layer is set to the Onto blend method, only the parts of the layer that are over a not transparent area will be visible.

Precisely: this is the same as the Composite blend method except that the transparency of the resulting color is set to be the same as the transparency level of layer B.

straight.png + synfigtux.png = onto.png

Straight Onto

This method is a combination of the two methods above. E.g. if an half-transparent object is set to Straight Onto, it will only be visible over a non-transparent area, and the non-transparent part under that object won't be visible.

Precisely: the resulting color is "(X-B)*amount + B" where X is A but with its transparency set to A's transparency times B's transparency.

So if amount is 1 the result is A, but with its transparency multiplied by that of B, and if amount is 0 the result is B. In particular, if amount is 1 and A is a very transparent color, the resulting color will be a possibly more transparent version of A; despite the fact that A is very transparent, none of B's color is used in the result.

(Yuck. Are these 'precisely' comments useful?)[Yes!]

straight.png + synfigtux.png = straightonto.png

Behind

This blend method makes the layer visible over transparent areas, and invisible over non-transparent areas, giving the impression that the layer is behind the other layers. It is often used for the "Shade" effect layer, to make a drop-shadow effect.

Precisely: this is the same as the composite blend method, but with A and B swapped. B is composited onto A instead of A being composited onto B.

straight.png + synfigtux.png = behind.png

Screen

This blend method is similar to the Screen Mode often found in 2D programs. It combines the colors of the screen mode layer and the ones behind it, and gives a lighter result in general.

straight.png + synfigtux.png = screen.png

Overlay

This is similar to PhotoShop - layer blend mode Overlay Mode

Precisely: define 3 new colours: RM = A * B; RS = 1-(1-A)*(1-B); RET = A*RS + (1-A)*RM then blend RET onto B as in the Onto method above(!)

Any idea what that is aiming to do? Or what the layer does in this PhotoShop program? This appears to emulate the effect of a cross-fade between the two layers if they are set to equal amounts - i.e like 'add', but maintaining the overall brightness of the image

straight.png + synfigtux.png = overlay.png

Hard Light

This is similar to PhotoShop - layer blend mode Hard Light Mode

For each of red, green and blue, if the component is in the top half of its range then calculate X=1-(1-(2A-1))*(1-B), otherwise calculate X=2AB, then blend X onto B as in the Onto method above.

Is this aiming to make bright colours brighter and dark colours darker?

straight.png + synfigtux.png = hardlight.png

Multiply

This is similar to PhotoShop - layer blend mode Multiply Mode

Precisely: the resulting colour is (((A*B)-B)*amount+B). The calculation is performed independently on red, green, and blue components. When amount is 0, this simplifies to B. When amount is 1 it simplifies to A*B.

straight.png + synfigtux.png = multiply.png

Divide

Describe me

Precisely: the resulting color is (((B/A)-B)*amount)+B.

When amount is 0, this becomes simply B.

When amount is 1, this becomes B/A.

A very small quantity is added to A before dividing by it to avoid a divide-by-zero condition. This causes the divide blend method to bias toward positive values, but the effect is really negligible.

straight.png + synfigtux.png = divide.png

Add

Describe me

Precisely: the resulting color is (B + A*A.alpha*amount). The calculation is performed independently on red, green, and blue components. The resulting color's alpha is B.alpha.

straight.png + synfigtux.png = add.png

Subtract

Describe me

Precisely: the resulting colour is (B-A). The calculation is performed independently on red, green, and blue components.

straight.png + synfigtux.png = subtract.png

Difference

Describe me

Precisely: the resulting colour is the absolute value of (B-A). The calculation is performed independently on red, green, and blue components.

straight.png + synfigtux.png = difference.png

Brighten

Describe me

Precisely: for each of the red, green, and blue components, compare A's value with B's value and use the higher of the pair.

straight.png + synfigtux.png = brighten.png

Darken

Describe me

Precisely: for each of the red, green, and blue components, compare A's value with B's value and use the lower of the pair.

straight.png + synfigtux.png = darken.png

Color

Describe me

Precisely: the resulting colour is obtained by adjusting B to have the same U and V values as A, while keeping Y the same.

straight.png + synfigtux.png = color.png

As this example looks just like the saturation one, perhaps a yellow gradient instead of a white one would be more illustrative

yellowgradient.png + synfigtux.png = coloryg.png

Hue

Describe me

Precisely: the resulting colour is obtained by adjusting B to have the same hue as A.

straight.png + synfigtux.png = hue.png

Saturation

Describe me

Precisely: the resulting colour is obtained by adjusting B to have the same saturation as A. Saturation is the magnitude of the UV vector.

straight.png + synfigtux.png = saturation.png

As this example is so similar to the 'Color' blend example, here it is with a yellow gradient -

yellowgradient.png + synfigtux.png = saturationyg.png

Luminance

Describe me

Precisely: the resulting colour is obtained by adjusting B to have the same Y (luma) value as A, while keeping U and V the same.

straight.png + synfigtux.png = luminance.png

Alpha over

Describe me

straight.png + synfigtux.png = alphaover.png


Languages Language: 

English • français • italiano • română • русский