Composite

This transformation combines an overlay color or image with the current image. It can create a range of effects from simulating multiple exposures to complex compositing operations like placing an object in one image behind an object in another.

If you don’t need to align the two images and they are the same size in pixels, you can use the simpler Blend transformation instead of Composite.

Amount

The Amount control lets you control how much of the composite transformation is applied to the input image. You can apply a percentage of the transformation to the entire image, or you can specify an amount mask to restrict the effects of the transformation to only part of the input image. This can be used to composite an object from the overlay image "behind" an object in the input image by protecting the object with a mask.

Overlay and Overlay Amount

This control lets you select the overlay color or image.

If you select a solid color, then that color is applied as a semitransparent overlay over those parts of the input image not protected by the amount mask. Picture Window will not let you enter an overlay amount mask or set alignment points.

If you select an overlay image, you can use an overlay amount mask to control what parts of the overlay image are painted over the input image and the transparency of the overlay image. This mask can be used to isolate an object in the overlay image and composite just that object over the base image. You can also reposition, rotate, scale, and even shear the overlay image to precisely align specific points in the overlay image with corresponding points in the input image.

Operation

This control lets you select how the overlay color or image is combined with the input image.

Operation

Blend -- the overlay image is blended with the input image with a transparency determined by the amount and overlay amount controls.

Lighten/Darken -- the overlay image is blended with the input image, but only where it lightens or darkens the input image.

Filter -- the overlay image is applied to the input image as a filter (similar to the way the Filter transformation works). This differs from a blend which is like a double exposure while filtering is like sandwiching two transparencies. Note that dark parts of the overlay image are opaque and light areas are transparent as opposed to subtractive filtering (see below).

Subtractive Filter -- the overlay image is applied to the input image as a subtractive filter (computed as the negative of filtering the negative of the base image with the negative of the overlay image). This is similar to the normal (additive) filtering except that light areas of the filter are opaque and dark areas transparent.

Soft Light -- a special effect is created where the overlay and input image are combined so as to reverse each other.

Hard Light -- a special effect is created where the overlay and input image are combined so as to reverse each other.

Negative -- a special effect is created where the overlay and input image are combined so as to reverse each other.

Add -- the overlay image is added to the input image. This can generate out-of-range pixel values (outside the range 0..255) which are then clipped.

Subtract -- the overlay image is subtracted from the input image. This can generate out-of-range pixel values (outside the range 0..255) which are then clipped

Absolute Difference -- similar to Subtract except that negative differences (i.e. where the overlay image is brighter than the base image) are converted to positive instead of being clipped to black.

Register -- the overlay image replaces the input image and the amount and overlay amount controls do not blend the input image. The areas of the input image which are not covered by the overlay image show up as black in the result image. This option is normally used to warp the overlay image into alignment with the input image.

Alignment

This control lets you select how the overlay image is aligned with the input image. Unless alignment is set to None, labeled alignment points are displayed over the input and overlay images. You can drag these points to position them -- to move all the alignment points at once, hold down both Ctrl and Shift while moving the cursor. A box is also displayed over the input image that shows where the edges of the overlay image will fall.

To select alignment points on the base image, click the Composite thumbnail in the image browser. To select alignment points on the overlay image, click the smaller thumbnail just to the right of the Composite thumbnail in the image browser to select the auxiliary input image. Setting the display to split screen is useful so you can position alignment points and see the result at the same time.

None -- the overlay image is placed directly over the input image with their lower left corners aligned.

1-point (shift) -- the overlay image is shifted (without being scaled or rotated) so as to bring one point in the overlay image into alignment with one point in the input image.

2-point (shift/rotate) -- the overlay image is shifted so as to bring control point 1 in the overlay image into alignment with control point 1 in the input image and the line between control point 1 and control point2 in the overlay image is aligned with the line between control point1 and control point 2 in the input image, thus shifting and rotating the overlay image without altering its size. Consequently, control point 2 in the overlay image is not necessarily brought into coincidence with control point 2 in the input image.

2-point (shift/rotate/scale) -- the overlay image is shifted, scaled, and rotated as necessary so as to bring two points in the overlay image into alignment with two points in the input image.

3-point (shift/rotate/scale/shear) -- the overlay image is shifted, scaled, rotated, and skewed as necessary so as to bring three points in the overlay image into alignment with three points in the input image. This type of warping is called an affine mapping.

4-point (perspective warp) -- the overlay image is warped using a perspective mapping to bring four points in the overlay image into alignment with four points in the input image.

This tool bar is only displayed if Alignment is set to something other than None. Clicking any of the arrow buttons shifts the locations of all the alignment points in the input image by one pixel in the direction of the arrow. If you hold down the Shift key while clicking one of the arrows, it shifts by half a pixel instead. This can be useful for fine tuning the locations of the input alignment points.

Using Composite with a Solid Color Overlay

This blends the input image with a solid color overlay. You can either vary the transparency of the color overlay uniformly over the entire image or you can apply the overlay only to selected parts of the input image. If you select Blend, the composite transformation is similar to making a double exposure. Selecting Filter simulates placing a colored filter over some or all of an image, or you can use the Filter transformation instead.

To composite with a solid color, click on the Overlay button and choose Solid Color. This pops up a color picker you can use to select the overlay color.

To blend the color overlay with the entire input image

Use the Amount control to select the opacity of the overlay. A 50% level gives a 50-50 blend of the input image with the overlay color; 100% gives the overlay only and 0% gives the input image only. Blending with a light color over the entire image gives a result like haze or a uniform light leak in your camera. Blending with a dark color darkens the entire image.

To blend the color overlay with part of the input image

Use the Amount control to create a mask for the input image that isolates the parts of the image you want to overlay. Use the double slider to select the opacity of the overlay over the masked and unmasked regions of the input image. Intermediate values blend the overlay color with the input image; blending with a dark color is a good way to darken parts of an image. Using a hard-edged mask, you can get results like pasting opaque cutouts of colored paper over the input image.

Using Composite with an Overlay Image

This technique lets you overlay one image over another. You can vary the transparency of the overlay image using the Overlay Amount control, and you can also restrict the effects to certain parts of the input image with the Amount control. Finally, you can also shift, rotate, scale, and/or skew the overlay image to make it line up with the input image.

Start by selecting the overlay image into the Overlay control.

By making the overlay image semitransparent, you get effects like a double exposure. By using a hard-edged mask to control the overlay transparency, you get an effect like a collage or photomontage. You can even place an overlay object "behind" an object in the input image by using two masks: one for the input image and another for the overlay image.

Overlay image and input image are the same size

This case is a lot like overlaying with a solid color (see above), except that you are now overlaying with a color image instead. Use the Amount control to set the opacity of the overlay image. Intermediate values give a blend of the two images. You can use a mask to selectively overlay different parts of the input image with different amounts of the overlay image. The Amount mask selects at each point the proportions of each image to blend into the result. A soft mask blends the images smoothly; a hard mask makes abrupt transitions between the images.

Overlay image and input image are different sizes

This case is a little more complicated because you need to define where to place the overlay over the input image as well as its opacity. To overlay an object in the overlay image with the input image, first use the Overlay Amount control to create a mask for the overlay image that isolates the object you want to overlay. Use the double slider to vary the opacity of the object and the background, and position the object over the input image using one or more Alignment points. To place the object behind an object in the input image, create a mask for the object in the input image and select it into the Amount control. This mask protects parts of the input image from the effects of the transformation.

For example, to overlay an image of the moon over a photograph of a mountain, start by creating two masks: one that precisely covers the moon in the overlay image and one that covers the silhouette of the mountain in the input image. Select the mountain mask into the Amount control, the image of the moon into the Overlay control and the moon mask into the Overlay Amount control. Adjust the double sliders as necessary to get the result you want. You can position the moon anywhere you want in the input image and even vary its size and orientation by selecting 2-point alignment.

Formulas

For information on the formulas corresponding to each Operation, see the Blend transformation.