Raising the contrast of a matte to make it look tidy, which crushes the soft grey values that made hair look natural.
Editing and retouching
Matte
Also called alpha matte, matte channel or trimap. Here is what it means, when it changes what you export, and what people get wrong.
Editing and retouchingRuns on your deviceUpdated 10 August 2026
The short answer
A matte is a greyscale image that records how opaque every pixel of a subject is: white for fully solid, black for fully transparent and grey for everything in between. Compositing software reads it to decide how much of the foreground to blend over a new background.
Updated
The short version
Matte at a glance
| Quick fact | Detail |
|---|---|
| Also called | Alpha matte |
| Appearance | A greyscale image, white solid to black empty |
| Comes from | A matting algorithm or a keyed studio shot |
| Strength | Hair, fur, smoke, glass, motion blur |
| Optional input | A trimap marking the uncertain edge band |
| Origin of the term | Film and optical compositing |
Slide the table sideways to see every column.
How is a matte different from a mask?
The two words describe the same greyscale data seen from different ends of the job. A mask is a control you paint, refine and switch on and off inside a layered document. A matte is the finished opacity map, the thing a compositor is handed and treats as data rather than as something to brush by hand.
The term comes from film. Before digital editing, a matte was a physical or optical blocker that stopped part of the frame exposing so another image could be printed into it, and a travelling matte was one that moved with the actor. Blue and green screen work inherited the vocabulary, which is why the process of extracting a subject is still called matting and the extraction algorithm is a matte puller.
Why does a matte handle hair and glass better than a hard cut-out?
Because it stores partial opacity. A camera does not record a clean boundary between subject and background: at the edge of a strand of hair, one pixel contains light from both. Compositing describes that pixel as a mixture of a foreground colour and a background colour, blended in proportion to the matte value.
A binary cut-out has to pick one side for every pixel, so it either eats the hair or keeps a rim of old background. A continuous matte can say a pixel is 30 per cent hair and 70 per cent background, and the compositor then mixes 30 per cent of the true hair colour over the new backdrop. The same reasoning covers smoke, netting, motion blur, and the semi-transparent bodies of glass and plastic bottles.
What is a trimap and why do matting tools ask for one?
A trimap is a rough three-region guide drawn over the photo: definite foreground in white, definite background in black, and a grey band of uncertainty over the edges where hair and soft detail live. The matting algorithm is told not to argue about the white and black regions and to spend all of its effort estimating opacity inside the grey band.
Classic matting methods needed a trimap supplied by hand, which is slow. Most current tools predict the uncertain region themselves from a segmentation step, so you get a matte without drawing anything, and only fall back to a manual trimap when the automatic result gets an edge wrong.
What people get wrong about matte
Each one is a real failure mode, not a style preference.
Assuming a good matte alone fixes the edge, when the foreground colours behind it still need decontaminating or a halo appears.
Drawing a trimap with the uncertain band too narrow, so the algorithm has no room to solve the hair it was meant to solve.
Questions people ask
Matte, answered
The follow-up questions people search for once they have the definition.
Is a matte the same as an alpha channel?
Nearly. A matte is the greyscale opacity map itself, and the alpha channel is where that map ends up when the image is saved. You can view a matte on its own as a black and white picture, whereas the alpha channel is a slot inside the file format.
Why does my matte look grey all over instead of just at the edges?
Usually because the subject is genuinely semi-transparent, such as glass or sheer fabric, or because the extraction has low confidence across a low-contrast area. Grey inside a solid product is a fault, and it shows up as a washed-out subject once composited.
Do I need a trimap to remove a background?
Not usually. Modern matting models estimate the uncertain edge region themselves. A hand-drawn trimap is worth the effort when an automatic result loses fine hair or keeps a chunk of background, because it tells the algorithm exactly where to concentrate.
Same cluster
More from the glossary
Neighbouring entries, so every term in the set is one hop from every other.
Megapixel
A megapixel is one million pixels.
Model quantisation
Quantisation stores and computes a model's weights at lower numeric precision, typically dropping from 32-bit floats to 16-bit floats or 8-bit integers.
Lossy compression
Lossy compression shrinks an image by permanently throwing away detail your eye is least likely to notice.
From the studio
Tools that do this
Each one runs on your own device, so the photo stays where it is.
Keep reading
Elsewhere on Photomop
Platform image rules
Image size, file type and background rules for 22 selling platforms, each one sourced.
Read itConvert image
JPG, PNG, WebP and AVIF, with an honest answer about HEIC and alpha.
Read itAll the tools
The whole toolbox in three families, with what each one needs before it can run.
Read itOn-device privacy
The pipeline in five steps, what leaves your device, and how to check it yourself.
Read it
Now do it to a photo
Photomop is a photo studio that runs on your own device. Resize, crop, compress, convert, batch edit, remove a background and change one all work in the browser tab you are reading this in, at full resolution and with no watermark. The on-device page shows you how to check that the photo stays put.