Split Color Channels
See a picture's RGB channels or CMYK plates side by side, and save any of them.
or browse filesdrop an image here or paste
.png
See a picture's RGB channels or CMYK plates side by side, and save any of them.
or browse filesdrop an image here or paste
Every photo on your screen is three greyscale pictures stacked on top of each other: one for red, one for green, one for blue. This page takes them apart. Drop an image and you get the whole set at once - the original and its red, green and blue channels side by side on one sheet, each one labelled. Switch Show from All to a single channel for colour isolation: that one plate on its own, at the picture's full size. Show in channel color is the difference between the plates as they are really stored (grey) and the tinted version most people recognise, and the download always follows what is on screen.
This is the splitter. If you want to push the channels around instead of separating them - more red, less blue, a colour cast taken out - that is the RGB channel editor.
The Color model switch decides what a channel means. RGB is the light the screen emits, so the plates are red, green and blue, and a picture that carries transparency gets a fourth one: the alpha mask, white where the image is solid and black where it is see-through. CMYK is the ink a press lays down, so you get the four separations - cyan, magenta, yellow and black - computed the plain way from sRGB, with no ICC profile and no dot gain, which is the right answer for a mock-up or a school assignment and not for a print run. For a real profiled conversion of the whole image use /rgb-to-cmyk/, and for the transparency mask on its own there is /alpha-channel-extractor/.
Photographers use the split to see which channel holds the detail: skin is smoothest in red, foliage separates in green, and a blue channel is usually the noisiest thing in a night shot. Designers use it to pull a clean mask out of a picture whose subject sits against one strong colour.
Downloading gives you exactly what the stage shows: the whole sheet as one image when All is selected, or that single plate at the source's own size when one is. PNG keeps the values exactly as they are - the right choice when the plate is going back into an editing program - while JPG and WebP are smaller and fine for showing someone. The fourth entry in the format menu is ZIP, one file per channel: it writes every plate of the current model as its own full-size PNG, named after your file, so a single click gives you photo-red.png, photo-green.png and photo-blue.png. You can also copy the view straight to the clipboard, or share it from a phone.
The channels of a photograph are the photograph - a red plate of a document or a passport is still readable, and an alpha mask still shows the shape of whatever was cut out. None of it leaves the device. The file is decoded, split and encoded again by your own browser, the sheet you see is drawn on a canvas in this page, and the ZIP is built in memory here as well. There is no upload step, and closing the tab is all the cleanup there is.
Because that is what a channel is. Each one stores a single number per pixel - how much red, how much green, how much blue - and a single number per pixel is a greyscale image. Tinting it red is a presentation choice, not the data, which is why the toggle is there. Turn Show in channel color off and you are looking at the actual stored values; turn it on and each plate is drawn in the light it stands for, so the three of them add back up to the original.
The black plate takes as much of the picture as it can - it is computed first, as whatever is missing from the brightest of red, green and blue - and the three colour plates only carry what is left over. A bright, colourful photo therefore lands mostly on black with light cyan, magenta and yellow on top. A press operator would redistribute that with under colour removal and a profile for the specific paper and inks; this page does the plain sRGB separation instead, so the numbers are predictable.
Not here - this page only takes an image apart. Swapping red for blue, turning one channel up or down against the others is the job of the RGB channel editor, which gives you a gain slider per channel and writes the result back as a normal picture. It does not swap channels or use one as a filter over another. Split first if you want to see what you are working with, then adjust there.
No. Your picture is read, separated and saved entirely inside this browser tab - there is no server that could receive it, so the plates of a private photo, a scan or a design never exist anywhere but on your own device.