ICDL Image Editing (ICDL) cheat sheet

Raster images store pixels on a rectangular grid; each pixel holds color values and optionally alpha, and enlarging beyond native pixel dimensions requires interpolation.

Vector images store paths and shape attributes; they remain resolution-independent until rasterized to pixels for display or output.

Image pixel dimensions are width × height in pixels; PPI relates those pixels to a physical inch when printing and is not an inherent physical size of the file.

Color depth is bits per pixel or per channel: 8-bit grayscale has 256 levels; 8 bits per RGB channel gives 24-bit color and about 16.7 million combinations.

RGB is additive light mixing for displays; CMYK is subtractive ink mixing for print; RGB-to-CMYK conversion depends on profiles and can shift out-of-gamut colors.

For normalized alpha compositing, C_out = αC_foreground + (1 − α)C_background; α = 0 is fully transparent and α = 1 is fully opaque.

A histogram plots pixel counts by tonal level; clipping at the shadow or highlight end indicates lost detail, but the histogram gives no spatial location.

Lossy compression discards image data and can accumulate artifacts after repeated decode-edit-encode cycles; lossless compression preserves pixel values exactly.

Metadata includes EXIF camera settings, IPTC captions, and XMP rights information; metadata can be preserved or stripped during export depending on settings.

The exposure triangle links aperture, shutter speed, and ISO; changing one factor alters exposure and also changes depth of field, motion blur, or noise.

Aperture f-number = focal length ÷ effective aperture diameter; a smaller f-number means a wider aperture and generally shallower depth of field.

A one-stop exposure change doubles or halves light: shutter 1/125 s admits half the light of 1/60 s, and f/2.8 admits twice the light of f/4.

Shutter speed controls exposure duration and motion blur; faster speeds freeze motion, while slower speeds blur moving subjects and increase camera-shake risk.

ISO amplifies the sensor signal; higher ISO permits faster shutter or smaller aperture but typically raises noise and reduces dynamic range.

White balance compensates for light color: tungsten is about 3200 K, daylight about 5500 K, and open shade is higher, so auto white balance can vary between frames.

Optical scanner resolution is true sampled ppi; interpolated resolution adds estimated pixels and does not add real detail.

Megapixels = pixel width × pixel height ÷ 1,000,000; cropping reduces both dimensions and therefore reduces total megapixels.

Camera JPEG is typically 8-bit lossy; RAW preserves sensor data and offers more latitude for exposure and white-balance edits.

Canvas size changes the working area without resampling existing pixels; enlarging adds space, while reducing crops content, whereas image size can resample pixel dimensions.

A new document needs pixel dimensions, color mode, bit depth, and background choice; print output also needs an intended ppi and color profile.

Layer stacking order determines occlusion; upper opaque layers hide lower ones, and blend modes change how layer pixels combine with underlying pixels.

A grayscale layer mask uses black to hide, white to reveal, and gray for partial visibility; masks are non-destructive because pixel data remains intact.

Selection feathering creates a gradual edge; anti-aliasing smooths jagged raster edges by partially covering boundary pixels.

Adjustment layers apply tonal or color changes non-destructively and affect layers below within the same group or stack, subject to masks and clipping.

Type layers remain editable text; rasterizing type converts glyphs to pixels and removes text-editing capability.

Vector shapes and paths are resolution-independent; paths can be converted to selections, masks, or stroked outlines.

Hexadecimal web colors use #RRGGBB, where each pair is 00–FF hexadecimal for red, green, and blue; #FFFFFF is white and #000000 is black.

Cropping removes pixels outside a boundary and reduces dimensions; rotating or straightening may require interpolation or cropping to avoid empty corners.

Resizing with resampling changes pixel count; resizing without resampling changes print dimensions and ppi while keeping stored pixels unchanged.

Transform operations include scale, rotate, skew, distort, and perspective; repeated raster transformations can soften detail through interpolation.

Clone stamp copies sampled pixels; healing brush copies texture while blending color and luminance with surrounding pixels.

Dodge lightens and burn darkens targeted tones; sponge increases or decreases saturation; all three can be applied through selections or masks.

Gaussian blur uses a radius; Unsharp Mask uses amount, radius, and threshold, where threshold limits sharpening to edges above a tonal difference.

Levels remaps input black, gamma midpoint, and white to output levels; Curves remaps tones along an editable curve for finer contrast and color control.

Selection operations include grow, similar, contract, expand, and feather; quick mask temporarily represents a selection as a paintable grayscale channel.

Non-destructive retouching uses duplicate layers, smart objects, masks, and adjustment layers so original pixels remain recoverable.

Format decision rule: JPEG for lossy photographic web output without alpha; PNG for lossless graphics with alpha; GIF for 256-color graphics or simple animation; TIFF for flexible lossless print or archive; PSD for layered editing; SVG for vector.

Save usually writes the current native format and may preserve layers; Save As can change format; Export creates a derivative, often flattened or optimized for web.

Print size in inches = pixel dimensions ÷ output ppi; required pixels = desired inches × ppi, so 3000 px at 300 ppi prints 10 inches wide.

Downsampling should use an interpolation method suited to content; sharpening after downsampling can restore apparent edge crispness lost in resampling.

JPEG does not support alpha transparency; PNG, WebP, and TIFF can carry alpha, while GIF supports binary transparency and a 256-color palette.

ICC profiles describe device color behavior; converting to an output profile and embedding it helps preserve intended appearance across systems.

Metadata such as EXIF, IPTC, and XMP can be retained or stripped at export; stripping removes camera settings, captions, and rights information.

Lossy quality settings trade file size against artifacts; progressive JPEG can improve perceived loading on slow connections but does not add transparency.

Web output is commonly sized by pixel dimensions and compressed to a target file size; physical ppi is irrelevant to on-screen pixel dimensions.