Lossy
Throws information away to get smaller.
JPEG, WebP and AVIF discard detail the eye is least likely to miss. Each save discards more, so repeatedly editing and re-saving a JPEG degrades it every time. Keep an original and export copies.
The words, explained.
Image terms glossary
Plain-English definitions of the image and file-format terms you meet when converting and compressing files, including lossy, alpha, DPI, EXIF and chroma subsampling, each with why it matters in practice.
Image terminology, in plain language, without the hand-waving.
Lossless
Smaller with nothing discarded.
PNG and GIF compress by describing repetition rather than by dropping detail. Decoding gives back the exact original pixels, which is why lossless files of photographs are large.
Alpha
The transparency channel.
A fourth value per pixel saying how opaque it is. PNG, WebP and AVIF have it; JPEG and BMP do not. Converting to a format without alpha replaces transparent areas with a background color, which is why File4Edit asks you to pick one.
Aspect ratio
Width divided by height.
Changing it without cropping means stretching, which looks wrong in a way people notice even when they cannot say why. "Cover" crops to fit; "contain" pads to fit; "stretch" distorts.
Megapixel
A million pixels.
Width times height. A 4000×3000 image is 12 megapixels. It matters because memory and processing time scale with pixel count, not with file size — a 2 MB PNG can be far heavier to work with than a 5 MB JPEG.
DPI / PPI
Only meaningful when printing.
Dots or pixels per inch is a note about intended physical size. On a screen it does nothing: a 1000px image is 1000px whether it is tagged 72 or 300 DPI. Changing DPI without resampling changes no pixels at all.
Compression artifact
Damage introduced by lossy compression.
Blocky squares, smeared edges and colored fringes around hard lines. They appear where the compressor had to choose what to discard, which is why text and line art suffer far more than photographs.
Chroma subsampling
Storing color at lower resolution than brightness.
The eye resolves brightness far better than color, so JPEG usually stores color at half resolution. Almost invisible on photographs; very visible on red text and sharp color edges.
Quantization
Reducing to a limited set of values.
Turning millions of colors into 256 for a GIF, for instance. Done naively it produces visible banding; dithering hides the banding by scattering the error.
Dithering
Trading resolution for apparent color depth.
Mixing available colors in a pattern so the eye blends them into one that is not in the palette. It is why a 4-color image can still read as a photograph.
Raster vs vector
Pixels versus instructions.
A raster image is a grid of pixels and blurs when enlarged. A vector image is a set of shapes and stays sharp at any size. SVG is vector; everything else here is raster. Converting vector to raster is one-way.
Bit depth
How many values each channel can take.
8-bit gives 256 levels per channel, which is 16.7 million colors. Lower depths band visibly in gradients. Higher depths matter for editing headroom, not for final display.
sRGB
The color space the web assumes.
A shared definition of what "red" means, so a color looks the same on different screens. An image tagged with a wider space and displayed as sRGB looks flat and desaturated.
EXIF
Metadata the camera embeds.
Camera model, exposure, orientation, timestamp, and often GPS coordinates. It travels with the file, survives copying, and is not visible in the picture — which is exactly why it catches people out.
Unsharp mask
Sharpening by exaggerating edges.
The image is blurred, and the difference between original and blur is added back. Named for a darkroom technique using an out-of-focus negative. Too much produces bright halos along every edge.
Data URI
A file encoded directly into text.
Base64 of the bytes, inline in HTML or CSS, so no separate request is made. It is about 33% larger than the file and cannot be cached on its own, so it only pays off for very small images.
How to use Image terms glossary
- Type the term you are looking for The filter box above the list matches against the term, its one-line definition and its full explanation, so you can search by a word inside an entry as well as by its name.
- Read the one-line definition Every entry opens with a single sentence — lossy is "throws information away to get smaller," alpha is "the transparency channel," DPI is "only meaningful when printing."
- Read what it means in practice Below the definition, each entry explains the consequence: why re-saving a JPEG degrades it every time, or why converting to a format without alpha needs a background color.
- Clear the filter to browse With the box empty the whole list shows again. If nothing matches what you typed, the page says so rather than leaving you with a blank list.
Image terms glossary: common questions
What is the difference between lossy and lossless compression?
Lossy compression throws information away to make a file smaller; lossless compression makes it smaller without discarding anything. JPEG, WebP and AVIF are lossy: they discard the detail the eye is least likely to miss. That is why repeatedly editing and re-saving a JPEG degrades it every time. PNG and GIF are lossless: decoding gives back the exact original pixels, which is also why a lossless photograph is a large file.
Does DPI matter for images on the web?
No. DPI or PPI is a note about intended physical size and does nothing on a screen: a 1000-pixel image is 1000 pixels whether it is tagged 72 or 300 DPI. Changing the DPI value without resampling changes no pixels at all. It only becomes meaningful when the image is printed.
What is the alpha channel?
Alpha is a fourth value stored per pixel saying how opaque that pixel is, and it is what makes transparency possible. PNG, WebP and AVIF have it; JPEG and BMP do not. Converting an image to a format without alpha replaces the transparent areas with a solid background color, which is why File4Edit asks you to choose one.