Treating the JPEG as the master file and re-saving it after every edit, which compounds generation loss.
File formats
JPEG
Also called jpg, .jpg or .jpeg. Here is what it means, when it changes what you export, and what people get wrong.
File formatsRuns on your deviceUpdated 10 August 2026
The short answer
JPEG is the standard lossy format for photographs. It shrinks files by discarding detail your eye is least likely to notice, which makes it small and accepted almost everywhere, but it has no transparency, no layers, and loses a little quality every time you re-encode it.
Updated
The short version
JPEG at a glance
| Quick fact | Detail |
|---|---|
| Standard | ITU-T T.81, published 1992 |
| Compression | Lossy, 8 x 8 block DCT with quantisation |
| Transparency | None |
| Bit depth | 8 bits per channel in the baseline profile |
| Colour | YCbCr, usually with chroma subsampling |
| Maximum size | 65,535 x 65,535 pixels |
Slide the table sideways to see every column.
How does JPEG actually throw data away?
A JPEG encoder does four things in sequence. It converts the image from RGB into YCbCr, separating brightness from colour. It usually halves the resolution of the two colour channels, a step called chroma subsampling, because human vision is far more sensitive to brightness than to hue. It splits the image into 8 x 8 pixel blocks and runs a discrete cosine transform on each, turning pixel values into frequency coefficients. Then it divides those coefficients by a quantisation table and rounds the results.
That rounding step is where the loss happens, and it is deliberate. High-frequency coefficients, the ones describing fine texture, are divided by large numbers and mostly round to zero. Whatever survives is packed with Huffman coding, which is lossless. Everything discarded in quantisation is gone permanently.
What does the quality slider really control?
The 0 to 100 scale is an encoder convention, not part of the JPEG standard, so quality 80 in one application does not necessarily match quality 80 in another. What the number does is scale the quantisation table: a lower setting multiplies the divisors, rounds harder, and keeps less.
The curve is steeply non-linear. Dropping from 100 to about 90 often removes half the file size with no visible difference. Somewhere around 75 to 85 is where most photographic images sit comfortably. Below about 60 you start seeing the 8 x 8 grid appear as blockiness, ringing along sharp edges, and muddy colour in saturated areas. Quality 100 is not lossless; it is simply a very gentle quantisation table applied to an already subsampled image, and the file can be several times larger for no perceptible gain. If you are trimming files for a marketplace, a compressor that lets you watch the result (Photomop's compressor does the encoding in your browser) beats guessing at a number.
Why does re-saving a JPEG degrade it?
Every decode and re-encode cycle runs quantisation again on data that has already been quantised, and the rounding errors accumulate. This is generation loss. Open a JPEG, make a small edit, save, and repeat, and after a dozen rounds the blocking and colour smearing become obvious.
Two things make it worse. Cropping by a number of pixels that is not a multiple of eight shifts the block grid, so every block gets re-coded from scratch. Changing the quality setting between saves means the second encoder cannot reuse the first one's decisions. The fix is to keep a lossless master (PNG, TIFF or the original camera file), edit that, and export a fresh JPEG each time rather than editing the JPEG itself. Pure rotations, flips and grid-aligned crops can be done losslessly by tools that manipulate the coefficients directly instead of re-encoding.
What can JPEG not do?
The gaps are worth knowing before you commit a workflow to it.
- No transparency. There is no alpha channel, so a cut-out saved as JPEG arrives with its background filled in, usually white or black.
- No layers. The file is a single flattened raster.
- No animation. A JPEG is one still image.
- Eight bits per channel in the baseline profile, so no HDR and no 16-bit editing headroom.
Colour management and metadata ride along in application segments rather than in the core codec: JFIF (ITU-T T.871) defines the common container conventions, EXIF carries camera data and orientation, and an ICC profile can be embedded to describe the colour space. CMYK JPEGs exist in prepress via an Adobe marker, but they are a niche and many web tools mishandle them.
What people get wrong about jpeg
Each one is a real failure mode, not a style preference.
Exporting at quality 100 in the belief that it is lossless, producing a much larger file that is still lossy.
Saving a cut-out product image as JPEG and being surprised the transparent background turns solid.
Assuming quality 80 means the same thing in two different applications, when the scale is encoder-specific.
Questions people ask
JPEG, answered
The follow-up questions people search for once they have the definition.
Is JPG the same as JPEG?
Yes, identical. The three-letter extension survives from older Windows and DOS systems that allowed only three characters after the dot. Both extensions describe the same format, and renaming a file between them changes nothing about its contents.
How many times can you re-save a JPEG before it looks bad?
It depends on the quality setting and whether the image is re-cropped, but visible damage usually appears within five to fifteen cycles at moderate quality. Keeping a lossless master and exporting fresh copies avoids the problem entirely, however many versions you produce.
Can a JPEG have a transparent background?
No. JPEG has no alpha channel, so transparency is filled with a solid colour when you export. If you need a cut-out, save as PNG or WebP. If the destination will not accept transparency, flatten deliberately onto the background colour you want rather than letting the exporter choose.
Why do my JPEGs look blocky in flat areas?
That is the 8 x 8 transform grid becoming visible because quantisation was too aggressive. Smooth gradients such as skies and studio backdrops show it first. Raising the quality setting, or exporting from the original rather than from an already compressed copy, usually clears it.
Is JPEG still the safest format for marketplace listings?
Yes, for broad compatibility. Every major selling platform accepts JPEG, file sizes are predictable, and no reviewer will reject it on format grounds. Newer formats compress better, but JPEG remains the option least likely to cause an upload problem.
Same cluster
More from the glossary
Neighbouring entries, so every term in the set is one hop from every other.
Layer
A layer is one image stacked on top of others in an editing document, like a sheet of glass in a pile.
Lifestyle shot
A lifestyle shot shows a product in real use and real surroundings rather than isolated on a plain background.
Interpolation
Interpolation is how software works out pixel values at positions that did not exist before, which is what happens whenever you resize, rotate or warp an image.
From the studio
Tools that do this
Each one runs on your own device, so the photo stays where it is.
Sources
Where the facts on this page come from. Every link opens the specification, standard or documentation the claim was read at.
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.