Why JPG compresses so well
JPG was designed around how human vision works. The eye notices changes in brightness far more readily than changes in colour, so the format stores brightness carefully and colour roughly. It also breaks the image into small blocks and keeps only the coarse patterns within each one.
This is why a photograph can lose ninety percent of its file size and still look unchanged, while a screenshot of text treated the same way develops visible fringes around the letters. The format is tuned for continuous tone images, not for sharp edges.
Choosing a target for a JPG
If a form dictates a size, use that number. If the choice is yours, think about where the image will be seen. A photo on a web page rarely needs more than 150KB. An email attachment is comfortable at 200KB. A photo you may want to crop later deserves more room.
The preview slider is the honest test. Drag it over the most detailed part of the image, usually a face or a patch of text, and decide whether the difference matters for the way the file will be used.
What to watch for
- Banding in skies. Smooth gradients are where heavy compression shows first. Raise the target if the sky turns into visible bands.
- Halos around text. If your JPG contains lettering, a low target creates a faint outline around each character.
- Already small files. A 60KB JPG pushed to 50KB gains almost nothing and loses real detail.
- Transparency. JPG does not support it. An image with a transparent background gets a white one instead.
Frequently asked questions
Is there a difference between JPG and JPEG?
None at all. They are the same format with two spellings, left over from an era when file extensions were limited to three letters. A tool that handles one handles the other.
How does JPG compression actually work?
It converts the image into frequency information and discards the fine detail the human eye is least sensitive to, particularly in colour. The quality setting controls how much is discarded, which is why the same photo can be stored at wildly different sizes.
Can I compress a JPG without losing anything?
Not through quality reduction, which always discards data. Some detail is removed every time. The practical goal is to remove only what you cannot see, which is what a well chosen target achieves.
Why does compressing the same file twice make it worse?
The second pass works on the already reduced version and removes detail again, compounding the artefacts. Always return to the original file rather than recompressing an output.
Does the output keep my camera information?
No. The image is redrawn onto a canvas before saving, so EXIF data such as camera model, date and GPS location is not carried into the new file.
Should I use JPG or WebP for my website?
WebP produces smaller files at similar quality and is supported by every current browser. JPG remains the safer choice for forms, email attachments and anything that may be opened by older software.