PNG to JPG converter
Your images never leave this device
This is the conversion that actually saves space. A photograph stored as PNG keeps every pixel exactly and gets enormous; as a JPG it is routinely a tenth of the size with nothing visible lost. Two things come with it and neither is reversible: JPG discards detail permanently, and it has no transparency, so anything see-through is laid onto white before the file is written.
How to convert PNG to JPG
What JPEG is good at, and what it mangles
JPEG works in eight-by-eight blocks. Each block is turned into a set of frequency coefficients, the high-frequency ones are divided down and mostly rounded away, and what survives is what gets stored. That is an excellent model for a photograph, where brightness changes gradually and the eye does not miss the finest variations. It is a poor model for a hard edge, because an edge is made of nothing but high frequencies, and discarding them leaves a faint ripple on either side. On a screenshot of text or a logo with a crisp outline you can see this at full zoom: a grey shimmer that was not in the PNG.
Colour is treated worse than brightness, deliberately. Most JPEG encoders store the two colour channels at half resolution horizontally and vertically, on the sound reasoning that human vision is far more sensitive to detail in luminance than in colour. It is a good bet for a photograph and a bad one for red text on a white background, which is about the worst input you can hand a JPEG encoder.
The white behind your transparent pixels
JPEG has no alpha channel, so a transparent PNG needs something painted behind it, and what this tool paints is white. That is a decision rather than a default: a canvas with transparent pixels encodes them as black if nothing is drawn first, which is how a transparent logo ends up in a black box on other converters. If white is the wrong background for your image, the format cannot help you and the file should stay a PNG.
Everything except the pixels is dropped
The conversion decodes the PNG into a canvas and encodes the canvas as JPEG. A canvas holds pixels and nothing else, so the output carries no EXIF, no XMP, no colour profile and no text chunks. On a screenshot that saves a few kilobytes and loses nothing. On a photograph it means the camera model, the capture date and any GPS coordinates are gone, which is often the reason people convert in the first place. It also means an image tagged Display P3 arrives as untagged sRGB: the browser converts the pixels into the canvas colour space on the way in, so saturated colour outside the sRGB gamut is compressed down to what sRGB can express.
What people use it for
- Getting a screenshot under an email attachment limit
- Meeting the file-size cap on a marketplace listing or an application form
- Cutting a folder of exported photographs down to something sendable
- Stripping camera and location data out of a picture before sharing it
- Making a phone screenshot small enough to send in a message
- Deciding, per file, whether a picture is better off as a PNG after all
Questions
JPEG compresses in eight-by-eight blocks and handles hard edges badly. For logos, line art and text, PNG is the right format and converting away from it is a step backwards.
It is replaced with white, painted before the image is drawn. JPG has no alpha channel, so a transparent PNG needs something behind it.
For a photograph, typically five to fifteen times. For a flat-colour graphic, sometimes nothing at all.
Flat colour with few distinct tones is PNG’s best case and JPEG’s worst. The saving figure shows a dash when that happens, which is the cue to keep the original.
85 covers almost everything. Below about 70 the block structure starts showing in skies and flat colour; above 95 you pay a lot of bytes for a difference nobody can see.
You can convert the file back, but the detail JPG discarded is gone for good. Keep the original PNG if you might need it.
No, and neither does anything else that is not a pixel. GPS coordinates included.
It is converted into sRGB while being drawn and saved with no profile attached. Colours outside the sRGB gamut are compressed to fit, which is visible mainly in very saturated reds and greens.
Yes. The shorter extension is a leftover from DOS filename limits.
No. The conversion happens in this page using the browser’s own image pipeline.