Image Size vs Quality: What Actually Shrinks a File
"Make this image smaller" hides three completely different operations, and picking the wrong one is why people end up with a blurry photo that is still too big to attach. There are only three levers that change an image file's size: its dimensions, its compression setting, and its format. This guide explains what each one does, which to reach for first, and how to hit a hard limit without visible damage.
Order of operations: reduce the dimensions first if the image is larger than it needs to be, then change the format, then lower the quality only as far as you need. Doing it in that order costs the least visible quality per byte saved.
Lever one: dimensions
Dimensions are width and height in pixels, and they are the most powerful lever by a wide margin, because pixel count grows with the square of the scale. Halving both the width and the height leaves you with one quarter of the pixels. A phone photo that comes off the camera at 4000 pixels wide has around 12 million pixels; the same photo at 1000 pixels wide has under a million. Nothing you do with a quality slider comes close to that kind of reduction.
The question is whether you need those pixels. Extra resolution is only useful if something will display or print it. A photo shown 800 pixels wide in a web page gains nothing from being 4000 pixels wide except load time. A photo destined for a large print gains a great deal. The practical test is: what is the largest size at which anyone will ever look at this image? Match the dimensions to that, allowing roughly double for high-density screens, and stop.
This converter does not resize. It re-encodes your image at its original pixel dimensions and changes only the format and compression, which keeps its behavior predictable. If you need different dimensions, do that first in your phone's photo editor, your operating system's built-in image tools, or a desktop editor, then convert.
Lever two: compression quality
For lossy formats, the quality setting decides how much fine detail the encoder is allowed to discard. In this converter that is the slider, which runs from 10 to 100 percent and defaults to 90. It applies to JPG, WebP, and AVIF. It disappears for PNG and ICO, because those are lossless and have nothing to trade.
The relationship between the slider and file size is not a straight line. Going from 100 to 90 usually cuts the file dramatically for almost no visible change, because the top of the range spends enormous numbers of bytes on detail nobody perceives. Going from 90 to 80 saves less and starts to be visible on close inspection. Below roughly 60, you are trading obvious quality for diminishing returns, and the artifacts described in our JPG vs PNG guide start to show up in skies, shadows, and around text.
A useful habit: change the slider, convert, and read the resulting size shown on the file row. Two or three passes will find the setting that meets your limit, and you will see exactly what each step costs.
Lever three: format
Choosing a different format can be nearly free, in the sense that it changes the size without changing the dimensions or your perception of quality. Moving a photograph from PNG to JPG typically produces a fraction of the original size. Moving a JPG to WebP typically saves another 25 to 35 percent at comparable quality. Moving to AVIF usually saves more still, at the cost of slower encoding.
Format choice can also work against you, which is why files sometimes grow. Converting a photograph to PNG makes it bigger, often several times bigger, because PNG has to record every pixel exactly. Converting an efficient HEIC from an iPhone to JPG usually makes it bigger too. Neither is a bug; both are the arithmetic of lossless or less-efficient storage.
What does not shrink a file
- Renaming the extension. Changing .png to .jpg relabels the file and does not re-encode anything. Some software will then refuse to open it.
- Changing DPI or PPI. That number is metadata describing intended print size. It does not change how many pixels exist, so it does not change the file size.
- Cropping a little. Trimming 5 percent off an edge removes 5 percent of the pixels. Useful for composition, negligible for size.
- Zipping a JPG. Already-compressed data has almost nothing left for a general-purpose compressor to find.
- Stripping metadata. It helps by kilobytes, not megabytes. Worth doing for privacy reasons, as covered in our metadata guide, but not as a size strategy.
Why two identical-looking photos differ in size
Content drives compression. A photograph of a plain wall compresses to a tiny file because there is almost nothing to describe. The same camera pointed at a tree in leaf produces a much larger file at the identical quality setting, because fine random detail is exactly what lossy compression cannot throw away without it showing. This is also why screenshots of interfaces compress so well as PNG and why photos of foliage, gravel, hair, and rain are stubbornly large in every format.
Hitting a specific limit
Upload forms often specify a maximum, commonly a few megabytes, occasionally as little as a few hundred kilobytes for a profile photo. A reliable sequence:
- Check the dimensions. If the image is far larger than the display will ever be, resize it first. This alone often solves the problem.
- Convert to JPG if the image is a photograph and it is currently PNG or HEIC.
- Convert at 90% quality and check the resulting size on the file row.
- Step down to 80, then 70, converting again each time, until you are under the limit.
- Look at the result at the size it will actually be seen. If it holds up, you are done.
If you have gone below 60 percent quality and you are still over the limit, the dimensions are the real problem and no amount of slider work will fix it.
Practical targets
| Use | Sensible width | Format and quality |
|---|---|---|
| Email attachment | 1600 to 2000 px | JPG at 80 to 85% |
| Web page body image | 1200 to 1600 px | WebP at 80% |
| Full-width hero image | 1920 to 2560 px | WebP or AVIF at 75 to 80% |
| Profile photo | 400 to 800 px | JPG at 85% |
| Screenshot of text | Native size | PNG (lossless) |
| Logo with transparency | As needed | PNG or lossless WebP |
| Photo for printing | Largest you have | JPG at 95% or PNG |
Treat these as starting points rather than rules. The right answer always depends on the picture and where it is going, which is why the converter shows you the before and after size on every file it processes.
Common questions
What is the difference between image size and file size?
Image size usually means dimensions, the width and height in pixels. File size is how many bytes the saved file occupies. They are related but not the same: two images with identical dimensions can differ enormously in file size depending on format, compression setting, and how much detail the picture contains.
Does lowering the quality slider always make the file smaller?
For lossy formats like JPG, WebP, and AVIF, yes, lowering it produces a smaller file. For PNG and ICO the slider does not apply at all and is hidden, because those formats are lossless and have no quality to trade away.
What quality setting should I use?
Start at the default of 90 percent, which is visually clean for almost all photographs. If you need a smaller file, step down to about 80, then 70, checking the result at the size it will actually be viewed. Below roughly 60 percent, artifacts become easy to see in skies, shadows, and around text.
Can this converter resize my image?
No. It re-encodes at the original pixel dimensions and changes only the format and compression. Resizing is a separate job handled by your phone's photo editor, your operating system's image tools, or a desktop editor. Converting after resizing gives you both savings at once.
Why did my file get bigger after converting?
You almost certainly converted to a lossless format. A photograph saved as PNG is often several times larger than the same photograph as JPG or HEIC, because PNG stores every pixel exactly instead of simplifying the picture. That growth is expected and is the price of lossless storage.
Find your setting: drop an image into the converter, try 90% and 80%, and compare the sizes shown on each row. Two passes usually settle it.