The short answer
To make an image smaller without changing its dimensions, compress the file instead of resizing it. Compression changes how the same pixels are stored. Resizing changes the pixel grid itself.
That distinction matters when a form says something like:
- "Image must be 600 x 600 pixels and under 200KB."
- "Upload a 2 x 2 inch passport photo, JPG only, max 240KB."
- "Profile photo must be at least 1000 x 1000 pixels, max 1MB."
- "Do not change the image dimensions."
In those cases, the width and height are part of the requirement. If you resize the image, you may solve the KB problem but fail the dimension rule. Use a compression tool first.
File size and dimensions are different
Image "size" is confusing because people use the same word for two different things.
| What people say | What it means | Example | How to change it |
|---|---|---|---|
| File size | Bytes on disk | 85KB, 240KB, 2.8MB | Compress, remove metadata, convert format |
| Pixel dimensions | Width and height | 600 x 600 px, 1200 x 800 px | Resize or crop |
| Print size | Physical output size | 2 x 2 inches, 4 x 6 inches | Depends on pixels and DPI/PPI metadata |
A 2000 x 2000 image can be 10MB or 300KB. The dimensions are identical. The storage method is different.
The reverse is also true: a 500 x 500 PNG screenshot can be larger than a 2000 x 2000 JPEG photo if the screenshot is saved inefficiently or includes transparency and metadata.
When you should not resize
Use compression first when the current dimensions are already correct:
- Government or school forms: the form may validate both pixel dimensions and file size.
- Passport or ID photos: the crop may already be correct, and resizing could break the face-size or head-position requirement.
- Profile photos: the platform may ask for a square image above a minimum pixel size.
- Website images already sized for a layout: the CSS slot may expect the same aspect ratio and pixel grid.
- Screenshots for documentation: changing dimensions can make UI text softer or harder to compare.
- Images used by another system: a template, database, print workflow, or batch script may expect exact dimensions.
Resize only if the destination allows smaller dimensions or if the image has far more pixels than needed. web.dev's image guidance separates these concerns: correctly sized images reduce unnecessary bytes, while compression tunes the file that remains.
What compression changes
When dimensions stay the same, file size can still change because the exported file changes in one or more of these ways:
- Quality level: lossy formats store fewer visual details as quality goes down.
- Format: WebP or AVIF may store the same pixel dimensions more efficiently than JPEG or PNG.
- Compression mode: lossless preserves exact pixels; lossy changes pixels slightly to save more bytes.
- Color information: some encoders reduce color precision or chroma detail.
- Metadata: camera model, GPS, preview thumbnails, DPI tags, and editing history may be removed.
MDN's canvas export documentation is a useful mental model for browser tools: the browser can export the same canvas bitmap into different file formats and quality levels. If the canvas is 1200 x 800 pixels, the exported JPEG or WebP can stay 1200 x 800 while the byte size changes.
Pick the method by image type
| Starting image | Keep dimensions? | Try first | Avoid |
|---|---|---|---|
| JPG photo over an upload limit | Yes | Lower JPEG quality in small steps | Re-saving the same JPG repeatedly |
| PNG photo with no transparency | Yes | Convert to JPG or WebP if accepted | Keeping it as PNG just because it started as PNG |
| PNG screenshot with text | Yes | PNG optimization or lossless WebP | Low-quality JPEG |
| Transparent logo or cutout | Yes | PNG or WebP, depending on destination support | JPEG, because it removes transparency |
| Website hero already sized correctly | Yes | WebP or AVIF with fallback | Very low quality that causes banding |
| Email attachment | Usually | JPEG for photos, PNG for graphics | WebP or AVIF if the recipient may not support them |
| Archive or master file | Yes | Keep original, optionally add a compressed copy | Replacing the only master with a lossy export |
This is where many generic compression posts go wrong. "Convert to WebP" is good advice for a website image you control. It can be bad advice for a court form, school form, print shop, or email attachment that only accepts JPG or PNG.
Quality settings that usually work
There is no universal quality number. web.dev recommends testing different compression levels until you find the best compromise between visual quality and file size. Use these as starting points, not laws:
| Image job | JPEG/WebP starting point | What to inspect |
|---|---|---|
| Passport, ID, profile, or document photo | 82-90 | Face detail, text labels, edge sharpness |
| Product photo | 80-88 | Fabric, labels, corners, color shifts |
| Blog or website photo | 75-85 | Skin, sky, shadows, gradients |
| Thumbnail or preview | 65-78 | Subject clarity at small display size |
| Background image | 55-72 | Banding and obvious blocks |
| Screenshot with text | Lossless PNG/WebP first | Menu text, icons, thin lines |
For AVIF, many tools use a different scale, so do not compare the number directly against JPEG or WebP. A "quality 60" AVIF is not the same as a "quality 60" JPEG.
The safest workflow is boring but reliable: export, compare, adjust, repeat.
Format conversion without resizing
Switching formats can reduce file size without changing pixel dimensions. The important question is whether the destination accepts the new format.
JPEG to WebP: often useful for website delivery. Google's WebP documentation reports strong average savings in its test sets, and WebP supports lossy, lossless, transparency, and animation. Still, a strict upload form may reject WebP even if browsers support it.
PNG photo to JPEG or WebP: often a large win if the PNG is a camera photo and does not need transparency. PNG is lossless, so it preserves every tiny camera-noise variation. JPEG or WebP can store the same dimensions with fewer bytes by using lossy photo compression.
PNG screenshot to JPEG: often a bad trade. The file may get smaller, but text and hard edges can look dirty. Try PNG optimization or lossless WebP first.
Transparent PNG to JPEG: only do this if you are comfortable flattening the background. JPEG cannot keep transparent pixels.
JPEG to JPEG again: works, but do it from the best source you have. Recompressing an already compressed JPEG multiple times can compound artifacts.
How to compress without changing dimensions in PhotoTools
Use this workflow when the width and height must stay the same:
- Open Compress, not Resize.
- Drop in the image.
- Check the input dimensions shown by the tool or your system's file info.
- Keep the same dimensions and choose an output format accepted by the destination.
- Lower the quality slider gradually for JPEG, WebP, or AVIF.
- Download the result.
- Re-check the downloaded file's pixel width and height before uploading it.
- If privacy matters, check whether EXIF/GPS metadata remains.
PhotoTools runs the compression workflow in your browser. The image is processed locally during the tool session rather than uploaded to a server for conversion.
Do not use the resize tool if the destination requires the current pixel dimensions. Resize is for changing width and height. Compress is for changing bytes.
Examples by real upload problem
| Problem | Better move | Why |
|---|---|---|
| 600 x 600 JPG must be under 200KB | Keep JPG, lower quality from high 80s toward low 80s | Dimensions and accepted format stay safe |
| 1200 x 1200 profile photo is 2MB, limit is 1MB | Compress JPG or convert to WebP if accepted | Plenty of room without changing pixels |
| PNG screenshot is 900KB, dimensions must stay | Try PNG optimization or lossless WebP | Text remains crisp |
| Transparent product cutout is too large | Try WebP with alpha if accepted | Keeps transparency with fewer bytes in many web cases |
| Website image is 4000 px wide but displays at 800 px | Resize first if allowed, then compress | Keeping 4000 px wastes bytes |
| Form says exact 413 x 531 px JPG, max 300KB | Do not resize; compress JPG only | The validator may reject any dimension change |
The hard cases are usually strict forms. If a form names exact pixels and a file-size cap, trust those requirements over generic image advice.
When resizing is also necessary
Sometimes "without resizing" is the wrong constraint. If the image has far more pixels than the destination needs, compression alone may force you into ugly quality settings.
A 4000-pixel-wide photo crushed to very low JPEG quality may still be heavy and will look worse. If the destination only displays the image at 1000 px wide and allows that size, resize first and then compress moderately. That usually looks better than keeping unnecessary pixels and destroying detail.
Use this rule:
- If the dimensions are mandatory, compress first.
- If the dimensions are only larger than needed, resize first.
- If you are not sure, check the upload instructions before changing dimensions.
Troubleshooting
The image is still too large
Try a more efficient accepted format, lower quality a little more, remove metadata, or start from the original file instead of a previously exported copy. If the file is still far over the limit and the destination allows smaller dimensions, resize first.
The upload form rejects the file
Check the allowed file type. Many forms accept JPG and PNG only. A WebP file can have the right dimensions and file size but still fail if the form does not list WebP as an accepted format.
The image looks blurry
The dimensions may be unchanged, but compression quality is too low or the wrong format was chosen. Use a higher quality setting. For screenshots, diagrams, and text, use PNG or lossless WebP instead of low-quality JPEG.
The transparent background turned white or black
The image was probably converted to JPEG. JPEG has no transparency. Export as PNG, WebP, or AVIF if the destination supports transparency.
The dimensions changed anyway
You probably used a resize, crop, or export preset that changed the pixel grid. Reopen the original and use Compress only. Check the output dimensions before submitting the file.
A final rule that saves quality
Keep the original. Compression is safest when you can export again from the best source file.
If you repeatedly download a compressed image, compress it again, convert it to another lossy format, and compress it again, artifacts build up. For one upload, that may be acceptable. For product photos, ID photos, or website assets you will reuse, keep a clean master and create delivery copies from it.