Image Compression, Resizing and Format Choices

Learn when to resize, crop, compress or convert PNG, JPEG and WebP, with realistic quality and transparency checks.

Four operations that are often confused

Resize changes pixel dimensions. Crop removes part of the scene. Convert changes the file's encoding format. Compress aims to reduce encoded file size. Sometimes one action also affects the others, but no single setting is right for every image.

Use Image Resizer to set dimensions, Image Cropper to select the content, Image Format Converter to change an encoding and Image Compressor to test an export trade-off.

Pixel size and file size are different

An image measuring 4000 × 3000 pixels contains 12 million pixels. A 1200 × 900 copy contains about 1.08 million pixels, but you cannot determine its final byte count from pixel dimensions alone. Encoding settings, visual detail, transparency and metadata all affect size.

Example: a photograph for an article may be unnecessarily large at 4000 pixels wide if it will appear at 900 pixels. Resizing first generally reduces the data the encoder needs to store. For a logo with thin lettering, resizing too far makes the text hard to read even if the file becomes tiny.

Choose the format by the content

JPEG is common for photographs but uses lossy encoding and does not preserve alpha transparency. PNG supports transparency and crisp graphics, though photographs can be relatively large. WebP can encode photos or graphics and may preserve alpha, but compatibility should be checked against the destination workflow. The best choice is the one that meets your destination's format rules and keeps the necessary detail.

Transparent backgrounds and white fills

A transparent PNG logo can appear to have a white background when previewed on a white page. Before converting to JPEG, view it on a contrasting background: JPEG cannot preserve the transparent region. If a white background is acceptable, explicitly choose it; otherwise export as PNG or a supported transparent WebP.

Practical example: preparing a product image

1. Keep the original file unchanged.
2. Crop out unwanted surrounding space while keeping the product completely visible.
3. Resize to the actual target dimensions, for example 1200 × 1200 if the platform asks for a square image.
4. Try WebP or JPEG for a photograph and compare it with a PNG only if the destination supports both.
5. Inspect fabric textures, hair, fine text and edge halos at 100% zoom. Measure the real output file size rather than guessing from the quality slider.

EXIF rotation and color management

Some source photos contain orientation metadata or embedded color profiles. A browser canvas export may normalize orientation and may not keep every metadata field. Review the exported image in the actual app where it will be used. If accurate color reproduction is critical, use a color-managed professional workflow and inspect the final export on a calibrated display.

When the image contains private content

Working in the browser can avoid uploading the original to a conversion server, but it does not prevent someone with access to your computer, browser or downloaded files from seeing it. Remove sensitive metadata or hidden material using a dedicated inspection tool when necessary. See our privacy disclosures.

Frequently asked questions

Does setting 90% quality always make a file smaller?

No. The original codec, resolution and visual detail matter. Compare the actual output bytes.

Can I restore detail after shrinking an image?

Resizing down discards detail. Enlarging the smaller file does not recreate the original information.

Which format should a transparent logo use?

PNG is a reliable choice when transparency and sharp edges matter; WebP is another option when supported.