Cropping changes what an image shows. Resizing changes how many pixels represent it. Knowing which operation you need helps you avoid cut-off subjects, soft exports, and unnecessarily large files.
Quick answer
Crop to remove unwanted edges or change composition. Resize to change pixel dimensions while keeping the chosen view. A crop can preserve the pixels inside its boundary when it uses an aligned, one-to-one pixel mapping. Resizing through resampling calculates a new pixel grid. Export is a separate step: JPG can add compression loss even when you have not intentionally resized anything.
Crop vs resize: what actually changes?
Cropping selects a smaller field of view
Imagine a 4000 × 3000 photograph with a person in the middle. Selecting a 3000 × 3000 square removes the sides. At native crop dimensions, the person does not need more or fewer pixels to represent their face. You simply discard content outside the selection.
The resulting image has fewer total pixels than the original, but that does not automatically make its retained detail worse. It does mean you have less room for another crop later. Keep the original rather than treating the cropped download as a replacement master.
Resizing changes the pixel grid
Resize the entire photograph to 1000 × 750 and you keep the same composition in fewer pixels. Downsampling combines information from the source. Enlarge it to 8000 × 6000 and interpolation supplies additional pixels, not newly captured detail. Either operation changes the sampling grid.
Changing only a displayed width in a webpage is different again. The browser can display a large file at a small size without rewriting that file. For a practical explanation of display-only framing, see how to crop an image in CSS.
Aspect ratio is not an output size
A ratio describes proportions, not resolution. A 1:1 square might be 300 × 300 or 3000 × 3000. A 16:9 crop might be 1280 × 720 or 2560 × 1440. Choosing a ratio tells the editor what shape to select. Choosing exact dimensions additionally tells it how large to render that selection.
Which operation should you use?
| Your goal | Operation | Main trade-off |
|---|---|---|
| Remove distracting edges | Crop at native selection size | Discarded content is gone from the output |
| Keep the whole photo but reduce dimensions | Resize proportionally | Fine detail is reduced |
| Fill a square avatar | Crop, then resize if required | Some composition is removed |
| Make a small photo larger | Upsample | More pixels do not restore missing detail |
For example, a landscape photograph cannot become a filled square while simultaneously keeping every edge and preserving proportions. You must crop, add padding elsewhere, or accept distortion. These crop tools select a region rather than adding borders.
How to crop with or without resizing
Use Crop Image to Size when a destination specifies pixel dimensions. The following controls are available in the current editor, rather than hypothetical settings found in other software.
- Upload the original. Choose Upload image or drop a file into the upload area. The tool accepts JPG, PNG, WebP, and AVIF inputs, provided your browser can decode them.
- Choose a Crop preset. Pick the proportions you need, then drag the crop selection and adjust its boundary around your subject. Presets start with an inset selection, so inspect the framing instead of assuming the whole image is selected.
- Decide whether dimensions must be exact. Leave Exact output size off for dimensions derived from the selected source area. Turn it on to enter Width (px) and Height (px). Editing either dimension enables exact output and updates the rectangular crop ratio.
- Check the output dimensions. Compare the selected source area with the requested output. If the selection contains roughly 900 × 600 source pixels, requesting 1800 × 1200 enlarges it. A small preview cannot prove that enlargement will look sharp.
- Choose Export format and download. Rectangular outputs offer PNG or JPG. Download once, then inspect the saved image at its actual pixel size before uploading it elsewhere.
For a profile picture, Crop Image Square provides square framing. The same distinction applies: the square selection determines composition, while exact output size determines whether the tool also resamples it.
Why export format still matters
The crop selection and the file encoder solve different problems. PNG uses lossless compression, preserving the raster given to its encoder without JPEG-style compression loss. Ordinary JPG export uses lossy compression. A high setting can look excellent while still changing image data.
WebP supports both lossy and lossless encoding, so the extension alone does not identify which was used. These tools can accept WebP input but export only PNG and JPG. There is no WebP export option here, even if the browser itself supports one.
Choosing PNG does not undo resampling, repair an already compressed photograph, or recover removed edges. It avoids an additional lossy encoding stage for the newly rendered canvas. A JPG converted to PNG still contains the visible consequences of its earlier JPEG compression.
Nor does cropping guarantee a smaller download. A small crop exported as PNG can exceed the size of a larger, efficiently compressed JPG. Dimensions, texture, noise, transparency, metadata handling, and encoding all influence bytes. Check the saved file rather than predicting its size from the crop percentage.
Full resolution, smoothing, and browser limits
The editor calculates its source rectangle from the loaded image’s natural dimensions, not the on-screen preview dimensions. With exact output disabled, it rounds the selected width and height to whole pixels. With exact output enabled, it draws the selection into the requested dimensions.
The preview is limited to a longest side of 320 pixels, but download rendering does not use that preview scale. A small preview therefore does not mean the download is limited to 320 pixels. “Full resolution” here means using the loaded source selection, not retaining the original image’s full dimensions after removing its edges.
The renderer enables smoothing and requests high smoothing quality. Browser support and implementation vary. Fractional crop coordinates, rounding, color handling, and the canvas rendering pipeline mean this is not a guaranteed bit-for-bit pixel-copy workflow. It is also not a JPEG bitstream lossless-cropping utility.
Exact-size exports are guarded at 16,384 pixels per side and 40 million pixels total. Those are application limits, not a promise that every allowed canvas will work. Browsers and devices have their own dimension and memory limits. Larger downloads can fail below the application ceiling.
Common problems
The crop looks softer after download
Check whether an exact-size preset enlarged a small selection. Then compare PNG and JPG exports at 100% viewing scale. Viewer scaling can soften either file, and a website may resize or recompress a perfectly adequate upload.
The subject is cut off in a thumbnail
A destination ratio may need a different composition from your original. Move the selection before downloading. For wide video artwork, use YouTube Thumbnail Crop and keep important faces and text away from vulnerable edges.
The output is too large or will not export
Reduce exact dimensions and try again. If bytes are the problem rather than pixel dimensions, JPG may suit an opaque photograph better than PNG. This editor does not provide a quality slider or a target-file-size guarantee.
Frequently asked questions
Does cropping reduce image quality?
Not inherently within retained pixels. An aligned crop at native dimensions can preserve their detail, but this canvas workflow may resample fractional selections and re-encodes the result. JPG export adds a separate lossy stage.
Is resizing the same as compressing?
No. Resizing changes pixel dimensions through resampling. Compression encodes those pixels into a file. Lossless compression preserves its input raster; lossy compression trades some information for a different encoding result.
Can I crop without changing the aspect ratio?
Yes. Choose a preset matching the original proportions and adjust the selection. You still remove content and reduce native crop dimensions, even though the output retains the same width-to-height relationship.
Will increasing dimensions make my crop sharper?
No. Ordinary interpolation adds pixels without recovering missing source detail. Begin with a larger original or retain a larger selection when the destination requires a larger output.
Should I crop or resize first?
Usually choose the final composition from the original first, then render once at the destination dimensions. This avoids processing pixels you will discard and reduces unnecessary repeated resampling. Keep an untouched original for future versions.
For a preservation-focused workflow, read how to crop an image without losing quality.