8 min read · updated August 2, 2026
Crop Image Online: Ratios, Exact Pixels, and Quality
Image Cropper
Crop images with custom dimensions — free, no signup
To crop an image, you select a rectangular region of the pixel grid and throw away everything outside it. That is the whole operation — no scaling, no stretching, no interpolation. It is the correct fix whenever the problem is framing: a distracting edge, a subject that is off to one side, or a platform that demands a 1:1 square when your camera shot 4:3.
The catch is that cropping interacts with two things people rarely think about until it is too late: aspect ratio requirements and file re-encoding. Upload a 4:3 photo where a site expects 1:1 and the site will crop it for you, usually badly. Crop a JPEG carelessly and you can lose a generation of quality you did not need to lose.
This guide covers what actually happens to the pixels when you crop, the specific dimensions the major platforms and common real-world specs expect, and how to decide between cropping and resizing when either could technically get you to the target size.
Cropping and resizing solve different problems
The two operations get conflated constantly, and mixing them up is the single most common reason an image comes out distorted. Resizing changes the dimensions of the whole image by resampling every pixel — the full frame survives, but every pixel is recalculated. Cropping changes the dimensions by discarding pixels — the surviving region is untouched, but the frame gets smaller.
A concrete example: you have a 4000 by 3000 photo and a site wants 1080 by 1080. Resizing alone cannot get you there without squashing the image, because 4:3 and 1:1 are different shapes. The correct workflow is crop first, then resize: crop a 3000 by 3000 square out of the original, then scale that square down to 1080. Crop decides what is in the frame; resize decides how many pixels describe it.
The decision rule is simple. If the shape of the image needs to change, or you want something out of the frame, crop. If the shape is already right and only the pixel count is wrong, resize. If both are wrong, crop to the target ratio first and resize second — doing it in the other order means you resample pixels you are about to throw away.
What actually happens to the pixels when you crop
A digital image is a grid of pixels, each storing color values. A crop is defined by four numbers: the x and y offset of the top-left corner of the crop box, plus its width and height. The software copies exactly that region into a new image and discards the rest. The pixels inside the box are not resampled, blurred, or recalculated in any way, which is why cropping is fundamentally a quality-preserving operation.
The quality risk comes from what happens next: saving the result. If the source is a lossless format like PNG, the cropped file is pixel-identical to the corresponding region of the original. If the source is a JPEG, the story changes, because JPEG is a lossy format that compresses the image in 8 by 8 pixel blocks. When you crop a JPEG and save it as a JPEG again, the file is decompressed, cropped, and recompressed — a second lossy pass. At quality 90 or above the difference is invisible in practice; at aggressive quality settings, or after several crop-save cycles, you get visible generation loss, especially around sharp edges and text.
Two practical takeaways follow. First, always crop from the original file, not from a copy that has already been re-saved two or three times. Second, if you plan further edits, export the crop as PNG and only convert to JPEG as the final step. One re-encode at the end is harmless; five along the way are not.
The dimensions real platforms and specs actually want
Most cropping jobs are not creative decisions — they are compliance jobs. Some target somewhere published a spec, and your image has to match it. These are the numbers that come up most often in practice.
Two details matter beyond the raw numbers. Aspect ratio is what platforms enforce; exact pixel counts are usually minimums, so cropping to the right ratio at a larger size and letting the platform downscale is fine. And for print or ID photos, the spec is physical: a 2 by 2 inch photo at 300 DPI means you need a 600 by 600 pixel square with the head sized per the requirement — this is where cropping to exact pixels, not just a ratio, earns its keep.
- Profile pictures almost everywhere (LinkedIn, GitHub, X, Slack): 1:1 square, displayed in a circle, so keep the face centered away from the corners
- Instagram feed: 1:1 at 1080 by 1080, or 4:5 portrait at 1080 by 1350 for more screen height
- Stories, Reels, TikTok: 9:16 at 1080 by 1920
- YouTube thumbnails: 16:9 at 1280 by 720, minimum width 640
- Open Graph link previews: roughly 1.91:1, commonly 1200 by 630
- US passport and many visa photos: 2 by 2 inches at 300 DPI, which is exactly 600 by 600 pixels
- Standard photo prints: 4 by 6 inches is 3:2, 5 by 7 is 7:5, 8 by 10 is 5:4 — three different shapes, which is why lab prints crop your photos when you do not
Composition: deciding where the crop box goes
Once the ratio is fixed, the remaining decision is placement, and a few working rules cover most cases. For portraits, crop so the eyes sit roughly a third of the way down the frame, and never crop exactly at a joint — mid-forearm or mid-shin reads fine, at the wrist or ankle reads amputated. Leave a little space on the side the subject is facing.
For square profile crops, remember that nearly every platform masks the square into a circle. Anything in the corners disappears, so a face that fills the frame edge to edge will lose its chin and hair. Crop slightly looser than feels natural and check that the head fits inside an imaginary inscribed circle.
For thumbnails and social cards, crop tighter than instinct suggests. A 1280 by 720 thumbnail is often viewed at under 200 pixels wide on a phone, and a wide shot with a small subject turns into noise at that size. If the subject does not read clearly when you squint, the crop is too loose. Also keep critical text and faces away from the outer 10 percent of the frame — several platforms overlay timestamps, duration badges, and UI controls on the edges.
Five cropping jobs and how to handle each
The same tool settings apply differently depending on the job. These are the scenarios that show up over and over.
- Job application headshot: crop to 1:1 from the shoulders up, eyes in the top third, then resize to around 800 by 800 — large enough for any site, small enough to upload fast
- Visa or passport photo: crop to the exact pixel spec (600 by 600 for US applications), with the head occupying the required fraction of frame height; get the ratio wrong and the application portal rejects the file outright
- Marketplace product photos: crop every listing to the same square ratio with the product occupying a consistent share of the frame, so your storefront grid looks deliberate instead of random
- Screenshot for documentation or a bug report: crop to just the relevant window or panel, removing your bookmarks bar, desktop, and notification icons — smaller file, no accidental leaks of private tabs
- Photo going to a print lab: crop to the print ratio yourself (3:2 for 4 by 6) before ordering, because the lab machine will otherwise center-crop automatically and it does not know where the faces are
Common questions
Image Cropper FAQs
- Does cropping an image reduce its quality?
- Cropping itself does not reduce quality, because the pixels you keep are copied unchanged rather than resampled. Quality loss only enters when the cropped result is re-saved in a lossy format like JPEG, which adds one round of recompression. Saving at high quality, or exporting to PNG, keeps that loss invisible.
- What is the difference between cropping and resizing an image?
- Cropping removes pixels outside a selected region and changes what is in the frame; resizing resamples all pixels to change the image dimensions while keeping the full frame. Use crop to change the shape or remove content, resize to change the pixel count. To hit an exact target like 1080 by 1080, crop to the ratio first and then resize.
- How do I crop an image to exact pixel dimensions?
- Use a cropper that accepts numeric width and height instead of only drag handles. Enter the target dimensions, position the fixed-size box over the region you want, and export. This matters for specs defined in pixels, like a 600 by 600 passport photo or an ad slot with a fixed creative size.
- How do I crop a photo to a square without stretching it?
- Lock the crop tool to a 1:1 aspect ratio before adjusting the box, and the selection can only ever be square. Stretching happens when software resizes a rectangle into a square instead of cropping it. A ratio-locked crop discards the excess width or height, so faces and objects keep their true proportions.
- Can I crop a JPEG without losing any quality?
- For practical purposes, yes: crop the original file once and save at quality 90 or higher, and the recompression loss is not visible. Perfectly lossless JPEG cropping is technically possible only when the crop aligns with the format's 8-pixel block grid, which specialist tools handle. For everyday use, avoiding repeated crop-and-save cycles matters far more.
- What size should I crop a profile picture to?
- Crop to a 1:1 square around 800 by 800 pixels, which exceeds the display size of every major platform. Keep the face centered and within an imaginary circle inside the square, since most sites mask profile photos into circles and cut off the corners. One well-made square works across LinkedIn, GitHub, X, and Slack without re-cropping.
Cropping is the rare image edit where the mechanics are simple and the judgment is everything. Get three things right — crop from the original, lock the aspect ratio before you drag, and crop before you resize — and the technical side takes care of itself. What remains is the eye: where the frame edge falls relative to eyes, joints, products, and the circle masks that platforms will apply after you upload.
The ToolDoor Image Cropper handles the mechanical half: aspect ratio presets for the common shapes, numeric fields for exact-pixel offsets and dimensions, and a preview of the result before you download. It is free, requires no signup, and exports your crop without watermarks.
Nearby doors