Crop vs Resize Video: Which One Do You Need?
By Mark Fulton · 2026-09-08 · 10 min read
Cropping cuts pixels out of the frame to change its shape; resizing keeps every pixel and changes the scale of the whole frame. If a vertical platform is rejecting your landscape footage, you need to crop it to a taller ratio. If a file is too big or too high-resolution for where it's going but the shape is already right, you need to resize it. Using resize when you meant crop gives you a smaller version of the wrong shape. Using crop when you meant resize throws away parts of the picture you wanted to keep.
Those two sentences solve most of the confusion, but the reason people mix them up is that both tools live under "change my video's dimensions" in every editor's menu, and both produce a video with different width and height numbers than the one you started with. The difference is what happens to the picture in between.
What does cropping actually remove?
Cropping draws a smaller rectangle inside your existing frame and throws away everything outside it. Nothing inside the rectangle is touched, scaled, or resampled. If your source is a 1920x1080 landscape clip and you crop it to 9:16 for a vertical feed, the tool keeps a 608-pixel-wide vertical strip out of the middle of that 1920-pixel-wide frame and discards the roughly 656 pixels of picture on each side. Whatever was standing at the left or right edge of the original shot is gone. It isn't blurred or squeezed into the new frame, it's simply not in the output file.
This is exactly what FFmpeg's own crop filter documentation describes: the filter crops (i.e. removes pixels from) the input video into a specified rectangle, without touching the pixels that remain. VidClip's crop tool works the same way. Pick 9:16, 4:5, 1:1, or 16:9, and it computes the largest rectangle of that shape that fits inside your source frame, centers it on both axes, and cuts to it. It rounds the box down to the nearest even width and height on both sides first, because H.264 encoding with standard chroma subsampling requires even dimensions to encode at all. An odd-pixel crop would simply fail to render.
The consequence worth remembering before you shoot anything: a subject dead center in a 16:9 frame survives a crop to 9:16 with room to spare. Two people standing at opposite edges of a wide shot do not. No amount of software recovers pixels that were never captured.
What does resizing preserve?
Resizing takes the entire original frame and redraws it at a different resolution, without removing any part of the picture. If you resize a 3840x2160 4K clip down to a 1080-tall output, you still see everything that was in the original shot, top to bottom and left to right; there are just fewer pixels representing it. Nothing is cut off the edges. This is the operation FFmpeg's scale filter performs, and its own documentation is explicit that when the aspect ratio should be preserved the filter recomputes one dimension from the other rather than stretching the frame to fit an arbitrary box, using the force_original_aspect_ratio behavior described in the scale filter reference.
VidClip's resize tool works on that principle. You choose a target height (480p, 720p, or 1080p), and the width is calculated automatically from your source's existing aspect ratio, so a 16:9 clip resized to 720p comes out 1280x720 and a 9:16 vertical clip resized to "1080p" comes out 1080 pixels tall by 608 pixels wide, not 1080x1920. That number surprises people who assume "1080p" always means the familiar 1920x1080 landscape frame; it doesn't. 1080p describes the height, and the width follows whatever shape your source already is.
One more thing resizing does not do: it only scales down. There's no version of "resize" that invents detail a camera never captured, so a resize tool that claims to make a small clip bigger is just spreading the same pixels across a larger canvas, not sharpening anything.
Why does my video look stretched?
Stretching happens when a video gets forced into a new aspect ratio without cropping or padding to compensate, so both dimensions get scaled by different amounts. A 1920x1080 clip squeezed into a 1080x1080 square by simply setting width and height independently comes out with faces a little wider and shorter than they should be, because the horizontal scale factor (1080/1920) and the vertical scale factor (1080/1080) are different numbers, and every pixel in the frame gets pushed by whichever factor applies to its axis.
Genuine resize tools, including VidClip's, don't do this: they lock the two axes together by deriving width from height (or vice versa) using the source's own ratio, so nothing gets distorted, and it's an aspect-ratio-preserving scale down to a specific pixel height, not a fit-to-box stretch. If your output looks stretched, one of two things happened: an editor let you type in a width and a height independently and you picked numbers that didn't match your source ratio, or a "fit to frame" feature filled a target box by stretching rather than cropping or padding. The fix in both cases is the same. Decide what your final ratio needs to be, then crop to that ratio (accepting some edge loss) or pad to it (accepting bars), and let the height or width follow proportionally rather than typing in both numbers by hand.
When do you need both operations, and in which order?
Plenty of real jobs need a ratio change and a resolution change at once, for example a 4K 16:9 screen recording that needs to become a 720p-scale 9:16 clip for Reels. Crop first, resize second. Crop while you still have the most source pixels to work with, so the tool has room to choose a clean rectangle and you can see exactly what survives. Resizing an already-cropped clip down to a smaller height afterward is just a straightforward downscale of a frame whose shape is already correct.
Doing it the other way around, resizing before cropping, doesn't break anything technically, but it throws away resolution you might have wanted for the crop step, and on a small resized frame it's harder to judge whether the crop box is landing on the right part of the picture. If your footage already has the right shape and only needs to be smaller, skip cropping entirely and go straight to resize. If it has the right resolution and only needs to be reshaped, skip resizing and go straight to crop.
There's a third operation worth naming here even though VidClip doesn't offer it: padding, sometimes called letterboxing. Padding scales the source down to fit inside a target box on its longer axis, then fills the leftover space on the shorter axis with solid color bars instead of cropping it away. It's how a widescreen movie shows up on an old square TV without cutting off the sides. VidClip's resize tool won't do this for you: if your clip's ratio doesn't already match the box you're publishing into, resizing alone gets you a smaller version of the original shape, not a padded fit to a new one. If you actually need bars, your practical options here are to accept a crop instead (see above) or to reshoot or re-lay-out the source at the target ratio in the first place. The comparison below lays out what each of the three does to the picture, side by side.
Crop vs resize vs pad
| Crop | Resize | Pad (letterbox) | |
|---|---|---|---|
| What happens to the picture | Outer pixels are cut away and discarded | Every pixel is kept, redrawn at a new scale | Every pixel is kept, plus solid color bars added |
| What happens to the ratio | Changes to whatever shape you pick | Stays exactly the same as the source | Changes to the target box; source sits inside it |
| What happens to file size | Usually shrinks (fewer total pixels to encode) | Shrinks in proportion to the resolution drop | Usually grows slightly (bars add encodable area) |
| Symptom that means you need this one | Platform wants a different shape than your footage | File or resolution is too big but the shape is already right | You want the full original frame visible with no cropping and don't mind bars |
| Available on VidClip | Yes, /crop-video (9:16, 4:5, 1:1, 16:9, centered) | Yes, /resize-video (480p/720p/1080p, downscale only) | No, not offered |
If your symptom is "the platform is rejecting my shape" or "important parts are getting squeezed," you need crop. If your symptom is "the file is too big" or "the resolution is overkill for where this is going" and the shape is already correct, you need resize. If your symptom is "I don't want to lose any of the frame and I'm fine with bars," you need padding, which means looking outside VidClip for that specific step, or accepting a crop instead.
Does either one reduce file size?
Cropping usually reduces file size because it reduces the total pixel count the encoder has to store per frame. Cutting a 1920x1080 frame down to a 608x1080 vertical crop removes roughly two-thirds of the pixels, so there's proportionally less data per frame even before any compression setting changes. The amount you actually save depends on how much the footage in the cropped-out area was contributing (a busy, detailed background compresses worse than a plain wall, so cropping it away saves more).
Resizing reduces file size more predictably, because the relationship between resolution and pixel count is a known multiplier: halving both width and height cuts total pixels to a quarter, and file size tends to fall with it before you've touched a single compression setting. That math and the resolution-to-file-size relationship are covered in more depth in our breakdown of video compression by the numbers, which is worth reading if the goal here is actually hitting a target file size rather than a target shape.
Padding is the outlier: it usually makes files slightly larger, not smaller, because the color bars it adds are extra pixels the encoder has to store, even though they're flat and cheap to compress compared to real footage.
None of this is a substitute for compression itself. Cropping and resizing both reduce the raw pixel count going into the encoder, which helps, but the actual bitrate and codec settings still do most of the work on final file size.
Try it without uploading anything
Both tools run the same way: drop a file into /crop-video or /resize-video, and the video is read into your browser's own memory and processed there with FFmpeg compiled to WebAssembly. Nothing goes to a server, there's no queue, and no account is required. Run the same clip through both tools and compare the two outputs side by side. It's the fastest way to see, on your own footage, exactly what each operation keeps and what it throws away, and if you're unsure which settings to reach for on a specific export, the recipe advisor will suggest one.
Frequently asked questions
Does cropping a video reduce its size?
Usually, yes. Removing pixels from the frame reduces the amount of picture data the encoder has to store per frame, so the output is typically smaller than the source at the same compression settings. How much smaller depends on how much visual detail was in the part you cropped away; cropping out a plain, static background saves less than cropping out a busy, moving one, because the busy area was costing more bits to encode in the first place.
How do I crop without losing quality?
Crop from the highest-resolution version of the file you have, not from a copy that's already been resized down or heavily compressed. The pixels inside the crop rectangle are untouched, so quality loss at the crop step itself comes only from re-encoding afterward. Keep your compression settings reasonable on export and the cropped region will look as sharp as the same region did in the source.
What's the difference between crop and trim?
Crop changes the shape of each frame in space, cutting pixels off the edges. Trim changes the length of the video in time, cutting whole frames off the beginning or end while leaving every remaining frame's picture untouched. They solve different problems and are frequently both needed on the same clip. If frame-accurate trimming without a re-encode is what you're after, see our guide to lossless trimming and keyframes.
Can I crop and resize in one pass?
Not in a single step on VidClip; crop and resize are two separate tools because they do two separate things to the frame. Run crop first while you still have the most source resolution to work with, download that result, then feed it into resize if you also need a smaller file. Doing the reshape before the resize means you're always judging the crop box against full-resolution footage rather than an already-shrunk copy.