← All guides

How to Resize an Image (Pixels, File Size, or Percentage)

imagesresizedimensions

“Resize” means three different things and people conflate them constantly. Knowing which one you actually need saves time and prevents quality damage.

The three things:

  1. Resize by pixels — make the image have specific width/height (e.g., 1920×1080)
  2. Resize by percentage — make the image bigger or smaller proportionally (e.g., 50% of original)
  3. Resize by file size — make the file smaller in MB/KB (which is actually compression, not resizing)

Most “resize” tools handle the first two. The third needs a separate tool. Here’s when to use each.

The fastest way to resize by dimensions or percentage

Use the Image Resizer. Drop in your image, pick dimensions or a percentage, click resize.

The flow:

  1. Open the Image Resizer
  2. Drag in your image
  3. Choose your method:
    • Specific dimensions: enter target width and/or height in pixels
    • Percentage: enter a scale (50% halves; 200% doubles)
    • Fit within bounds: enter a maximum width/height; image is scaled to fit
  4. Choose whether to maintain aspect ratio (almost always yes)
  5. Click resize
  6. Download the resized image

The conversion runs in your browser. No upload.

To reduce file size (compression, not resize)

Use the Image Compressor instead.

Resizing makes the image smaller dimensionally. Compression keeps the same dimensions but reduces file size by re-encoding with less detail. These are separate operations.

For most “my image is too big” situations, you want a combination: resize the dimensions down if they’re huge (over 2000 px on the longest side is rarely needed for screen viewing), then compress.

Aspect ratio: why you almost always want “maintain aspect ratio” on

Aspect ratio is the relationship between width and height. A 1920×1080 image is 16:9 (1920 ÷ 1080 = 1.78). If you resize to 1000×500, you’ve changed the aspect ratio to 2:1 — and the image will look stretched/squished.

Always keep “maintain aspect ratio” turned on unless you specifically want to distort. Visual content (photos, design work, screenshots) looks wrong when proportions change.

If you need a specific aspect ratio that doesn’t match your input, use the Image Cropper — it cuts off parts of the image rather than stretching, which is what you actually want.

Common target dimensions

What you should resize to depends on where the image is going:

For email or sharing:

  • 1600 px on the longest side is plenty for typical viewing
  • 800 px on the longest side for tiny thumbnails

For social media (always check current specs — these change):

  • Instagram feed (square): 1080 × 1080
  • Instagram feed (portrait): 1080 × 1350
  • Instagram Stories: 1080 × 1920
  • Twitter/X feed: 1200 × 675 or 1080 × 1080
  • LinkedIn feed: 1200 × 627
  • Facebook feed: 1200 × 630

For print:

  • 300 DPI is the rule. For a 4×6 inch print, that’s 1200 × 1800 pixels.
  • Higher print sizes scale up: 8×10 at 300 DPI = 2400 × 3000.

For web pages on your site:

  • Hero images: 1920 × 1080 maximum (or 2× for retina = 3840 × 2160 but file sizes get huge)
  • Inline images: usually 800-1200 px wide
  • Thumbnails: 400 × 400

For “I just need it smaller” (most common reason):

  • 1600 px on the longest side
  • Quality 0.85 compression
  • That’s it. Done.

Upscaling: when you can and can’t

The Image Resizer can scale images larger (1.5×, 2×, etc.). The result is mathematically interpolated — the algorithm guesses new pixel values from neighboring pixels.

This works OK for:

  • Modest upscaling (1.5× to 2×)
  • Simple graphics with flat colors (logos, icons)
  • Already-large source images that just need slight scaling

This doesn’t work for:

  • Aggressive upscaling (4× or more) — the result looks blurry / mushy
  • Recovering detail from compressed JPGs — what’s gone is gone
  • Making low-res scans usable for high-quality print — fundamentally limited

For real upscaling — taking a small image and producing a high-quality larger one — you need AI-based tools (Topaz Gigapixel, Upscale.media, ESRGAN) that use machine learning to invent plausible new detail. The basic interpolation in browser resizers can’t do this.

For most use cases, don’t upscale — find a higher-resolution source if you need a larger image.

Common pitfalls

Resizing PNG instead of converting to JPG. PNGs of photos are 3-5× larger than equivalent JPGs. If you’ve got a 20 MB PNG that “needs to be smaller,” the right move is convert to JPG with PNG to JPG, not resize. Resizing a PNG just gives you a smaller PNG that’s still inefficiently formatted.

Resizing already-compressed JPGs. Each JPG save degrades quality slightly (it’s a lossy format). If you resize a JPG, you’re saving it again, picking up another generation of compression artifacts. To preserve quality, resize once from the highest-quality source you have, save the result, work from that.

Resizing screenshots to fit social media. Screenshots often have crisp edges and text. Aggressive resize/compression makes text fuzzy. For screenshots specifically, either resize less aggressively or save as PNG.

Resizing for a print order without knowing DPI. Print quality depends on pixels-per-inch. A 1000×1000 image prints fine at 3×3 inches (~333 DPI) but looks pixelated at 10×10 inches (100 DPI). Always know the print size in inches plus the printer’s DPI requirement.

Resizing without cropping first. If your final destination is 1080×1080 (Instagram square), and your source is 4032×3024 (iPhone photo, 4:3), resizing maintains the 4:3 ratio — you get 1080×810 or similar. To actually get 1080×1080, crop first to square, then resize.

Resize vs crop vs compress

Quick decision tree:

  • Wrong dimensions? → Resize (Image Resizer)
  • Wrong shape (aspect ratio)? → Crop (Image Cropper)
  • File too big but dimensions are fine? → Compress (Image Compressor)
  • Wrong shape AND wrong size? → Crop, then resize
  • Wrong everything? → Crop, then resize, then compress (the full workflow)

The full “prepare an image for sharing” workflow

For a typical “make this iPhone photo emailable” task:

  1. (Optional) HEIC to JPG if the source is HEIC
  2. (Optional) Image Cropper if you want a specific aspect ratio
  3. Image Resizer to 1600 px on the longest side
  4. Image Compressor at quality 0.85
  5. (Optional) EXIF Stripper if there’s sensitive location data

Total time: under a minute per photo, sometimes seconds with batch input.

Final result: a JPG that’s typically 200-400 KB instead of 4-5 MB, looks visually identical to the source on any screen, and is comfortable to email or upload anywhere.

Privacy

The Image Resizer (and Cropper, Compressor) all run in your browser:

  • Image read with the File API
  • Resized in a canvas element via JavaScript
  • Output is a browser blob you download

Nothing uploads. For private images (personal photos, work mockups, screenshots of sensitive UIs), this matters.

TL;DR

  • Resize by dimensions or percentageImage Resizer
  • Reduce file sizeImage Compressor — that’s compression, not resize
  • Change aspect ratioImage Cropper — cropping, not resize
  • Common target: 1600 px longest side for general “make it smaller” use
  • Always keep aspect ratio on unless you specifically want to distort
  • Don’t upscale heavily — basic interpolation can’t recover real detail