Compress Image

About Image Compression

Image compression is the process of reducing file size while maintaining acceptable image quality. This tool uses advanced algorithms to compress images efficiently without significant quality loss.

What is the difference between lossless and lossy compression?

Lossless compression rearranges and deduplicates data so the original file can be reconstructed bit-for-bit when decompressed — PNG, WebP-lossless, FLIF, and ZIP all use this approach, typically shrinking images by 20-50%. Lossy compression permanently discards data the human eye is least likely to notice (high-frequency detail, subtle color shifts) and trades quality for far smaller files — JPEG, WebP-lossy, AVIF, and HEIC can reduce file size by 80-95% with no visible change at typical viewing distances. Use lossless for screenshots, line art, logos, anything with sharp text, and any image you will edit further. Use lossy for photographs and any final-output image where bandwidth matters. Once an image has been lossy-compressed, the discarded data is gone forever — always keep a lossless master.

Will compressing an image multiple times degrade its quality?

Yes, but only for lossy formats. Each time you re-encode a JPEG or lossy WebP, the encoder applies another round of quantization on top of the already-quantized data, accumulating artifacts: blocky 8x8 patches, color bleeding around edges, and a soft "crunchy" texture in smooth areas. This is called generation loss, and it is why you should never repeatedly save and edit a JPEG through a workflow. Lossless formats (PNG, WebP-lossless) can be re-encoded indefinitely without any quality change because no information is being thrown away. Best practice: keep the original in a lossless master (PSD, TIFF, PNG) and only export to JPEG/WebP/AVIF for final delivery. If you must re-edit a JPEG, do all your changes in one session and save once.

Which compression quality setting should I use for JPEG?

Quality 85-92 is the sweet spot for almost all photographic content — most viewers cannot distinguish it from quality 100, but file size is 40-60% smaller. Below 75, blocking artifacts become visible on smooth gradients (skies, skin); below 60, compression noise is obvious. Above 95, file size grows rapidly with almost no perceptual gain. Different encoders disagree on quality scales: Photoshop's "12" is roughly equal to standard quality 95, while Google's MozJPEG at default 75 is comparable to standard 85 thanks to better trellis quantization. For web hero images use 80-85; for thumbnails 70-75 is fine; for printable archive use 95+ or move to a lossless format. Always preview side-by-side at 100% zoom before committing.

How do I keep transparency when compressing PNG images?

Transparency is the alpha channel — a per-pixel value from 0 (fully transparent) to 255 (fully opaque). PNG, WebP, AVIF, GIF (binary only), and HEIC all support alpha; JPEG does not. If you compress a PNG-with-transparency to JPEG, the transparent areas will be filled with a solid background color (usually white). To preserve transparency, stay in PNG or move to WebP/AVIF which give you alpha plus much better compression. Quantizing a PNG-24 (16M colors) to PNG-8 (256-color palette) with tools like pngquant typically shrinks the file by 60-80% while keeping perfect alpha — modern pngquant uses dithered alpha to preserve smooth edges. For maximum savings on transparent assets, use WebP-lossless or AVIF, both of which support full 8-bit alpha at roughly half the size of PNG.

How do AVIF, WebP, and JPEG XL compare for compression efficiency?

All three are modern successors to JPEG, but with different tradeoffs. AVIF (based on the AV1 video codec) gives the smallest files — typically 50% smaller than JPEG at the same perceptual quality — but is slow to encode and has limited support in older software. WebP (Google, 2010) is well-supported in all modern browsers and gives roughly 25-35% savings over JPEG, with fast encoding and both lossy and lossless modes. JPEG XL (2021) is the newest standard, supports progressive decoding, lossless JPEG transcoding (saving 20% on existing JPEGs without quality loss), and excellent quality, but browser support is incomplete as of 2026. For maximum reach today, serve WebP with a JPEG fallback via the picture element. For new projects targeting modern browsers, AVIF gives the best size; JPEG XL is the format to watch for archival use.

What does "progressive JPEG" mean and should I use it?

A baseline JPEG stores image data top-to-bottom in a single pass, so the browser reveals the picture line by line as bytes arrive. A progressive JPEG splits the data into multiple scans of increasing detail — first a blurry full-image preview, then progressively sharper passes — so the user sees the whole picture immediately, even at low quality, and watches it sharpen. For images over about 10 KB, progressive JPEGs are typically 2-10% smaller than baseline AND feel faster to load on slow networks, which is why Google PageSpeed and most CDNs recommend them. The encoding cost is slightly higher and decoding requires holding the whole image in memory before display, so for very small thumbnails or memory-constrained embedded devices, baseline can win. Modern encoders like MozJPEG default to progressive.

How does color quantization in PNG-8 affect image quality?

PNG-8 uses an indexed color palette of at most 256 colors, chosen by an algorithm that picks the colors most representative of the image. For graphics with a small palette (logos, icons, UI screenshots), 256 colors is plenty and the file becomes 50-80% smaller than PNG-24. For photographs, 256 colors causes visible banding in skies and skin tones; dithering (Floyd-Steinberg, Atkinson, or modern blue-noise dither) trades a grainy texture for the appearance of more colors by alternating palette entries in checkerboard patterns. Tools like pngquant and ImageMagick let you control palette size (16, 32, 64, 128, 256) and dither strength. For photographs, prefer JPEG, WebP, or AVIF instead; PNG-8 is for graphics.

Why does my compressed image look worse on some devices than others?

Several factors stack here. First, color profile handling: if the image has an embedded Adobe RGB or P3 profile and the viewer's app or browser ignores it, colors render incorrectly. Always tag with sRGB for web. Second, device pixel ratio: a 400-pixel-wide image scaled to 800 CSS pixels on a 2x Retina display looks soft, even at quality 100, because the browser is upscaling. Always serve images at 1x, 2x, and 3x densities using srcset. Third, OLED versus LCD displays render dark grays and pure blacks differently, exposing JPEG blocking in shadow areas that looks fine on LCD. Fourth, ambient light: a low-quality image looks worse outdoors with reduced screen contrast. Test your compression on a phone in sunlight, not only on your color-calibrated desktop monitor.

Is this compressor private, and does it upload my images anywhere?

No upload happens. This tool runs entirely client-side in your browser using the HTML5 Canvas API — your image is read into memory, decoded, re-encoded, and offered back as a download without a single byte leaving your device. That makes it safe for confidential screenshots, client work, medical or legal documents, and anything under NDA. The trade-off versus server tools is that quality depends on your browser's built-in JPEG/WebP encoder rather than a tuned library like MozJPEG or pngquant, and very large images are auto-resized to a 4096px maximum to stay within browser memory limits. Because the canvas redraws pixels, all EXIF, GPS, and camera metadata is stripped automatically as a side effect — handy for privacy, but keep an original if you need that data.

Why does compressing a PNG sometimes make the file bigger, and which format should I download for the web?

PNG is a lossless format, so the quality slider does nothing to a PNG — there is no perceptual data to discard. When you re-encode a PNG through a canvas, the browser's encoder is often less efficient than the tool that originally created the file (it cannot do palette quantization like pngquant), so the result can actually be larger than the source. This tool detects that case, keeps your original bytes, and reports 0% rather than bloating the file. For the web, the practical move is to convert: this tool encodes WebP and JPEG from the same canvas at your chosen quality and flags the smaller one as Recommended. As a rule, choose WebP for photos and graphics with transparency (typically 25-35% smaller than JPEG), JPEG for broad compatibility on photographic content, and keep PNG only for sharp-edged graphics, logos, or screenshots where lossless matters.