You know forwarded images get blurry, but not why. Does every forward lose quality? Why do some images degrade fast while others are slow? Why do blurred images show blocky artifacts rather than uniform blur?
This article explains JPEG compression science in plain language — from pixels to frequencies, quantization to artifacts — so you fully understand why and how images degrade.
01 The JPEG Compression Pipeline
JPEG splits images into 8×8 pixel blocks, performing DCT (Discrete Cosine Transform) on each — converting spatial pixel brightness into frequency domain coefficients. Simply put: transforming "what color is at what position" into "what brightness variation frequencies exist in this area."
Then the critical step: quantization. Each frequency coefficient is divided by a corresponding value in a quantization matrix, then rounded. Larger quantization values discard more information — smaller files, less detail. Low-frequency data (large color blocks, gentle gradients) is preserved; high-frequency data (edges, textures, sharp transitions) is discarded first.
Quantized data undergoes Huffman coding for further lossless compression, producing the final JPEG file. In the entire process, only quantization is lossy — and this "lossiness" is the root of JPEG artifacts and quality degradation.
02 Why Does Each Forward Lose Quality?
When a JPEG is saved, it undergoes a complete "decode → modify/don't → re-encode" cycle. Even if you change nothing — just open and re-save — the quantization step rounds again, accumulating new rounding errors.
First save loses ~1% information; second save loses ~1% from the already-degraded version; third save loses from the twice-degraded version — errors snowball. This is "generation loss."
This explains why repeatedly forwarded images show increasingly prominent block artifacts — each 8×8 block, after multiple quantization rounds, has internal color differences gradually flattened while inter-block boundaries become increasingly jarring. The result is that distinctive "mosaic + blur" hybrid effect.
Fun math fact: a 10-megapixel JPEG at quality 75 discards ~40% of DCT coefficient information during quantization. By the 10th re-save, cumulative loss can exceed 90% of original information.
03 Which Images Degrade Fastest? Slowest?
Fastest degradation: detail-rich, high-contrast images with fine lines and text (screenshots, design mockups). These contain abundant high-frequency information, losing much with each quantization. Text-heavy memes are especially prone to "patina."
Slowest degradation: uniformly colored, softly gradiated, no-sharp-edge images (sky photos, solid backgrounds). These are predominantly low-frequency, losing little during quantization.
Social platforms accelerate degradation further — WeChat, Weibo, etc. compress on upload (typically quality 60–75), adding an extra generation loss beyond manual forwarding. This is why group chat images degrade so quickly.
FAQ
Do PNG images degrade when forwarded repeatedly?
PNG is lossless — theoretically no quality loss from re-saving. But social platforms convert PNG to JPEG on upload, so PNGs forwarded through social media still degrade — the culprit is the platform's transcoding, not PNG itself.
Is JPEG quality 100 lossless?
No. Even at quality 100, quantization still occurs — the quantization matrix divisors just become smaller (~1), but DCT coefficients are still rounded. Each save still has infinitesimal information loss, just imperceptible to human eyes.
Can JPEG degradation be prevented?
Best prevention: don't use JPEG — save as PNG or WebP (lossless mode). If JPEG is necessary, avoid re-saving: use lossless formats for intermediate versions, only export to JPEG as the final step.
Is WebP more resistant to re-saving degradation than JPEG?
WebP's lossy mode is similar to JPEG — re-saving also degrades quality. But WebP generally achieves better quality at equivalent compression ratios (higher SSIM), so WebP degrades slightly less than JPEG over the same number of re-saves.
Why do social platforms compress images?
To save storage and bandwidth. A platform with 100M daily users may receive billions of images daily — without compression, server costs and loading speeds would be unsustainable. Compression is a tradeoff between user experience and operational costs.
Try the Tool Now
This article explains JPEG compression science in plain language — from pixels to frequencies, quantization to artifacts — so you fully understand why and how images degrade.