概述
本指南比较了JPG和PNG在最重要的标准上的表现，以帮助您选择适合您需求的格式。
JPG最适合照片；PNG最适合具有透明度或文本的图形。
Head-to-Head Comparison
文件大小
JPG: 由于有损压缩，JPG在照片中的文件大小要小得多。
PNG: PNG是无损的，文件较大，但对于具有透明度的图形是必要的。
Winner: JPG
质量
JPG: 由于有损压缩，JPG每次保存时都会失去质量。
PNG: PNG是无损的, , 保存时没有质量损失。
Winner: PNG
透明度
JPG: JPG不支持透明度。
PNG: PNG支持完整的alpha通道透明度。
Winner: PNG
最佳用途
JPG: 最适合照片和具有数百万种颜色的复杂图像。
PNG: 最适合徽标、图标、截图和具有锐利边缘的图像。
Winner: 平局
网页使用
JPG: JPG加载速度更快，适合照片 - 有利于页面速度。
PNG: PNG文件较大，但在需要透明度时是必要的。
Winner: JPG
Technical Foundations
JPG (JPEG) is a lossy, block-based image format. Its encoder divides the image into 8x8 pixel blocks, applies a discrete cosine transform (DCT) to convert each block from spatial pixels into frequency coefficients, and then quantizes those coefficients, discarding the high-frequency detail to which human vision is least sensitive.[1] The surviving coefficients are compressed with entropy coding. PNG (Portable Network Graphics) is a lossless format that stores every pixel exactly. It applies per-row prediction filters and then the DEFLATE algorithm, a combination of LZ77 dictionary matching and Huffman coding, so the original image is reconstructed bit-for-bit on decompression.[3][4]
Color and Transparency
JPG encodes images in the YCbCr color space and commonly subsamples the chrominance channels (typically 4:2:0), storing color at lower resolution than brightness because the eye is more sensitive to luminance.[2] JPG has no alpha channel and therefore cannot store transparency. PNG supports 8 or 16 bits per channel and a full alpha channel, allowing variable per-pixel transparency.[3] PNG also supports an indexed-color mode with a palette of up to 256 colors, used for simple graphics to reduce file size.
Compression Behavior
Because JPG discards data, it achieves much smaller files than PNG for continuous-tone images such as photographs, where gradual color transitions compress efficiently and the discarded detail is not readily perceived.[1] PNG produces smaller files than JPG for images with large areas of flat color, sharp edges, or text, where its lossless prediction filters find strong redundancy and JPG's block transform introduces visible artifacts. For a typical photograph, a lossless PNG is several times larger than a visually comparable JPG.
Artifacts and Generation Loss
JPG compression produces characteristic artifacts: blocking along the 8x8 grid, and ringing (halos) around high-contrast edges such as text. These artifacts increase as the quality setting is lowered. Each time a JPG is decoded, edited, and re-saved, the lossy compression is reapplied, so quality degrades cumulatively, a property known as generation loss.[1] PNG, being lossless, exhibits no artifacts and no generation loss: repeated saving leaves the image unchanged.
Standardization and Support
JPEG was standardized as ITU-T Recommendation T.81 and ISO/IEC 10918-1, with the JFIF file convention defining how the compressed data is stored as a file.[1] PNG is defined by the W3C and was published as ISO/IEC 15948; it was created in the mid-1990s partly as a patent-free replacement for GIF.[3][4] Both formats are supported by all major web browsers, operating systems, and image-editing software.
Typical Applications
JPG is used predominantly for photographs and complex natural images, including digital camera output and most photographic content on the web. PNG is used for graphics requiring transparency, sharp edges, or exact reproduction, such as logos, icons, screenshots, line art, and images composited over varied backgrounds. Many workflows store a master image in a lossless format (PNG or another) and export a JPG copy for distribution where small file size is the priority.