Download JPG sample images

The JPEG standard was created in 1992 by the Joint Photographic Experts Group, a joint committee of the International Organization for Standardization (ISO/IEC JTC 1/SC 29/WG 1) and the International Telecommunication Union (ITU-T Recommendation T.81 / ISO/IEC 10918-1). Its objective was establishing an open, cross-platform standard for compressing continuous-tone grayscale and color still images.

The distinction between .jpg and .jpeg stems entirely from early PC operating systems. MS-DOS and Windows 3.1 enforced the FAT16 8.3 filename limitation (maximum 8-character name and 3-character extension), requiring the format name to be truncated to .jpg. Conversely, Unix, Linux, and classic Mac OS supported 4-letter extensions and retained .jpeg.

Today, .jpg and .jpeg are 100% interchangeable synonyms representing the exact same underlying binary bitstream, MIME type (image/jpeg), and decompression algorithms.

Color Space Conversion: RGB pixel values are transformed into the YCbCr color model, separating luminance (Y) from chroma blue-difference (Cb) and chroma red-difference (Cr). Because the human visual system possesses substantially higher spatial acuity for brightness than color, chroma channels can be subsampled (e.g., 4:2:0 or 4:2:2) to halve color data with imperceptible loss.

8x8 Block Partitioning & FDCT: Image channels are divided into non-overlapping 8x8 pixel blocks. Each block undergoes Forward Discrete Cosine Transform (FDCT), converting spatial pixel intensities into 64 frequency coefficients: 1 DC coefficient representing average block brightness and 63 AC coefficients representing high-frequency visual details.

Quantization & Huffman Encoding: High-frequency AC coefficients are divided by values from a human-calibrated quantization table and rounded to integers, discarding psycho-visually redundant information. Finally, non-zero coefficients are zig-zag scanned and compressed using Run-Length Encoding (RLE) and Huffman entropy coding.

Baseline JPEG decodes image data sequentially from top to bottom, one horizontal scan line (MCU row) at a time. On slower network connections, users see a blank screen followed by a descending curtain of imagery.

Progressive JPEG reorganizes the bitstream into multiple progressive scans. The first scan renders a low-resolution, blurry preview of the entire image almost instantaneously. Subsequent scans inject higher-frequency spectral coefficients, progressively refining image sharpness until reaching 100% clarity.

Progressive encoding typically produces 2% to 10% smaller overall file sizes than baseline encoding for photographic content due to improved entropy coding efficiency across multi-pass coefficient groups.

A raw JPEG file is structured as a sequence of marker segments, each starting with the byte 0xFF followed by a marker code. The file begins with SOI (Start of Image, 0xFF 0xD8) and terminates with EOI (End of Image, 0xFF 0xD9).

JFIF (JPEG File Interchange Format): Encapsulated within the APP0 marker (0xFF 0xE0), JFIF defines pixel aspect ratio, resolution density (DPI), and basic RGB color space constraints.

Exif (Exchangeable Image File Format): Encapsulated within the APP1 marker (0xFF 0xE1), Exif was standardized by JEIDA to store digital camera metadata, including camera make/model, shutter speed, aperture, ISO sensitivity, focal length, GPS coordinates, and embedded thumbnail previews.