Skip to content

JPEG Compression

Definition

A lossy codec that divides an image into 8×8 pixel blocks, applies the discrete cosine transform, quantises the coefficients, and entropy-codes the result. Every save round at a given quality setting introduces a new quantisation layer whose artefacts are detectable.

Compression type
Lossy (information is discarded)
Block size used
8 by 8 pixels
Forensic suitability
Not suitable for archival or evidence preservation

Common questions

Why can't you use JPEG for archival forensic images?+

JPEG is a lossy format that discards fine-detail information permanently every time you save. Once that detail is gone, it's irrecoverable, making JPEG unsuitable for evidence that needs to be preserved exactly as captured.

What happens each time you re-save a JPEG image?+

Every save introduces a new quantisation layer. The artefacts from this quantisation are detectable and represent data loss from the previous version. If you save a JPEG multiple times, each round adds another layer of degradation.

How does JPEG compression actually work?+

JPEG divides an image into 8x8 pixel blocks, applies a discrete cosine transform to convert the pixel data, quantises the resulting coefficients to reduce them, and then entropy-codes the result. This process discards information at the quantisation step.

Related terms

Bayer CFA
A colour filter array arranged in a repeating 2×2 mosaic of red, green (×2), and blue filters over the sensor. Each photosite...
BSA 2023 § 63
The electronic record admissibility provision (replacing IEA § 65B), requiring a certificate from a person in a responsible official position confirming the...
Colour Space
A defined mapping between numerical RGB values and real-world colours. sRGB is the web default with a smaller gamut; Adobe RGB covers...
Daubert Standard
The US federal evidentiary standard (Daubert v. Merrell Dow Pharmaceuticals, 1993) requiring that expert testimony be based on scientifically valid methods with...
Demosaicing
The algorithm used to interpolate full-colour pixel values from the single-channel raw sensor data. Common methods include bilinear interpolation, gradient-based algorithms (AHD,...
DNG (Digital Negative)
Adobe's open-standard RAW format providing a vendor-neutral archival container for camera sensor data; accepted as an archival format by SWGDE and several...
EXIF (Exchangeable Image File Format)
A metadata standard embedded in JPEG and TIFF files by capture devices. EXIF fields record camera make and model, lens data, exposure...
Fixed-Pattern Noise (FPN)
Any repeatable, spatially correlated noise in an image sensor, as opposed to random shot noise. PRNU and DCNU are both forms of...
Image Enhancement
Processing of a digital image to improve the visibility of features already present in the original data; permitted under SWGIT, FBI, and...
MD5
A 128-bit hash algorithm widely used in legacy forensic systems; known collision vulnerabilities (Wang and Yu, 2004) mean SHA-256 or SHA-3 are...
R V. Tobi (2019)
UK Court of Appeal decision affirming that forensic image enhancement is permissible provided the original is preserved, the enhancement steps are documented,...
RAW File Format
The unprocessed output of a digital camera sensor, stored before demosaicing, white balance, or tonal curve application. Preserves full bit-depth and all...

Explained in these topics

Your journey to becoming a forensic professional starts here.

Practice with mock tests, learn from structured notes, and get your questions answered by a global forensic community, all in one place.