Limited time offer - Save 30 percent on DxO's powerful photo editing software bundle. Learn more






Essential characteristics of noise

Sunday May 10 2009

Sensor Insight
Introduction | Noise in shadows | Noise in mid-tones | Noise in highlights | Summary | Annex
Essential characteristics of noise

Noise is an integral attribute of camera sensor performance and understanding it has been essential to establish a reliable and objective DxOMark Sensor Score. This section discusses the nature of the three principal kinds of noise: photonic, thermal or electronic, and pixel response non-uniformity (PRNU), and explains how shape of Signal-to-Noise Ratio (SNR) curves relates to three metrics of DxOMark—dynamic range (low illumination), SNR (mid-tones), and color sensitivity (highlights).

“Know your enemy.”

—Sun Tzu, The Art of War (6th century BCE)

Serious photographers often find themselves battling with noise. They visit websites, read manuals, and consult with one another about the best ways to fight noise in shadows, in mid-tones, and in highlights. Although mastering specific techniques is crucial to success, it can also be very useful to have a greater technical understanding of what noise is, and what its principal causes are. With DxOMark, we shed light on some metrics that are derived from noise characteristics.

Noise is an integral attribute of camera sensor performance, and is a complex phenomenon that cannot be summarized by a single value. Even though different cameras may produce significantly different amounts of noise, signal-to-noise ratio (SNR) curves (which display SNR as a function of the signal value) always have the same shape: SNR, expressed in decibels (dB) starts from very negative values, increases very fast for small signal values, then increases less and less, and eventually seems to stabilize at a limit value. The definition of SNR and why it is important for image quality can be found here.

SNR Curve