[Readout Noise] – What is Readout Noise ?

time22/05/13
leibieProduct

When a camera measures the amount of light it has captured in each pixel of an acquisition, there is always some degree of error. This inaccuracy is called Read Noise, or Readout Noise.

As signals comprising different numbers of photons are captured and converted into electrical signals measured in electrons, the readout noise is specified in numbers of electrons (e-). Due to the precision of electronics in modern scientific cameras, this readout noise is typically very small, on the order of 1 to 3e- for low-light imaging cameras.

For applications with high light levels, such as where thousands of photons are captured by each pixel, this error bar is miniscule compared to the signal, so read noise of less than 5e- can effectively be ignored. For example, compared to a signal of 2,000 photoelectrons, read noise of even 10e- would make a less than 3% difference to the signal-to-noise ratio, and likely be imperceptible. However, for low light levels where photon counts may be in the tens of photons, low read noise can play a significant role in the signal to noise ratio and image quality.

Due to their parallel architecture, all CMOS cameras exhibit a distribution of read noise values from pixel to pixel. Therefore, sometimes two values for the read noise in e- are quoted on specification sheets. The Median value is specified so that 50% of pixels have a read noise value at or under this figure, and provides insight into the typical read noise value for the camera. The Root-Mean-Square (RMS) value specifies the root mean square of the whole read noise distribution, providing insight into the extent of high-read noise pixels that are not included in the Median measurement.

Some specialized low-light imaging cameras have a low noise mode called Correlated Multi-Sampling mode, or CMS. In this mode, some reduction in frame rate is traded for more precise signal measurement, leading to read noise figures of only around 1.1e- (Median) / 1.2e- (RMS).

share to :
  • twitter43black
  • Youtube-01
  • ins (2)_black
topPointer
codePointer
call
Online customer service
bottomPointer
floatCode