It is important to understand which camera type suits your application. However, it is as important to configure the camera to the needs of your application. The right configuration can result in an optimization of dynamic range, maximum full well or sensitivity. A parameter that can heavily influence these specifications is the analog gain of the sensor. Some image sensors allow for the adjustment of analog gain. Most Adimec cameras implement this feature, so the analog gain can be adjusted to fit the needs of the application. The first part of this blog series describes the advantages of using analog gain instead of digital gain. The second part focuses on the application possibilities of different analog gain levels.
The Programmable Gain Amplifier (PGA) is an amplifier that can control the analog gain to different levels. Some cameras, like the Q-21A230, have a PGA built into the sensor architecture. A block diagram has been made to show a simplified view of the signal path and where the PGA is located.
To demonstrate the impact of analog gain in comparison to digital gain, we will compare measurement data of the Q-21A230 using either 1x gain, 4.8x analog gain or a 4.8x digital gain. We have performed EMVA1288 measurements to show the difference in a few important camera parameters. The results can be found in table (1).
|Read noise (e–)||16.0||3.3||16.0|
|Full well (ke–)||33.6||7.0||7.0|
Table 1: Measurement data Q-21A230 analog vs digital gain at 529 nm in 12 bit
As can be seen in table 1, the 4.8x analog gain outperforms the 4.8x digital gain in several important parameters. We will analyse these differences in more detail.
Table 1 shows that the measured full well size decreases from 33.6 ke– to 7 ke– when the gain is increased to 4.8x. This is the same for both analog and digital gain. In reality, the full well of a sensor cannot change, but the full well size is not measured in electrons directly. The EMVA measures the video level in digital numbers (DN) and from there the full well size is calculated. So, applying a gain means you need less photon converted electrons to reach the maximum video level. Therefore, we say that applying a gain decreases the effective full well size.
Read noise, also called temporal dark noise, is the noise source related to the image sensor and readout circuits. It is a factor that determines the sensitivity of the camera, because the lower read noise, the lower the minimum number of photo-electrons that can be detected. Table 1 shows that applying a digital gain has no effect on the read noise. However, when an analog gain is applied, the read noise decreases from 16 e- to 3 e-.
Maximum Signal-to-Noise Ratio
As the read noise decreases when an analog gain is used, you might think that the Signal-to-Noise Ratio (SNR) would also be better for the analog gain compared to the digital gain. However, the maximum SNR is only marginally better for the analog gain because the read noise only has a minor contribution to the maximum SNR. The shot noise, which is approximately equal for both the analog and digital gain, is the dominant factor influencing the SNR. This can also be seen in this formula:
As the effective full well size decreases when using gain, either analog or digital, the relative effect of the shot noise on the SNR becomes larger. Therefore, the maximum SNR decreases almost the same for analog and digital gain.
Increasing the digital gain does not improve image quality. But some applications require the use of gain to increase brightness. In that case, analog gain can actually improve the quality of your images. As shown here, using an analog gain increases the sensitivity of your camera by decreasing the read noise, making it possible to measure smaller differences in light intensity. This can be seen as an improvement in image quality because you can then observe signals that without a gain were not possible.
However, improving the lighting or increasing exposure times will give better results when it comes to SNR, as this is mainly determined by the full well size. The drawback is that a longer exposure time reduces the maximum framerate and increases dark signal non-uniformity. In addition, it is not always possible to improve lighting. Therefore, in these situations applying a gain offers a solution. In which case, it is much preferred to use the analog over digital gain as was made clear in this blog.
In the second part of this blog series, we will focus on the different analog gain modes of the Q-21A230 CoaXPress camera, and for which application types they are best suited.