Charge Binning
Charge Binning is a pixel-level technique in CMOS (and CCD) image sensors where charges from multiple adjacent pixels are combined on the sensor before readout, effectively acting like a pixel “merger” to improve sensitivity and/or frame rate.
How It Works
- Photon Capture: Each pixel collects photo-generated charge during exposure.
- Charge Combination (on-sensor): Instead of reading each pixel individually, the sensor combines the charges from a group of pixels (e.g., 2×2 pixels) in the charge domain before conversion to a digital signal.
- Single ADC Conversion: The combined charge is sent to the readout amplifier/ADC as if it came from a single, larger pixel.
Effects & Benefits
- Higher Sensitivity: By adding the charges together, you get more signal (more electrons) per readout, which improves the signal-to-noise ratio (SNR), especially in low light.
- Reduced Read Noise Impact: Since multiple pixels are read through one conversion, read noise is applied only once per combined group, not per pixel.
- Higher Frame Rates: Reading fewer “binned” pixels reduces the amount of data that must be transferred, allowing higher frame rates.
- Effective Larger Pixel Size: Example: 2×2 binning combines four pixels into one, effectively doubling pixel pitch in each dimension.
Trade-Offs
- Loss of Spatial Resolution: Combining pixels reduces image resolution. A 2×2 binning mode produces ¼ the original pixel count.
- Potential Saturation: The combined charge may exceed the well capacity of the binning node, leading to clipping in very bright scenes.
- Color Accuracy Impact (for Bayer sensors): Mixing pixels of different colors in charge domain binning can alter color fidelity unless special algorithms are used.
Charge Binning vs. Digital Binning
- Charge Binning (on-sensor): Combines electrons before ADC → better SNR improvement.
- Digital Binning (after readout): Averages/combines pixel values in post-processing → no read noise benefit.