Calibration Data for Prime95B on instrument "Salix" from Steven F. Lee Lab

Each camera is calibrated using the mean–variance (photon transfer curve) method: dark/bias frames and a series of uniformly-illuminated flat-field frames at increasing intensity are used to fit, per pixel, the linear relationship between signal variance and signal mean. Five per-pixel calibration maps are produced, each saved as a single-channel float64 TIFF at full sensor resolution: - offset.tif — mean dark/bias signal per pixel. Units: ADU (raw analogue-to-digital camera counts). - variance.tif — per-pixel signal variance measured across the dark/bias frames. Units: ADU². - gain.tif — per-pixel conversion gain, i.e. the slope of variance vs. mean signal across the flat-field intensity series. Units: ADU per photoelectron (ADU/e⁻). - readnoise.tif — per-pixel read noise, computed as sqrt(variance) / gain. Units: photoelectrons RMS (e⁻). - rqe.tif — relative quantum efficiency: each pixel's gain-corrected flat-field response divided by a locally-smoothed (uniform filter) version of the same map. This is a dimensionless ratio, centred around 1.0, capturing only pixel-to-pixel non-uniformity in relative sensitivity — it is not an absolute quantum efficiency measurement. These maps are used to convert a raw acquisition frame (in ADU) to photoelectron counts via: photoelectrons = (raw_ADU − offset) / gain / rqe All five maps are stored as 64-bit floating-point (float64) TIFFs.