Sensor analysis - Touptek native driver incorrect readings. Works fine with Ascom driver.

[discophil]

Sensory Analysis giving inaccurate readings using the native Touptek driver, due to Black Level switching /jumping to 794 every time you run the sensor analysis.

To reproduce > Black level set to 0 for example. Then open sensor analysis and click start and black level jumps to 794. Unable to keep it at 0 causing incorrect results for dynamic range , read noise measurements etc.

Using the Ascom drive works properly correctly. See measurements in the attached image.

* V4.1.12516.0
* Touptek ATR2600C
* Windows 11
* No logs, sorry

[admin]

Hi,

the sensor analysis deliberately sets the black level to an appropriate value that will ensure that the histogram is separated from the left hand side (no zero value pixels), even at high gains. This is necessary for correct measurements throughout the process. Unfortunately, ASCOM drivers don't allow SharpCap to adjust the black level directly, so the user selected black level will be used, even if it is inappropriate.

I'd be pretty confident here that the read noise values for the ASCOM driver are the incorrect ones - if the black level is set too low then you get 'black level clipping' in the captured frames. This occurs when there are significant numbers of pixels at pixel value zero - some would have had lower values except that the zero is the minimum value possible. When this happens, things like read noise measurements are artifically reduced, giving incorrect values.

I do have the mono version of that camera, so I will check the analysis, but if you look at the equivalent ZWO ASI2600MC, the published analysis data (https://www.zwoastro.com/product/asi2600/) shows a read noise of about 1.3 at the point where the camera switches into HCG mode (100 gain on the ZWO), which fits much closer with the 'Native' results than the 'ASCOM' results.

cheers,

Robin

[discophil]

Makes sense, but an offset of 794 is way to high for this camera.

Offset to avoid clipping is about 150/160 (so says cuiv the lazy geek). I used Nina to check for clipping/offset and got a similar value. I'm setting about 200 offset in Nina for my camera. Basically all the same as Cuiv as per his video guide here https://youtu.be/mCYQoVSuEaI?si=wieW-MvV9vO8MfRf

Is sharpcap calculating the offset incorrectly at 794 as this doesnt match nina stastitics for calculating the offset to avoid clipping?

[admin]

Hi,

the value chosen for black level is picked to ensure that there is no black level clipping at maximum gain, then that same value is used for all gain level measurements. If you look at the histogram at max gain in HCG, 16 bit, then you will see that actually 794 might even be a bit on the low side. At lower gains (the sort that you would realistically use for deep sky imaging), lower black levels are sufficient.

cheers,

Robin