I'm using SkyScan 1172. According to one of the Bruker's manual named "SkyScan 1172 How to set up a scan?", for 4K camera mode with No Filter,
- Voltage = 40--50 kV: but it turned out this resulted in some star-like ray noises. As per the suggestion from the technician, I raised it to 70 kV and it got better.
- rotation step <= 0.3 deg: "Preferably lower the rotation step (instead of increasing frame averaging) for low dense samples when the signal to noise ratio is too low."
- frame averaging 4--8 frames: "Preferably increase the frame averaging (instead of decreasing the rotation step) for high dense samples when the signal to noise ratio is too low."
Considering my sample is of very low density material (which does not allow me to use even a thin filter), it is likely that I should decrease rotation step a lot and use moderate number of frames for averaging. However, there's yet another potential consideration that I am suspecting: motion blur. My sample is quite susceptible to the vibration, I can see the blur when I press the "Grab" button immediately after rotation or translation. Therefore, I am suspecting that during the actual scan the first image frame for each position can be slightly blurry. If this is true, increasing frame averaging can be quite important to reduce the noise due to blur. I usually take 0.1--0.15 deg rotation step and 3-5 frame averaging but maybe I should try increasing the frame averaging a bit more.