>>3913451>>3913452The idea is that when stacking multiple photos of dim deep sky objects, every subexposure is limited by the exposure length and the target will always be underexposed, so you want an ISO low enough to not overexpose the stars but high enough for the signal to drown out the bias noise present at the bottom of the histogram (picrel)
The whole iso invariance thing doesnt really matter for single exposures though, for that just expose the photo as much as you can without making the stars trail, and use a high iso that gets you the shot
like
>>3913440 said, the 500 rule can help to tell how long an exposure you can take before the stars trail. There's also the NPF rule (
https://www.lonelyspeck.com/advanced-astrophotography-shutter-time-calculator/), which takes into consideration that parts of the sky closer to the celestial poles move across the sky slower. But honestly these are just good starting points, and you'll get the most accurate result for your camera and lens by trial and error
