Hacker News Reader

Comment by pietroq | Hacker News Reader

Let's put this to rest for once and all. The average radius of Earth is 6,371 km. Initial sats will be at 550 km height (many will be higher, many somewhat lower, but does not make much diff.) So the surface of the 6,931km radius globe is ~603,672,875 km2 (~604 million square kilometers). Don't have exact size for a Starlink sat, but a good upper bound is 3 x 2 m, i.e. 6 m2 = 0.000006 km2. Let's say there will be 40,000 such sats at peak, so altogether they occupy 0.24 km2 area. Altogether, they will occupy 0.00000003976% of the sky. QED.

View on HN

Looks like you've never done astronomy. A single satellite passing through the shot can saturate the sensor and ruin a multi hour exposure.

The satellite would not be passing through the shot for the entire exposure.

If you had software to detect and delete frames with a satellite moving through them, how many frames from a multi-hour exposure would have to be deleted?

I assume it can even be done digitally in post-production, but even if you had to actually close the aperture ahead of the satellite crossing through your frame, we know exactly where the satellites are going to be so we could calculate when they would enter and exit the frame.

We studied this technique in the graduate-level observation class I took 30 years ago. It is not trivial for a lot of reasons, especially for modern wide-field telescopes with multiple satellites in the field of view at all times.

I appreciate the expert feedback. A lot of very complex things often seem trivial from an outside perspective.

I think everyone can agree this will be at least somewhat a burden on astronomy and will at least marginally increase the cost in obtaining certain imagery.

I guess it’s possible that there are certain types of images which may even become prohibitive or impossible to obtain, but I personally haven’t read a scientific analysis making this claim.

I also think most everyone is agreeing that we attack this problem from both sides. Limit the reflectivity of the satellites and work on enhancing software to mitigate the effects.

In the end there is a tremendous benefit to society both from astronomy as well as having affordable worldwide broadband internet.

I’m not going to go so far to say that we need to sacrifice one for the other, but I think the capabilities and opportunities provided by Starlink to billions of people are too great to pass up. Even if there is a significantly increased cost to perform certain observations. Even if it means some observations can only be made from orbit in the future. Even if some observations become entirely impossible? Hopefully that can be avoided.

You had digital sensors thirty years ago?

Yes. In the case of the system I used in class, it was a CCD intended for a TV camera, hence pretty low resolution. Pre-HD. But this was something appropriate for a one meter telescope, not a big one.

Er . . .

A long exposure isn't the same as frame stacking. It isn't taking a traditional video and stacking the frames. Even if a shot is made out of stacked frames, each frame may be 5+ minutes of exposure, enough time for the chance of it being ruined by a satellite to be very high.

Looks like you have never done computers. It is trivial to substract the well known light effect of the sat.

From a saturated pixel on a CCD?

I once made the mistake of shooting a twilight flat when it wasn't dark enough yet, and it took me 20 minutes to pump all of the electrons out of the CCD.

Reply on news.ycombinator.com