Astrophotography with G9

[un hombre]

Hi all,

According to this site, 600/500/400 rule is not really a good starting point for all sensors because of their physical sizes, pixel density and because star trailing takes different time for different sensors: https://starcircleacademy.com/2012/06/600-rule/

So I did all the calculations from the site for my G9 sensor and 15mm PL 1.7 and it seems that it takes 1.5 seconds for a star to shift from one pixel to another and start star trailing on Panasonic G9 and 15mm lens:

Panasonic G9 sensor – 17.3mm x 13mm, resolution 5184 x 3888

t = 0.00333/(30mm/13750) where 30 is ff equivalent.

If I want decently sharph photos I should probably allow for 3pix star trailing tops and this means around 5 seconds of exposure max... And that is what I did and I can confirm that with pixel peeping on my phone - so imagine what would happen if one applied 500 rule and allowed for 10-15 seconds long exposure.

Anyway, I am currently on Tenerife and I went to see El teide and did some astrophotography. I had completely clear skies and beautiful milky way but because I allowed myself only for 5 seconds tops exposure (I will do image stacking anyway so I made lots of photos) I had to bump up iso to 3200-6400.

I cannot judge the photos on the small screen of my telephone so I am attaching examples here for you to have a look if you are curious.

Here are raw files with jpeg from camera:

https://www.dropbox.com/s/ubtmrtbxfxs0jzm/teide.zip?dl=0

If anyone has time, can you see how much quality image you can get from post processing?

 

[boxerman]

Your drift is right, but I think you're going a little too far. I did roughly the same calculation here calculating start trails and came out with about the same numbers. But if you look at the display you are going to use, you might come out with different conclusions. On a 13" display, I think 7-10 pixels is enough so that the trails will not be bad. I got into this because the star trails on a 24" wide print bothered me. But, I came to the conclusion that 5 pixels would probably be sufficient for this 24" print (tho I didn't re-do the shot by flying back to New Zealand!).

(In my post linked above and subsequent posts in that thread, I show some images on which you can make some visual judgments.)

In general, I've adjusted my personal rule of thumb down to 150 / f (for m4/3; equiv. 300 / f for full frame). I think that's about 5 pixels shooting in the equatorial plane (worst case).

One of the things I noticed was that bright stars were already 3-4 pixels in diameter, so keeping the trail down to 1 pixel is overkill. But, look at some of your pix, and make your own judgment.

The penalty you pay for tiny exposure times in noise is quite horrendous. I haven't stacked, yet, but even with my (generous by your standards) exposure times I'd need some stacking to satisfy me fully. 4/3 sensors, as yet, are at the hairy edge of astrophotographic competence.

 

[un hombre]

Tkanks for your input.

I guess 5 pixels might be ok for my needs which would increase the exposure to 8 seconds and maybe allow me to decrease iso to 3200.

I am tempted to drive to teide tonight again and try different settings.

How do the pictures look on the big screen?

Can anyone evaluate?

 

[boxerman]

I guess 5 pixels might be ok for my needs which would increase the exposure to 8 seconds and maybe allow me to decrease iso to 3200.

I am tempted to drive to teide tonight again and try different settings.

How do the pictures look on the big screen?

I took a quick, informal look. There is no discernible trailing at any magnification I tried. So, I think you are significantly into overkill range. Expand to look at the pixel level, and take a look, especially, at the size of bright stars relative to your (and my) computed pixel trail length.

The 6400 shots look fine at normal size (on a 15" laptop). However, to me, they get really scary when blown up; the noise is quite bad. "Wormy" irregularities in the noise, pattern, too. 3200 is much better, even if not optimal for Milky Way at that exposure duration. Some "wormy" pattern in the noise, but much less.

Personally, I'd say you can go up to my personal rule of thumb, 150 / FL, which would be 10 secs exposure. Again, if you're going to print large, you might still have some trail worries. For screen viewing, I think you'd be fine.

 

[Bill Wallace]

Why not go and shoot perhaps the MW start at iso1600, as wide an f stop as your lens will allow at about 20 seconds and adjust from there. Doing it your way I would have analysis by paralysis as the saying goes. I shot some Astro/MW last summer with my Em 1 which is not the best for this type of work and got very passable work. I used the 12 f2 and the FE 1.8. I also shot with my em5. Needless to say from the above I am a seat of the pants shooter.

My thoughts

Bill

Not sure if my shots are in my galleries here or not but I do have a gallery on my site titled Grand Marais night sky. I also shot some in Death Valley last spring.

 

[boxerman]

Here's a sequence of night exposures at 200 ISO (!), 20 mm lens, 30 seconds, with successively greater post processing. You can do quite a lot, I discovered, with fairly severe curves adjustments. There's software that will try to keep the sky background dark, too. You might look into that.

In these, I exposed so that the brightest star was JUST at clipping levels. WAY underexposed for Milky Way, but you will see, PP makes it usable. 30 seconds makes star trails visible at the edges, but really not terrible, especially if you're aiming for screen display. This is an E-M5, so you're undoubtedly better off!







--
The BoxerMan

 

[Mark9473]

How do the pictures look on the big screen?

Can anyone evaluate?

I looked at a few jpegs and they have ugly blotches in the sky background and also what seems like severe jpeg compression. It's not a pretty sight unfortunately. The lens is also affected by a bit of coma and astigmatism, especially towards the corners, and chromatic aberration on the brighter stars.

I would suggest trying a bit less extreme camera settings, stopping the lens down to f/2 or f/2.8, and stack a bunch of images.

 

[un hombre]

Thanks for suggestions, Mark.

I think I will go tonight and experiment with some other settings.

I attached jpegs just for the record, I do not intend to rely on them and expect the raws to be much more usable.

 

[Vlad S]

If I want decently sharph photos I should probably allow for 3pix star trailing tops and this means around 5 seconds of exposure max... And that is what I did and I can confirm that with pixel peeping on my phone - so imagine what would happen if one applied 500 rule and allowed for 10-15 seconds long exposure.

In selecting the exposure time you will have to compromise between sharpness and noise. I find that a little less sharp, but a lot less noisy images are far more satisfying. For 20mm lens I usually go with 15 seconds exposure, I think you can do at least that with a 15 — probably even longer, and still produce beautiful images.

Vlad

 

[Gary from Seattle]

I get what I find to be very good images with the 8mm FE F1.8 at 25 seconds ISO 1600. The math I use is 250/8 = 30 seconds and that also looks fine but I can get by exposure-wise with 25 seconds.

I think with the 8mm FE, though, the stars are very pinpoint and motion is probably harder to see on the image, as is overexposure of any particular stars. As the FL goes to 15mm I would think the stars will be more critical for motion and overexposure. I would shoot 15mm @ ISO 1600-3200 (max) at 250/15 = 17 seconds or go less than that if I liked the results.

 

[Adrian Harris]

I normally shoot astro at around 7-9mm and for about 40-50secs. On a GX8 at iso1600 (preferably).

I do get star trails but have learned a technique from YouTube that not only reduces the trails very successfully but also appears to half the noise.

It requires the use of Photoshop and is well worth googling for.

Best wishes.

 

[l3gion]

I normally shoot astro at around 7-9mm and for about 40-50secs. On a GX8 at iso1600 (preferably).

I do get star trails but have learned a technique from YouTube that not only reduces the trails very successfully but also appears to half the noise.

It requires the use of Photoshop and is well worth googling for.

Best wishes.

 

[Adrian Harris]

I normally shoot astro at around 7-9mm and for about 40-50secs. On a GX8 at iso1600 (preferably).

I do get star trails but have learned a technique from YouTube that not only reduces the trails very successfully but also appears to half the noise.

It requires the use of Photoshop and is well worth googling for.

Best wishes.