Perseids Attempt - Another Learning Experience

[Mike Lewis]

So I think one of the best things about FocalWorld is the great resource that we have here of talented, experienced, and helpful photographers of almost every genre. This comes in handy when one's outing does not produce the results hoped for, as there is almost always at least someone here who is able to chime in with some good input on improving things for the future. So in that spirit I am posting 3 images from my night recent out in RMNP trying to get some images of the Perseids this year. I have 2 versions of my final composite image to show, as well as a single 'raw' frame that illustrates the conditions and colors I was seeing right out of the camera. I feel like for me this represents an improvement over past efforts at shots like this, while still falling far short of where I eventually want to get to.

Background:

• I planned on getting up to Rainbow Curve in RMNP to shoot some of the Perseids. This seemed like a decent location, although in practice there was some strong light pollution from the cities out to the east of the park.
• I had 2 lens choices, a 24mm f/1.4 Rokinon (not really as wide as I wanted) and my EF 16-35 f/4 zoom (not really as fast as I wanted). I went with the Rokinon.
• I went with 8 seconds at ISO 4000. I probably could have pushed both of those a little harder, but I am really a stickler for round stars and am new to having a camera with decent high ISO so I was being overly conservative there.
• I could not get out on the peak night for the shower, ands so I had what were greatly reduced rates over what would have been possible to see the night before. Additionally, the night I got out Friday night/Saturday morning) I had a commitment until about 9 pm in Loveland, so did not get up to the park and all set up until well after dark at about 11:15. This prevented me from really setting up much of a composition or from getting any sort of an initial foreground exposure prior to full darkness. The moon was setting just as I arrived so no moonlight in the shot.

The next 2 images are composites, with a single frame as the background and then all the meteors added by layer masking in Photoshop. The first version I did I did not like the way the white balance edits to the images seemed to destroy the natural color I had captured for the meteors themselves. I also realized I wanted to pick one of images that had M31 showing as the frame for the background (M31 rotated out of the FOV as the night wore on...). Here is that 1st version.

I went back and reworked the image to try to keep the meteor colors while still ending up with the background frame having a decent color to the sky. This was much more challenging, but here is the result from that set of edits:

The 2nd image, with the more pronounced meteor colors really does represent the meteor color captured right out of the camera. But also right out of the camera the sky had that characteristic brown coffee look to it from light pollution. Here is a single frame to show what I mean:

So I consider this to be another learning experience image, and not anything too impressive. I would love to get some feedback from those at FW more experienced with these type of images.

I ended up with over 700 frames from the outing - I may try to make them into a time-lapse at some point in the future. Of course, the meteors will go by so quickly as to be invisible, but the sky motion might be pretty cool anyway.

ML

 

[Jameel Hyder]

I really like your first image Mike in that the sky has a good color and luminosity. The second image has the sky much bluer and the meteor colors don't go well with that sky color (to my eye anyway). The first image of the sky with the meteors a bit warmer than it has but less warm compared to the second will be my choice.

Thanks for sharing your experience however. The Rokinon worked out nicely. I normally go as high as 15s with it wide open. Perhaps somewhere between your 8s and 15s with ISO pushed to 6400 would have been ideal. The noise clears up pretty well even in ACR. If your goal is not timelapse you can also set the High ISO NR to on which doubles the exposure subtracting the dark frame from every exposure.

Overall nice work.

 

[AlanLichty]

Another thumbs up for that first image - that one feels like the winner for my eyes.

 

[Ben Egbert]

I am going with the 2nd image. I like the darker sky and the way the meteors stand out. Thanks for posting this with all your info, it all helps when I am out to know what others are doing.

 

[JimFox]

You are being way to negative Mike with your results. They look pretty good I would say.

I would say 8 secs is way way way too short. If you are going to pixel peep that much you need to start using a tracker then. 8 seconds is just way too short as Jameel already said. But I would say even 15 seconds is too short for a 24mm. Personally I use 20 seconds normally to a max of 25 seconds if I shoot at 24mm. But then I don't pixel peep. I of course look at it at 100%, and I process it to look the best it can, but I grade my Night time star and Milky Way images the same way I grade my normal landscape images, how does it look printed? And how does it look viewed at a normal viewing distance for prints?

But as I have mentioned before, there is no right or wrong way really, so if you want to shorten your shutter speed and increase your ISO, that's your choice to make. But if you end up having to do median stacks of multiple images to deal with the noise, you might as well just use a star tracker then.

As for the WB, I don't see what the issue is? For stacking meteors like this, you process your base (master) image according to the WB and exposure that appeals to you. Then go and process each of the meteor streak shots (usually as batch) to make them look their best. Then blend those ones in.

The last you one you posted, that color of sky isn't uncommon at all. It's pretty normal for it to get that coloring at some point during the night.

For the timelapse? I don't know why you are saying it's impossible to see the meteors in a timelapse. I see them in mine, and I have done quite a few timelapses with meteors, so I am not sure why you think it's impossible to have them show up. If you have 700 frames processed, you should create a timelapse with them. It would be great to see you do one.

Back to your meteor images, I like #2 the best as the meteors have the best color. You have kind of lost the tail of the Milky Way in it. It would be nice to see it more pronounced. The silhoutte of the ground is okay, but it would be nice to see the mountains with some detail. You can't overexpose the image a couple of stops and pull the shadows up 100%? That's what I do on many of my images from my D850.

 

[Mike Lewis]

Thanks for all the feedback. I will have to play with some longer exposures versus trailing versus ISO the next time I am out for nightscape.

I had some foreground images that were taken at longer exposure but I just didn't really like them. Perhaps it was more about the subject itself than anything else. I have started using the DXO PureRaw program (instead of Topaz DeNoise) on a number of my higher ISO images. I think it does a better job. The constraint is that it has to be used on the raw file (think of it as a preprocessor) and it outputs a DNG as output. That approach does not work for astrophotography images, but does work well for nightscape or even some higher ISO wildlife images. The R5 noise isn't really too objectionable at 4000 ISO but it cleans up fantastically well on the DeepPrime setting of PureRaw.

I appreciate the opinions on the 2 images. I lean towards the 2nd one myself as I really like the meteor colors to look more as they were captured. As an astrophotographer I am WELL aware of the normal color to the sky with any sort of LP. I was posting that shot mainly to illustrate the color of the meteor straight out of the camera as well as the extra work involved with getting the masked meteors into the composite when the sky colors differ so much.

I will continue to pursue the time-lapse. Perhaps it will create an interesting result.

ML