Incense

by Bart Arondson

submit your photo


Hall of Fame
View past winners from this year

Please participate in Meta
and help us grow.

Take the 2-minute tour ×
Photography Stack Exchange is a question and answer site for professional, enthusiast and amateur photographers. It's 100% free, no registration required.

I'm new to taking star trails, and fairly new to photography in general.
After my first attempt at taking a star trail, I composited them in Photoshop and when I merged the layers I got this pattern over the image. I'm assuming it's the result of the lens but am unsure, and I have had a hard time defining it which makes searching the internet difficult.

enter image description here

The pictures were taken at the Sunshine Coast in Australia. I was shooting with a 9-18mm f4-5.6 M.Zuiko lens on an E-P3 body. I was shooting to the south east so I'm getting the curvature from both poles in my trails. That's not what I'm worried about however. Its the moire like pattern that results from stacking the individual photos.
The photos were taken at 18mm, with a 30sec exposure, an ISO of 4000 and an aperture of f5.6.
I first tweaked the exposure and reduced noise in Adobe Bridge, then imported the photos as layers to Photoshop CS5 where I used the lighten filter on each layer to show the trails.
The moire like pattern only becomes apparent on flattening the image. I tried another series of images that were taken facing south and didn't get the strange effect.

After doing some more reading this morning I realise I should be exposing for at least a minute and use a combo of screen and lighten to get smoother trails but again, it's the weird moire that has got me stumped.

share|improve this question
    
Are you trying to compile multiple images? If so, can you provide your settings? –  Rob Clement Dec 5 '13 at 0:05
    
From where and pointing where are you shooting? The star trails are really strange to me (changing curvature --- are you on the Equator or near there?) –  Rmano Dec 5 '13 at 1:02
1  
Googling around I see this happens now and then when images are stacked. This heavily depends on the software you use and on the edits you have made (lens correction, contrast/brightness, noise reduction, etc...) before you stack the images. So could you please describe your workflow as complete as possible? –  Bart Arondson Dec 5 '13 at 1:10
1  
I added the extra info you guys were after into the main body of my question. Cheers Mek –  mekugi Dec 5 '13 at 5:04

2 Answers 2

I'm guessing this has to do with two overlays that were very slightly misaligned from each other such that the small variations in each image due to the Bayer matrix become apparent.

If so, this is a rare case where working in raw is actually hurting. Put another way, the raw data has some regular high frequency content due to the Bayer matrix. Normally you don't see this and don't care because the frequency is high. However, if you merge two versions of a picture where one has very slightly different scaling than the other, you get a low frequency beat signal.

To fix this, the best answer is to not try to merge two versions of a picture that have very small differences in their scale. Your description is vague in what you actually did, but if you compoosited two separate pictures, possibly the dimensions of your sensor changed very slightly between the two pictures due to a change in temperature. This would only apply if you separately scaled the two pictures to overlap them. The high frequency Bayer noise of any two raw images from your camera would be the same, even if the angle of view they represented changed slightly due to dimension changes.

Try filtering each picture down by 2x before compositing. That should eliminate the Bayer high frequencies so that no beating results after compositing.

share|improve this answer
3  
possibly the dimensions of your sensor changed very slightly between the two pictures due to a change in temperature Sounds interesting. Any source or reference on this? –  Bart Arondson Dec 5 '13 at 16:22
    
@Bart: Physics says this should happen, and at the resolution of modern sensors the effect could well be pixel-sized. However, on thinking about this more I realize that this wouldn't matter unless the resulting pictures were scaled by looking at their contents, like a program that finds overlaps might do. –  Olin Lathrop Dec 5 '13 at 18:06
    
Cheers Olin, I'll try shrinking the pictures and compositing them again. I didn't scale any of the images, just dumped them into PS from Bridge. –  mekugi Dec 5 '13 at 20:37
1  
It definitely looks as though the images were stacked as a RAW bayer grid, rather than RGB pixels...and that the grid was misaligned. I would offer that first converting the images to TIFF, then stacking, would probably solve the problem. –  jrista Dec 6 '13 at 2:51

At first look it appears to be a very slight moir caused by the lens, which only comes to light when it's amplified with the stacking effect. Have you got a different lens that you can try, and then take similar images? Or even try similar images but at somewhere other than 18mm? Try at 9mm and about 14mm. Then do all the rest of the process in exactly the same way as you did for this image to see if the problem is recreated. It's possible that even changing to 16 or 17mm could eliminate the problem.

share|improve this answer
4  
Lenses don't cause moire as they are analog devices, moire is caused when you have two grids of close but different sizes that interact, such as a tight weaved fabric and camera sensor pixel array. The lens can only influence moire by blurring the image of the grid so that it no longer interacts with the sensor. –  Matt Grum Dec 5 '13 at 10:00
    
Thanks Laurence, I'll wait for another clear night and try some different lenses/focal lengths and see what happens. Luckily that's the view from my front balcony. –  mekugi Dec 5 '13 at 20:39

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.