[Paul2660]

One of the main reasons I invested in Pentax, was the Astrotracer feature. I have not had good skies until last weekend and finally was able to try it out.

I used the Pentax 15-30, @F3.2 and 15mm, for exposures between 40 seconds and 2 minutes. What I found was that in the center of the image, the solution works OK, (where there would be the least amount of trailing), but as you move towards the edge of the frame, the astrotracing failed 100%. See in the images below, where on the left I have a 30 second image non astro tracked and then a 2 minute image astrotraced on the right, as you can see there are full star trails and the Milky Way is blurred. The camera did attempt to track as the river bluff and trees are blurred as to be expected. However it appears that the Astrotracing feature can't handle a wide lens very well, or least on my camera. Have others used the 15mm focal length or 14mm etc with Astrotracer and had better results? In the crops, the first image is the right edge of both captures and the 2nd is the center of both.

I also noted a band of different color towards the top of each frame, that was consistent throughout all the astrotracer shots and pixel shifted shots I took.

Thanks
Paul C
Photos Of Arkansas

Attached Images

   

---

Last edited by Paul2660; 08-06-2016 at 07:28 AM.

[monochrome]

Hi Paul. Welcome to the Forum. I own a K-1 but I have yet to experiment with Astrotracer since where I live is affected by strong light pollution.

There has been extensive discussion of wide angle lenses and astrotracer on both K-1 and APSc cameras. Here is a link to a specific post from just a few weeks ago describing edge-trailing using extreme-wide lenses on K-1. Scroll to the top and read the entire thread also - very informative.

You can also search on the subject. In the upper right, click 'Search the Forums' box; click Advanced Search link; on the upper left of the next dialog box enter 'astrotracer wide'*; scroll down and click 'Search Now'. The system will return a string of threads containing those terms.

You will need to sort through the posts using the O-GPS1 external module on APSc cameras earlier than K-3II but there's quite a lot of informative usage and workaround help in those threads.

Maybe a couple hours spent on a cloudy evening . . . . .



* Be sure to use the proper product term Astrotracer, not astrotracker as you have styled it in your post.

[Paul2660]

Hello Monochrome,

Thanks for the catch, will fix that!!. And the links. Looks like it's known issue.

Paul C

[mikeSF]

Paul,
the issue you are having is not with the Astrotracer, as i assume it is tracking the earth's rotation correctly and in a linear fashion. When you use a lens that demonstrates extreme radial distortion, the lens is not capable of rendering the perimeter stars in the direction they are actually traveling, and Astrotracer cannot help you. You are trying to defeat physics.
Use a longer lens with less distortion and the results are outstanding.
I do star panos with the FA31 and the tracking and lack of distortion is very nice.
Good luck but please understand the limits of your lens.

---------- Post added 08-06-2016 at 07:51 AM ----------

...
I also noted a band of different color towards the top of each frame, that was consistent throughout all the astrotracer shots and pixel shifted shots I took.

Thanks
Paul C
Photos Of Arkansas

I didn't see an example of this but from your description, that sounds like possible light leakage into your viewfinder, ie from a streetlight or any other stray light behind you. It is a good practice when shooting night images to block the viewfinder for long exposures.

[monochrome]

@MikeSF
As I understand his lament, part of Paul's disappointment is caused by his location in Arkansas trying to capture the entire Milky Way. At our latitude and longitude (I'm in Missouri) he believes that can be accomplished only with a 14 or 15 on FF; which leads to the limitations of lenses.

[beholder3]

This simple high school physics at work. Stars in the corners of a very wide rectilinear lens will show more movement than those in the center.
The corners can't be better using this type of technology and angle of view and shutter times. There are dozens of threads and posts about this limitation out there.

It is just like expecting the coma of a lens to be the same in the center and the edges. There will always be more on the edges.

[Paul2660]

I bet you are right on light leakage. Where I was shooting dos have a street light behind me. I will bring some gaffers tape to cover the eye piece next time.

And thanks for the explanation I figured it was as much which is unfortunate. But I can try a vertical pan next time with the Sigma 35mm.

Paul C

[stevebrot]

And thanks for the explanation I figured it was as much which is unfortunate.

You might want to try a fisheye. The circular projection should help with the exaggerated edge trailing.


Steve

[dave2k]

The K-1 and other recent Pentax cameras are shipped with a viewfinder cover for the light leakage issue. So you can use that instead of gaffers tape if you wish.

[jbinpg]

I found this the best link for understanding why rectilinear ultra wide lenses will always show star trailing at the peripjery of the frame even with the astrotracer function enabled:

https://www.pentaxforums.com/forums/172-pentax-k-3/315335-16mm-samyang-lens-a...ing-k3-ii.html

The post by Nass explains it clearly.

[Paul2660]

Thanks again to all the replies. I have seen this in star trail work, just for some reason did not expect it. I fully realize the effect the rectilinear lenses have especially at the edges. I may try the fisheye and will move up to 24mm or even 35mm to see if I can get a better result. The images are very clean from noise as I had hoped.

Paul C

[interested_observer]

This "problem" of star trailing when using Astrotracking at the extreme edges and corners of wide angle lenses has been irritating me since this time last year. I spent a fair amount of time earlier this year when the Milky Way came back and available again, trying to figure it out. ... and I think I may have identified the problem (well actually Pete_XL).

• The longer the exposure the worse the trailing becomes. I use an 18mm lens. At 60 seconds, really no trailing. At 120 sec it's starting, 240 sec it's very prevalent. To me that has the tell tail feel of error accumulation (40+ years of being an embedded real time software systems engineer).
• An 18mm lens' FOV is 77 by 48 degrees. That is a large chunk of sky, but when you compare that to Nass' diagrams of addressing sidereal along with the direction and shapes of star-trailing, the FOV of the lens should be manageable. I do think that this is a contributing factor or component of the problem - but not the main and only problem.

• https://www.pentaxforums.com/forums/172-pentax-k-3/315335-16mm-samyang-lens-a...ml#post3552604 post #2
• https://www.pentaxforums.com/forums/172-pentax-k-3/315335-16mm-samyang-lens-a...ml#post3552636 post #4

The main reason for my position that sidereal and the direction and shapes of the star trails are not the main component(s) to the problem is that, if it were, we would be reading about this problem when other folks are using their motorized mounts with wide angle lenses - the iOptron SkyTracker, Vixen Optics Polarie Star Tracker, etc. No problem reports from this group, other than when the equipment is not properly polar aligned. Polar alignment, or in the case with the O-GPS1 / Astrotracing in its various embedded implementations, that comes down to 1) the electronic compass and 2) how the elevation angle or angle of inclination are obtained.

In another thread - that I did not come across or read until today, Pete_XL identified the problem....

P.S. With O-GPS mounted I get about 10 times more trail-free exposure time than without it. Last week I bought a cheap small "Skywatcher Star Adventurer" tracking mount to get more precise tracking with long lenses. It seems to be about 10 times more precise than astrotracer if the polar alignement is made accurat.

If you look at the extremely high degree of precision that is necessary (but possible) to align this system to the earth rotation axis with an optical polarscope you understand that O-GPS1 really is limited. Because of the boundaries of the sensitivity of the integrated sensors and the measuring method itself it is impossible to get into near of this precision. Lack of precision in alignment results inevitably in startrails, which are nothing else but also "pixel shift".

• https://www.pentaxforums.com/forums/172-pentax-k-3/294121-questions-re-k-3-ii...ml#post3237771 post #50

Electronic compasses have a wide performance margin based on a number of factors, especially location on the earth. I cherry picked this passage from the following paper (section 3.3.3.2 Magnetic Dip Angle Effects on Accuracy)

At the magnetic equator, where the earth's field is primarily horizontal, each degree of tilt of the compass sensor has very little impact on the directional accuracy of the compass. In contrast, at high magnetic latitudes, each degree of tilt may be 3-5° of direction error or more, as shown in the following table.

• http://www.telemar.no/files/technical_manuals/kvh/compass/compasstech.pdf

The solution based on what is referred to as SWaP+C (Size, Weight and Power, plus Co$t) and is a better pointing and elevation measurement (compass and elevation detection). How large do you want the camera body, its weight and power requirements (battery size and longevity) along with cost. The mechanical trackers haven't really solved this either. The iOptron SkyTracker, Vixen Optics Polarie Star Tracker, etc., have worse user alignments and calibrations procedures. Pentax's Astrotracking is fairly user benign but still a sore point for a lot of folks. There is really no cheap inexpensive small fully automated easy to use mechanical tracker that does everything on its own. So, there is really no good solution here, at this time.

[jbinpg]

Pete was specfically mentioning the improvement with long lenses, not wides, when using a tracking mount. The wider the field of view, the more sidereal motion will enter the equation.

[interested_observer]

Pete was specfically mentioning the improvement with long lenses, not wides, when using a tracking mount. The wider the field of view, the more sidereal motion will enter the equation.

Longer focal lengths will certainly reduce the problem, however for me when looking at and reading about the tracking mounts - no one is having (at least reporting) the edge/corner tracking issues (star trailing) that the Astrotracker is experiencing. The tracking mounts with wide lenses bypass our problem here, and I'll add the caveat - when properly aligned. I don't dispute that greater sidereal motion will be encountered with wide angle lenses, however with the same lens focal length(s), the tracking mounts appear to not encounter this problem (to the extent that the Astrotracker does). That is the specific point I am attempting to identify.

[Gimbal]

It's the distortion of the wide rectilinear lenses that causes this problem when using the astro tracer, but not when using a tracking mount.

I'll bet that if you take a long exposure with a wide lens at the stars without using any tracking devises at all, that the stars at the corner of the frame will have moved (well, appeared to have moved) more than the stars in the middle of the lens. This is due to the distorsionen of the lens, everything is enlarged the further away from the center of the frame you get in order to keep lines straight, everything else on the other hand will get distorted.

The wider the lens, the more extreme the distortion gets. Take a really wide lens and people standing at the edge of the frame will look disfigured with one arm twice as big as the other arm. Therefore when the lens is still (for instance when using astro tracer) and the stars move across the frame, the stars at the edges will appear to move faster then the stars in the middle. There is no way the image sensor can compensate for this.

A tracking mount however will rotate the whole body with lens, and thus the stars stay in the same spot on the lens the whole time. The star in the corner stays in the corner.