[GeoJerry]

I tried using the astrotracer on my K-1 recently for the first time and got an unexpected result. In the first picture, the foreground is of course blurred in this 4 minute astrotracer exposure, but the stars in the center look pretty good. But on the edge of the picture, at upper left, you can clearly see well developed star trails. They're also present at upper right but the stars up there are too dim to see well. In the second picture, taken 10 minutes later, I used a 30 sec exposure. The star trails aren't really evident. The lens I'm using in both cases is the Samyang 14 mm prime.

So it appears that the astrotracer will distort star trails on the fringes of the field of view using wide angle lenses. Has anyone else seen this? At what focal length does the problem start to diminish? Or is 4 seconds just too long? Looks like it wasn't designed for this type of use.

Attached Images

View Picture EXIF

File Date

 

View Picture EXIF

File Date

 

[Medex]

If you expect to see dots of stars in the corner of the frame you will be always disappointed. Ultra wide angle lenses always make streaks instead of dots in the frame's corners because spherical picture should be projected on plain sensor. Even very highly corrected lenses do that.

[Kevin B123]

I have had that with a DA 21 on the K-3ii when the exposure is too long.

With that lens you could make a 30 second exposure without astrotracer and get a non-trailing result (500/14=35.7 sec).

Generally 4 minutes is too long, 2 minutes would be better, take several of them and stack. Take a dark frame too, ISO 1600 and wide open.

You were using f8 and that negates the light collecting value of the large aperture and ISO 200 but that would be fine for the first sharp foreground image focused on the building, with astrotracer off. After that you should go to infinity on the lens, astrotracer on, ISO1600 and wide open. Then take 5 or 6 2 minute exposures (check the first one is not trailing in the corners) .

Do a precision calibration before you start too.

Then it's PP time This will get you started:

You Tube

You Tube

Its a start, and more experienced members with the 14mm may offer better insights.

[MaineNative]

I have the K-1 and Samyang 14 and I see the same thing. I agree it is a distortion issue because the trails always radiate out from the lens center. If it were an Astrotracer issue, they should go in the same direction or rotate around a center point.

I live with the streaks in the corner because IMO they look better than coma “bat wing” artifacts and because the starts in the center look so darn good with the Samyang 14. (I usually shoot around 80-90 seconds at f3.5).

I have looked at some shots from the Irix 15 mm and you can see the same streaks in the corner.

I know some people stitch together multiple shots to get the full Milky Way. Using that method lets you just use the good center part of the frame, but I have not tried it myself.

[photoptimist]

Yes, astrotracer trailing with UWA lenses is an known problem caused by the discrepancy between the spherical coordinate motion of the sky and the central perspective rectilinear coordinate motion of the sensor.

The amount and pattern of trailing depends on the focal length and is very sensitive to the pointing direction of the lens. Pointing the lens toward the celestial equator has worse trailing than pointing it toward either of the celestial poles. Note that if the lens is pointed directly at the pole, you should get no corner trailing at all if the calibration is good.

If I've done my math right, the amount of this kind of trailing is inversely proportional to focal length. So if a 14mm lens has tolerable trailing at 30 seconds, a 28 mm lens will have tolerable trailing at 60 seconds.

[SteveinSLC]

4 minutes is too much for astrotracer...I haven't heard of anyone getting good results at that time length. 1-2 is more like it.

If you want to go longer, you'll probably need a tracking mount.

[DeadJohn]

That can happen even with good tracking and zero distortion. Stars follow curves of varying radii depending on how far they are from the poles. Stars further from the poles follow less curved paths. Those curves cause the distances between stars to change from our viewing perspective during a long exposure.

There's only so much the astrotracer can do to correct that.

This photo (not mine) is an example. The stars on the upper-left and lower-right are farther apart at the start of the trail, temporarily get closer, then move apart again. The effect is most noticeable with ultrawide lenses. The center stars in this photo move in a straight line because they are above the equator.

[awscreo]

I had the same experience. Also, I had the same "warped" star trails with UWA lenses as image above. I instead started stacking star exposures, cleaner results and much less corner trailing

[GeoJerry]

If you expect to see dots of stars in the corner of the frame you will be always disappointed. Ultra wide angle lenses always make streaks instead of dots in the frame's corners because spherical picture should be projected on plain sensor. Even very highly corrected lenses do that.

Yes that sounds right, thanks. Should have thought of that!

---------- Post added 09-09-18 at 04:43 PM ----------

I have had that with a DA 21 on the K-3ii when the exposure is too long.

With that lens you could make a 30 second exposure without astrotracer and get a non-trailing result (500/14=35.7 sec).

Generally 4 minutes is too long, 2 minutes would be better, take several of them and stack. Take a dark frame too, ISO 1600 and wide open.

You were using f8 and that negates the light collecting value of the large aperture and ISO 200 but that would be fine for the first sharp foreground image focused on the building, with astrotracer off. After that you should go to infinity on the lens, astrotracer on, ISO1600 and wide open. Then take 5 or 6 2 minute exposures (check the first one is not trailing in the corners) .

Do a precision calibration before you start too.

Then it's PP time This will get you started:
Sequator - FREE PC Milky Way Stacking software that reduces noise - YouTube

Its a start, and more experienced members with the 14mm may offer better insights.

Thanks Kevin, some great advice! I watched the video and downloaded Sequator, looking forward to giving it a whirl!

---------- Post added 09-09-18 at 04:45 PM ----------

I have the K-1 and Samyang 14 and I see the same thing. I agree it is a distortion issue because the trails always radiate out from the lens center. If it were an Astrotracer issue, they should go in the same direction or rotate around a center point.

I live with the streaks in the corner because IMO they look better than coma “bat wing” artifacts and because the starts in the center look so darn good with the Samyang 14. (I usually shoot around 80-90 seconds at f3.5).

I have looked at some shots from the Irix 15 mm and you can see the same streaks in the corner.

I know some people stitch together multiple shots to get the full Milky Way. Using that method lets you just use the good center part of the frame, but I have not tried it myself.

Thanks MaineNative! Do you still see the trails even with 80-90 second exposures?

---------- Post added 09-09-18 at 04:46 PM ----------

Yes, astrotracer trailing with UWA lenses is an known problem caused by the discrepancy between the spherical coordinate motion of the sky and the central perspective rectilinear coordinate motion of the sensor.

The amount and pattern of trailing depends on the focal length and is very sensitive to the pointing direction of the lens. Pointing the lens toward the celestial equator has worse trailing than pointing it toward either of the celestial poles. Note that if the lens is pointed directly at the pole, you should get no corner trailing at all if the calibration is good.

If I've done my math right, the amount of this kind of trailing is inversely proportional to focal length. So if a 14mm lens has tolerable trailing at 30 seconds, a 28 mm lens will have tolerable trailing at 60 seconds.

Thanks very much, photoptimist! It all makes very good sense.

---------- Post added 09-09-18 at 04:47 PM ----------

4 minutes is too much for astrotracer...I haven't heard of anyone getting good results at that time length. 1-2 is more like it.

If you want to go longer, you'll probably need a tracking mount.

Thanks Steve, others agree with you. Not sure why I chose 4 mins, really kind of arbitrary!

---------- Post added 09-09-18 at 04:48 PM ----------

That can happen even with good tracking and zero distortion. Stars follow curves of varying radii depending on how far they are from the poles. Stars further from the poles follow less curved paths. Those curves cause the distances between stars to change from our viewing perspective during a long exposure.

There's only so much the astrotracer can do to correct that.

This photo (not mine) is an example. The stars on the upper-left and lower-right are farther apart at the start of the trail, temporarily get closer, then move apart again. The effect is most noticeable with ultrawide lenses. The center stars in this photo move in a straight line because they are above the equator.

Great, thanks DeadJohn! I've seen pictures before like the one you attached but hadn't given them much thought (except maybe that it was some kind of PP trick)!

---------- Post added 09-09-18 at 04:49 PM ----------

I had the same experience. Also, I had the same "warped" star trails with UWA lenses as image above. I instead started stacking star exposures, cleaner results and much less corner trailing

Thanks! Not sure I understood... will stacking remove that UWA trailing effect? Or are you stacking using longer lenses?

[MaineNative]

At 80-90 seconds I still get the trails in the corners but the stars in the center are still very fine dots of light. I am usually shooting the Milky Way facing south with a view about 45 degrees above the horizon.

I always do a precise calibration once I get on location then check the first few shots for trails and recalibrate if I see trails in the center.

[awscreo]

Yes that sounds right, thanks. Should have thought of that!

---------- Post added 09-09-18 at 04:43 PM ----------


Thanks Kevin, some great advice! I watched the video and downloaded Sequator, looking forward to giving it a whirl!

---------- Post added 09-09-18 at 04:45 PM ----------



Thanks MaineNative! Do you still see the trails even with 80-90 second exposures?

---------- Post added 09-09-18 at 04:46 PM ----------



Thanks very much, photoptimist! It all makes very good sense.

---------- Post added 09-09-18 at 04:47 PM ----------



Thanks Steve, others agree with you. Not sure why I chose 4 mins, really kind of arbitrary!

---------- Post added 09-09-18 at 04:48 PM ----------



Great, thanks DeadJohn! I've seen pictures before like the one you attached but hadn't given them much thought (except maybe that it was some kind of PP trick)!

---------- Post added 09-09-18 at 04:49 PM ----------



Thanks! Not sure I understood... will stacking remove that UWA trailing effect? Or are you stacking using longer lenses?

I think Astrotracer makes the stretching worse, I don't see the effect that much in my normal exposures. I use AT with longer focal lengths, it works really well, but with UWA I just stack single exposures.

Examples (at similar focal lengths):

with AT

Winter stars by Timur Dzhambinov, on Flickr

single exposure:

Light painting under the stars by Timur Dzhambinov, on Flickr

stacked:

Milky Way over the old fort by Timur Dzhambinov, on Flickr

[DeadJohn]

I think Astrotracer makes the stretching worse, I don't see the effect that much in my normal exposures.

Agreed, especially if the astrotracer exposures are running for a longer period of time. (P.S. 3 nice photos)

For the benefit of others, view the full size version of his first photo (https://flic.kr/p/22voHRM). Once you're at Flickr click to zoom in. Center stars are sharp while stars at the corners appear to move in different directions. That's not distortion or tracking errors. The stars are actually changing their distance from the center of the photo during the exposure.

[bobphoenix]

First time K-1 Astrotracer user here!!!
OMG!!! Thanks for this post... I went out this evening with my K-1 set to ISO 3200 with Buld and the FA 28-105 that came with it a couple of weeks ago (This is the first clear sky night we've had since I got the K-1).
So, this is shot at 45 sec f5.0 with Astrotracer on. I was facing about due West with a tree line at my right (I was trying to capture my section of the Milky Way). The non distortion seems to be in the upper right quadrant of the image.
So, my question is... Is this as much at I should expect for this combination of settings with Astrotracer, or can I do better? (I do have an original SMC M-24 f2, I can use...)