[reivax]

I have a K-1 and I've been trying to push myself to do some more astrophotography and night shooting. It's probably the most difficult thing for me because of the time and patience required (I'm used to shooting sports and people), but I really enjoy it once I get my mind to settle down and focus. I'm hoping that you all can help me to better understand a couple of things:

1) Shooting astrophotography with longer focal lengths like 50mm and 85mm. Is it possible without heavy star trails?
I have three lenses that I normally use" Samyang 20 1.8, Samyang 24 1.4, and Pentax 24-70 2.8
I've found that the 20 1.8 has been my favorite so far, but I'd like to try longer focal lengths as well and that brings me to my next point.

2) Astrotracer. Will this work on longer focal lengths? Or are there focal lengths I should stick to?

Thanks! Any other information/guidance you can offer would be greatly appreciated.

[Adam]

I have a K-1 and I've been trying to push myself to do some more astrophotography and night shooting. It's probably the most difficult thing for me because of the time and patience required (I'm used to shooting sports and people), but I really enjoy it once I get my mind to settle down and focus. I'm hoping that you all can help me to better understand a couple of things:

1) Shooting astrophotography with longer focal lengths like 50mm and 85mm. Is it possible without heavy star trails?
I have three lenses that I normally use" Samyang 20 1.8, Samyang 24 1.4, and Pentax 24-70 2.8
I've found that the 20 1.8 has been my favorite so far, but I'd like to try longer focal lengths as well and that brings me to my next point.

2) Astrotracer. Will this work on longer focal lengths? Or are there focal lengths I should stick to?

Thanks! Any other information/guidance you can offer would be greatly appreciated.

The longer the focal length, the shorter the exposure can be before you'll see star trails. You can still use the astrotracer, just expect those trails to become visible a little bit earlier. It will definitely boil down to some trial and error, but I'd say with those focal lengths you should be able to get good 30-60 second exposures.

[photoptimist]

Yes, it will work with longer focal lengths but the maximum shutter time will be proportionally shorter.

[DeadJohn]

Up to 100mm is fairly easy with some practice. Do a precise calibration to improve tracking. Sample below. 60 seconds and I probably could have gone longer.

Going beyond 100mm, as much as 300mm, can work. It gets more finicky. Composition becomes more of a challenge if you don't know exactly where to aim.

[reivax]

Up to 100mm is fairly easy with some practice. Do a precise calibration to improve tracking. Sample below. 60 seconds and I probably could have gone longer.

Going beyond 100mm, as much as 300mm, can work. It gets more finicky. Composition becomes more of a challenge if you don't know exactly where to aim.

That's a pretty amazing picture! I'm assuming you went to a pretty dark place? I've been trying to do some night photography in the small city where I live, but I get so much light pollution because we have a major city right next to us. Closest place that's considered dark enough for astrophotography is about two hours away. Any tips on dealing with light pollution?

[DeadJohn]

If you can make out even a hint of the Milky Way, you have enough darkness for astrophotography.

My sample above was taken from outer suburbs of New York City. If you look at a color-coded light pollution map, it's a yellow zone. What city is near you?

Light pollution map

I always shoot DNG, not JPG. Typical processing is to boost white point, lower black point to extend dynamic range and improve contrast. Then I gently boost shadows and clarity.

Note that narrower lenses make it easier to process out light pollution. At 100mm, I am looking at a small enough slice of sky that it's even across the frame. At wide angles, the frame often includes horizon and near zenith, with light pollution gradients that become tougher to equalize.

[reivax]

If you can make out even a hint of the Milky Way, you have enough darkness for astrophotography.



My sample above was taken from outer suburbs of New York City. If you look at a color-coded light pollution map, it's a yellow zone. What city is near you?



Light pollution map



I always shoot DNG, not JPG. Typical processing is to boost white point, lower black point to extend dynamic range and improve contrast. Then I gently boost shadows and clarity.



Note that narrower lenses make it easier to process out light pollution. At 100mm, I am looking at a small enough slice of sky that it's even across the frame. At wide angles, the frame often includes horizon and near zenith, with light pollution gradients that become tougher to equalize.

Looks like I’m in a white/purpleish area. There’s a yellow area about an hour away. I think I read that the Milky Way isn’t a good direction to shoot at this time of year, but that Orión is supposed to be pretty good. Hopefully, I can free up some time pretty soon and find out.

[aaacb]

Below is a 60s astrotracer exposure with 100mm lens (macro wr) This was with a k3ii, not a k1, I don't have a sample without astrotracer, but before I calibrated the gps a second time star trails were very noticeable.

[DeadJohn]

...I think I read that the Milky Way isn’t a good direction to shoot at this time of year, but that Orión is supposed to be pretty good...

Yes, correct. I suggest downloading Stellarium. It's free planetarium software that shows what's up at any time. There's a plug-in (included with the download) that lets you specify a camera sensor and focal length to simulate framing.

[AstroDave]

2) Astrotracer. Will this work on longer focal lengths? Or are there focal lengths I should stick to?

Here's an astronomer's take on this:

If the system is properly calibrated (a big IF, perhaps), the focal length should not matter - depending on how astrotracer really works.

Why do we need astrotracer (or any other tracking system)?

Well, as you all know the earth rotates, so the stars seem to move. But things are a bit more complicated.

How can we compensate for the rotation? A simple attempt involves moving the camera (or, equivalently, just the sensor) along the apparent motion of the sky. However, unless the compensating motion is around the rotation axis of the earth, the field of view will also rotate. This rotation occurs because the stars actually move at a different linear rate across the field of view because they are moving on circular paths with different effective radii. Perhaps this can be best envisioned by thinking of a star at the celestial equator and one at the celestial north pole (i.e. Polaris , more or less). If you tried to take a picture encompassing both stars (a moderately wide angle lens could easily do this), the star at the equator would move during the exposure, while Polaris would stand still. If you tried to follow the equatorial star by moving your camera along with the star, then Polaris would get streaked out.

A normal equatorial tracking system does exactly the compensating motion you need, so no field rotation, and clearly, any focal length lens is OK.

Since your camera is probably mounted on a tripod instead of a tracking mount, things get a bit more complicated. The simplest compensation is to move the camera/sensor left-right/up-down. This allows you to track the center of the field, but does not correct for field rotation.

The amount of field rotation is proportional to distance from the center of the field as the center is tracked. This is an important fact. It means that the effect on a photograph will have the same fractional effect independent of focal length! This is because the field of view depends on focal length in exactly the same way. Alternatively, think of the longer focal length image as a crop of the central part of a wide angle image - blown up to the size of the original. Any rotational amount will be the same angular amount about the center in any crop.

So, if the astrotracer applies the necessary field rotation*, you should be fine with ANY FOCAL LENGTH LENS. And, because of the scaling effect noted above, exposure lengths can be the same for all lenses, as well.

* while I could not find any definitive statement about astrotracer rotation, the K1 is able to rotate the sensor. Camera specs quote for horizon correction: “correction up to 2 degrees.” (P. 117 of manual)

For some more info on field rotation, see RASC Calgary Centre - Field Rotation

All modern large (mirrors more than a few meters) professional optical telescopes use az/el mounts (like your tripod). They all have cameras which apply the necessary field rotation. So this really can work for any size optical systems - if it’s done right!

[photoptimist]

Here's an astronomer's take on this:

If the system is properly calibrated (a big IF, perhaps), the focal length should not matter - depending on how astrotracer really works.

A camera designer's take on this:

Astrotracer does rotate the sensor to compensate for the rotational motion of the Earth.

Unlike the typical tracking mount, Astrotracer has a very limited range of motion so it can only compensate for a short duration of motion. Also, unlike the typical tracking mount, Astrotracer's range is measured in pixels, not angular degrees. Because the rate of motion of the stars in pixels per minute depends the declination (stars at 0° = the celestial equator move fastest and stars at ±90° = the poles don't move at all), astrotracer runs out of tracking room sooner the closer the celestial equator is to any part of the image. And because the rate of motion of the stars in pixels per minute depends on focal length (the stars move faster in a telephoto image than in a wide-angle image) astrotracer runs out of tracking room sooner with longer focal length lenses.

Unlike a tracking mount which moves the camera in spherical sky coordinates, Astrotracer moves the sensor in rectilinear image coordinates. For long focal length lenses, rectilinear can be a very close approximation to spherical. But the approximation gets worse for shorter focal lengths. The result is that Astrotracer with ultrawide angle lenses suffers from some uncompensated star trailing in the corners that looks really strange because it looks like the star field is being stretched in strange directions.

[AstroDave]

Unlike the typical tracking mount, Astrotracer has a very limited range of motion so it can only compensate for a short duration of motion. Also, unlike the typical tracking mount, Astrotracer's range is measured in pixels, not angular degrees. Because the rate of motion of the stars in pixels per minute depends the declination (stars at 0° = the celestial equator move fastest and stars at ±90° = the poles don't move at all), astrotracer runs out of tracking room sooner the closer the celestial equator is to any part of the image. And because the rate of motion of the stars in pixels per minute depends on focal length (the stars move faster in a telephoto image than in a wide-angle image) astrotracer runs out of tracking room sooner with longer focal length lenses.

No!

As I pointed out, it’s the amount/rate of field rotation that counts. For a given position in the sky and your location, there is a unique field rotation rate (i.e. 1 degree in 5 minutes, perhaps). Since the camera can rotate the sensor, that is all that it needs to do (well, it does have to translate as well). How much it can rotate the sensor and how far it can move it determine how long an exposure can be for that particular circumstance.

It is immaterial whether your lens is long or short focal length (except as you point out in your last paragraph). If astrotracer is doing its job properly, a star image should remain on a single pixel during the exposure, again independent of focal length.

The field rotation during an exposure is different from the sky rotation due to the Earth. The web site I cite has a lengthy discussion of this.

It can be much different - consider a substantial example: consider the point directly overhead (the zenith). If you are trying to observe an object that passes through your zenith, your camera has to do an instantaneous 180 degree rotation at the instant of transit (the moment when a star goes directly overhead)!!!!! Prior to transit, the star is rising (i.e. your camera would be elevating to follow it), after transit, the star is going down - your camera now must start moving down as well - the tracking direction has gone from elevating to descending in an instant.

If you happened to be at the North Pole, however, no problem - your camera would be aimed straight up, but would need to rotate the field at the sidereal rate only - not very fast at all: 1 degree in 4 minutes.

Hence, you can’t simply say (assuming you are at more normal latitudes) that the field rotation is worse at the equator, and zero at the pole. In fact, as noted on the cited web site, the least field rotation occurs when observing more or less due east or west for objects near the horizon (and worst near the zenith, as discussed above).

You are quite correct about very wide angle lenses, basically because you can not map the spherical sky onto a flat surface (your camera sensor). Another way to think of this is that the scale (arcseconds per pixel) changes as you go out radially from the center of the image frame. This is a classic problem for map makers! The wide angle lens result is similar to the distortion in a Mercator map projection - where Greenland (and other land masses near the poles) are exaggerated in apparent size.

[photoptimist]

No!

As I pointed out, it’s the amount/rate of field rotation that counts. For a given position in the sky and your location, there is a unique field rotation rate (i.e. 1 degree in 5 minutes, perhaps). Since the camera can rotate the sensor, that is all that it needs to do (well, it does have to translate as well). How much it can rotate the sensor and how far it can move it determine how long an exposure can be for that particular circumstance.

It is immaterial whether your lens is long or short focal length (except as you point out in your last paragraph). If astrotracer is doing its job properly, a star image should remain on a single pixel during the exposure, again independent of focal length.

The field rotation during an exposure is different from the sky rotation due to the Earth. The web site I cite has a lengthy discussion of this.

It can be much different - consider a substantial example: consider the point directly overhead (the zenith). If you are trying to observe an object that passes through your zenith, your camera has to do an instantaneous 180 degree rotation at the instant of transit (the moment when a star goes directly overhead)!!!!! Prior to transit, the star is rising (i.e. your camera would be elevating to follow it), after transit, the star is going down - your camera now must start moving down as well - the tracking direction has gone from elevating to descending in an instant.

If you happened to be at the North Pole, however, no problem - your camera would be aimed straight up, but would need to rotate the field at the sidereal rate only - not very fast at all: 1 degree in 4 minutes.

Hence, you can’t simply say (assuming you are at more normal latitudes) that the field rotation is worse at the equator, and zero at the pole. In fact, as noted on the cited web site, the least field rotation occurs when observing more or less due east or west for objects near the horizon (and worst near the zenith, as discussed above).

Hmmm... We seem to have a bit of disagreement over some aspects of this...

The pattern of motion of the stars in an image is a combination of rotation, translation, and distortion caused by viewing spherical motion with a rectilinear lens. The nature of the motion and sensitivity to focal length depends on the pointing direction of the camera.

If you attach a camera to a tripod and point it directly toward the polestar (e.g., declination of +90° if you are in the Northern Hemisphere), the all the stars in the image appear to rotate around the exact center of the image. The rotation rate is 360.9856° per solar day or about 1° every 4 minutes. In this situation, you are right that the focal length does not matter. Astrotracer will rotate the sensor to track the rotation of the star field.

But if you point the same camera-on-tripod at the celestial equator (i.e., declination of +0°), then all the stars appear to move in nearly straight lines across the image with no rotation in the image at all. In spherical coordinate terms, the equatorial stars are still moving at the same 1° every 4 minutes created by the Earth's rotation. But in rectilinear image coordinate terms on the sensor, the image of the equatorial stars will be translating across the sensor at about 4.4 microns per minute per mm of focal length. The longer the focal length, the faster the apparent motion in the image, and the sooner that astrotracer runs out of room.

Pointing the camera somewhere between the celestial pole and equator creates an imaging geometry that has some combination of the rotation and translation. The closer you point to the pole, the more the image motion will be just rotation and the focal length will matter less. The closer you point toward the celestial equator, the more the image will include translation and the focal length will matter more.

If you don't believe me that focal length affects the speed of star motion in astrophotography images, then ask yourself why one of the most basic rules of astrophotography is the "500 rule" (How to Avoid Star Trails by Following the '500 Rule') that states that for a fixed-mount camera, the longest shutter time that avoids visible star trails is 500 divided by focal length. The stars generally move much faster when using higher magnification.


P.S. Ricoh has a nice table of approximate Astrotracer tracking times (Specifications | GPS UNIT O-GPS1 | RICOH IMAGING) for the O-GPS-1 that shows both the effects on declination and focal length for different cameras. Similar (but "tentative") tracking time numbers for the K-1 are found on Ricoh's K-1 FAQ (PENTAX K-1 | FAQ | Support | RICOH IMAGING).

[Gimbal]

I guess the field rotation AstroDave is referring to is what is left if you have an alt-azimuth goto-mount that will track the center star perfectly for hours.

[photoptimist]

I guess the field rotation AstroDave is referring to is what is left if you have an alt-azimuth goto-mount that will track the center star perfectly for hours.

I bet you're right. That also explains his comments about tracking at the zenith which can be a serious pain for a telescope mount.