Another first moon shot. (Tamron 180mm+2X TC)

Bill Boehme wrote in post #10259367
There are times when stacking moon images is useful, as you said. Rather than looking at stacking lunar images as a technique for noise reduction, view it instead as a tool for refining edge details. Most of the time, people only consider mean value stacking, but that will tend to soften edges at the expense of reducing noise. I use median value stacking on moon images. It is not very effective at noise reduction, but its main value lies in enhancing edges. The median filtering can result in some image distortion if the quality of the images being used are not especially good.

Just so users of astrophotography-specific software don't get confused (I did when I first read this), I believe Bill is talking about the "Focus Stacking" feature in Photoshop CS3 and later, in which the user can select a median blending mode. Registax doesn't allow the user to select the blending mode for stacking. DeepSkyStacker does, but it can't stack lunar images. I use Iris software for astrophotography, and it will stack lunar images, and allows the user to select from several different blending modes, including Median and Mean. Median usually produces better results for lunar photos, but due to better noise rejection, not edge enhancement. My Photoshop is CS2, which doesn't have the Focus Stacking feature, so I will take Bill's word that it helps with edge enhancement.

Bill Boehme wrote in post #10259367
One other point when using a good solid tripod is that I would recommend keeping ISO as low as possible and let the shutter speed get as low as 1/50 second before increasing the ISO to a higher value. With my camera and lens (Canon 7D and 400/5.6 lens plus 2X and 1.4X TC for a FL of 1120 mm), the moon which has an angular diameter of 0.5 degrees represents a 2300 pixel width and height on the camera's sensor. With an approximate lunar rate of 0.0042 degrees/second, it would require a shutter speed slower than 1/20 second for motion blur to equal one pixel. (The angular resolution of the sensor with a 1120 mm FL lens is approximately 0.0002 degrees/pixel. At 1/80 second shutter speed which i used for my images yesterday, the moon traverses 0.0000525 degrees -- much less than the resolution of the sensor and probably somewhat less than the resolution of the lens/TC combination used.

Good advice if you have a rock solid tripod. I do, but most often when I'm shooting with only a camera lens I use a less than rock-solid but much easier to handle tripod. At long focal lengths, any movement of the camera, including that caused by a moderate breeze, will blur the image. If this is your situation too, decreasing the exposure time to a value that would be appropriate for hand-held imaging can save the results. We're not concerned about motion blur caused by the earth's rotation at sub-second exposures so much as motion blur caused by camera movement.

Bill Boehme wrote in post #10259367
That definitely can be an issue if one wastes too much time capturing the images and does not do some planning ahead. However, by planning the time frame when the moon is near its highest elevation (essentially due south if you are in the northern hemisphere) and setting up the tripod mount so that most of the rotation will be in azimuth, a conventional tripod mount can be used with very little field rotation if all of the images are captured in just a few minutes.

Actually field rotation of a given object has the maximum rate when the object is at it's culmination (due north or south and at it's highest elevation). And field rotation always causes the field to rotate about the center of the field, so I'm not sure what you mean by "setting up the tripod mount so that most of the rotation will be in azimuth". Points on opposite sides of the center of the field move in opposite directions.

Don

DonR wrote in post #10270757
Just so users of astrophotography-specific software don't get confused (I did when I first read this), I believe Bill is talking about the "Focus Stacking" feature in Photoshop CS3 and later, in which the user can select a median blending mode....

If you are not talking about the type of focus stacking done in macro photography then I think that you would be correct. Also, I believe that you are right about noise rejection when using median stacking. Mean stacking smooths over details probably because of a fairly wide standard deviation in the data when using a large number of images. I think that maybe skewness of data accounts for much of the difference in results between mean and median. I played with a different technique last night using mean value stacking and then created a new layer with median value stacking and an edge mask. The result is that it seems to produce comparable results to median stacking for cleaner edge details along with what seems like less noise in the relatively featureless areas. The downside of this is that some fine detail is lost in the relatively featureless areas. I also believe that I unintentionally used luminosity blending tuned for emphasizing midtone contrast edges. I don't think that was the best thing to do in this case. Probably better original images would be the best thing.

DonR wrote in post #10270757
Good advice if you have a rock solid tripod. I do, but most often when I'm shooting with only a camera lens I use a less than rock-solid but much easier to handle tripod. At long focal lengths, any movement of the camera, including that caused by a moderate breeze, will blur the image. If this is your situation too, decreasing the exposure time to a value that would be appropriate for hand-held imaging can save the results. We're not concerned about motion blur caused by the earth's rotation at sub-second exposures so much as motion blur caused by camera movement.

Very good informtion. My good tripod had a mechanical failure due to a casting flaw and I used a lightweight tripod for several months. Just breathing on a lightweight camera tripod can produce objectionable motion blur ... not to mention that the vibration from the camera's shutter is enough to cause significant problems even though the mirror is locked up. Even a wired shutter release can be problematic.

DonR wrote in post #10270757
Actually field rotation of a given object has the maximum rate when the object is at it's culmination (due north or south and at it's highest elevation). And field rotation always causes the field to rotate about the center of the field, so I'm not sure what you mean by "setting up the tripod mount so that most of the rotation will be in azimuth". Points on opposite sides of the center of the field move in opposite directions.

Strictly from a mathematical point of view, a situation with increasing rate and sign change at culmination could not be the case since that would require the existence of a singularity exactly at culmination which means that the rate would be undefined at that point and hence would not be differentiable.

Tracking a full moon near the summer solstice when it is close to its highest point in the southern sky using only the azimuth axis of a camera tripod head is relatively easy to do when taking a sequence of images. Both az and el can remain fixed for several shots -- then slew the az and repeat focus if the previous focus doesn't seem sharp enough and then shoot another series. The el adjustment only needs occasional tweaking.

The situation would be much different if shooting something due north at a similar elevation angle.

Bill Boehme wrote in post #10284897
Strictly from a mathematical point of view, a situation with increasing rate and sign change at culmination could not be the case since that would require the existence of a singularity exactly at culmination which means that the rate would be undefined at that point and hence would not be differentiable.

I guess I didn't describe it very well, Bill. The direction of field rotation doesn't change at culmination, it continues in the same direction but the rate, having reached its maximum at culmination, begins decreasing. The direction of field rotation changes when the object crosses points due east and due west of the observer, at which times the rate of field rotation is momentarily zero. Note that for objects in the southern sky, these points of zero field rotation may occur below the horizon. Objects that do not cross between the northern and southern skies do not have a point of zero field rotation. Their field rotation rate changes with their altitude, but never stops and never reverses direction.

There is, in fact, a singularity, which occurs only for objects that cross the zenith at culmination. Note that most objects never do this, and the moon only does it twice a year at any given location, and never at locations outside of the tropics. The rate of field rotation momentarily approaches infinity at that point, i.e., cannot be calculated. If you have ever tried to track an object accurately through the true zenith with an ALT/AZ mount, you will understand this, as it is physically impossible. At the exact moment the object crosses the zenith, the azimuth changes from 90 degrees to 270 degrees, and it is not possible to rotate the mount through 180 degrees of azimuth instantaneously.

Here is a site that describes the mathematics of field rotation very well, and with lots of diagrams - a picture is worth a thousand words:

http://calgary.rasc.ca​/field_rotation.htm

Bill Boehme wrote in post #10284897
Tracking a full moon near the summer solstice when it is close to its highest point in the southern sky using only the azimuth axis of a camera tripod head is relatively easy to do when taking a sequence of images. Both az and el can remain fixed for several shots -- then slew the az and repeat focus if the previous focus doesn't seem sharp enough and then shoot another series. The el adjustment only needs occasional tweaking.

Actually, in the northern hemisphere the full moon's elevation at its culmination closest to the summer solstice is at its lowest point of the year. On the summer solstice the sun reaches its highest elevation of the year for those of us in the northern hemisphere, but the full moon, being on the opposite side of the earth from the sun, is at its lowest elevation - specifically, at approximately (90 minus observer's latitude) minus 23.5 degrees. The full moon reaches its highest latitude of the year in the northern hemisphere at its culmination closest to the winter solstice, specifically (90 minus observer's latitude) plus 23.5 degrees.

Bill Boehme wrote in post #10284897
The situation would be much different if shooting something due north at a similar elevation angle.

The rate of field rotation for an object due north is of the same magnitude and opposite direction of the rate of field rotation of an object due south at the same elevation.

Don

I am aware of the singularity at the zenith since my career before retiring was in guidance, navigation, and control of space vehicles. Dealing with singularities is easily solved in GN&C by changing coordinate systems to a different frame of reference.

I guess that it may have sounded like I was saying that the moon was the highest point of the year at the summer solstice, but I meant that when the moon is at it highest elevation (the culmination) near the summer solstice, it is fairly easy to track with an az-el gimbal head oriented to local level (i.e., mounted on a camera tripod) because it is low in the sky, but I had a senior moment and didn't mention the "why" part. The reason "why" of course is the low elevation of the moon as you stated.

The problem with looking due north at about the same elevation as the moon (about 34°) was related primarily to using a camera tripod with an az-el gimbal head rather than a telescope mount oriented to the earth's rotation axis. For the example mentioned, it would put me looking almost directly at Polaris. If there were an extended object there, it would be impossible to maintain a fixed orientation of the object or track it with one axis only with a camera tripod.

Bill Boehme wrote in post #10286359
The problem with looking due north at about the same elevation as the moon (about 34°) was related primarily to using a camera tripod with an az-el gimbal head rather than a telescope mount oriented to the earth's rotation axis. For the example mentioned, it would put me looking almost directly at Polaris. If there were an extended object there, it would be impossible to maintain a fixed orientation of the object or track it with one axis only with a camera tripod.

True, but nevertheless the rate of field rotation when the camera is pointed at Polaris is exactly the same as when it is pointed due south at the same altitude as Polaris. So in a sense, it is easier to "track" an object centered on the celestial pole with an ALT-AZ mount - no movement of the mount is necessary, and the effect of field rotation is the same.

This fact can give some perspective about field rotation to people who have viewed and possibly even created star trail photos by pointing the camera near Polaris and taking photos over a period of time. No matter where you point the camera, you will experience field rotation with an Alt-Az mount tracking the subject. The rate of rotation is greatest for a given altitude when the camera is pointed due north or south, ant the rate decreases to zero momentarily and the direction reverses as the object crosses due east or west. The field rotation rate increases for a given azimuth as the altitude increases, except for the rare case of an object tracking along a line from due east to due west through the zenith. In this special case, the rate of field rotation is zero at all times except for the zenith crossing, at which point it is momentarily infinitely high.

Don

DonR wrote in post #10287960
...... The rate of rotation is greatest for a given altitude when the camera is pointed due north or south.....

If viewed from the perspective of a photographer located near mid northern latitude and using an ordinary camera tripod and thinking of orientation in terms of local level and wanting to shoot star trails, field rotation would not look the same if his camera was pointed due south as it would when pointed to the same elevation due north.

Bill Boehme wrote in post #10288484
If viewed from the perspective of a photographer located near mid northern latitude and using an ordinary camera tripod and thinking of orientation in terms of local level and wanting to shoot star trails, field rotation would not look the same if his camera was pointed due south as it would when pointed to the same elevation due north.

Actually it would look very nearly the same if the camera were centered on and TRACKING a star in ALT-AZ mode, if the star's altitude above the southern horizon were approximately the same as the altitude of Polaris above the northern horizon. The star at the center of the field would remain at the center of the field and all other stars would appear to rotate around it, with a radius equal to their distance from the center star, and at the same rate that stars equally distant from Polaris appear to rotate around Polaris.

The only difference is that the center of this field in the southern sky would only remain due south of the observer momentarily, before and after which the rate of field rotation would be lower.

You were probably thinking of the situation where the camera position is fixed on a point due south at the same elevation as Polaris. The star trails in this case would appear completely different, of course, but the term "field rotation" doesn't strictly apply to this situation. Field rotation occurs when an ALT-AZ mount or camera tripod TRACKS an object, not when the mount or tripod is stationary. In the case of Polaris, it makes very little difference whether the mount is tracking or stationary, since Polaris is always less than one degree from the stationary north celestial pole. So the circular star trails seen when a camera is centered on Polaris very closely resemble the field rotation seen when a camera tracks a celestial object in ALT-AZ mode.

Don

DonR wrote in post #10289034
You were probably thinking of the situation where the camera position is fixed on a point due south at the same elevation as Polaris.

Yes, that is what I was talking about since most photographers who are just casually into astronomy would not track anything other than manually keeping the moon centered in the field of view.

Bill Boehme wrote in post #10289233
Yes, that is what I was talking about since most photographers who are just casually into astronomy would not track anything other than manually keeping the moon centered in the field of view.

True, but even keeping the moon centered in the field of view from shot to shot amounts to tracking, and tracking the moon in ALT-AZ mode, e.g., with a normal camera tripod, causes field rotation.

In sub-second exposures you don't need to worry about field rotation blurring individual frames. But in a series of exposures spanning a few minutes, if you re-center the moon before each exposure, the moon will appear to rotate in the field from frame to frame, and the rotation can easily amount to over one degree in as little as four minutes. On a given night, imaging the moon near culmination, at its highest altitude due south (or due north in the southern hemisphere), guarantees the worst field rotation case.

If you don't compensate for the rotation when stacking multiple exposures, blurring can result. This was the point of my second post in this thread. Software designed for stacking lunar images, like Registax, should de-rotate the frames automatically, but other software you might use for stacking, like Photoshop, will not automatically de-rotate. De-rotating can be done manually in Photoshop, and though it's not difficult, it probably qualifies as an "advanced" technique with which many Photoshop users are not familiar. If your goal is a tack-sharp, crisp image of the moon showing the most possible detail at your focal length, you need to be aware of this.

Don