Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
04-28-2015, 07:26 PM - 1 Like   #46
Pentaxian
Sagitta's Avatar

Join Date: Feb 2011
Location: Maine
Photos: Gallery
Posts: 2,939
QuoteOriginally posted by jatrax Quote
I'm a little confused by this, can you explain why this is causes problems? I thought only the sensor moved, not the lens. What am I missing?
As far as I know, nothing.

I use my 10-20 with the tracker all the time. What you DO get at the 10mm end is trailing along the corners, since you're only tracking the stars in the center of the image, and as such the edges move a bit more than the center will due to the sheer amount of sky you're shooting. It has nothing to do with the weight of the lens unless you have a bad tripod mount and your lens sags over the course of the shot.

This is 29 seconds @ 16mm - if you pixel peep you'll notice that the stars trail in from each corner towards the center of the shot.



EDIT: Looking at it, I think this was a static shot (no tracking involved). If you were to use the tracker, the trailing could possibly be even more pronounced at the edges depending on what part of the sky you're shooting.

EDIT II: Here's one I know used the tracker. You can see the trailing here as well. This was 10mm / 90 seconds.



04-28-2015, 10:46 PM   #47
Site Supporter
K David's Avatar

Join Date: Oct 2012
Location: SF Bay Area
Photos: Gallery
Posts: 1,312
QuoteOriginally posted by jatrax Quote
I'm a little confused by this, can you explain why this is causes problems? I thought only the sensor moved, not the lens. What am I missing?
Center of mass, in a nutshell. With a small lens, the weight of the lens-camera combo is near or within the camera's body. With a longer or heavier lens it is further forward, maybe into the lens. If your central mass is over the tripod's center point or near it, then you have a more stable support. If the central mass is far forward and away from the tripod's center point, the system becomes unstable.

QuoteOriginally posted by DeadJohn Quote
If a heavy lens is causing problems, you might need a heavier duty tripod head.
Exactly right. And maybe a more stable tripod. I have an old surveyor's tripod I picked up just for this purpose but haven't finished fixing it yet.
04-28-2015, 11:24 PM   #48
Site Supporter
Pete_XL's Avatar

Join Date: Mar 2010
Location: Germany
Photos: Gallery
Posts: 207
QuoteOriginally posted by Ian Stuart Forsyth Quote
What you are seeing here is the accuracy at which the receiver is calculating the global position of your setup. With this kind of GPS the accuracy that we would see is in the ball park of 10m.


Ian, you are right, the accuracy of the GPS is within an acceptable range and not the problem in my point of view. More important for the calculations of the rotation than the exact coordinates is the azimutal angle and the inklination of the optical axis of the lens. The crucial negative factor with O-GPS1 is the fluctuation of the azimutal angle that can be seen in the picture. And that is what I wanted to show with that example.


O-GPS determins the azimutal angle by measuring the earth magnetic field. Obviously this not precise enough and/or varies in time. The inclination is determined by acellerator sensors in the camera and should also be quite unprecise in comparison with the accuracy that is normaly aspired in polar alignment when shooting astrophotos.


O-GPS1 is fine for easily tracking stars especially with short lenses. But if it comes to long lenses and/or long exposure times it reaches ist limits. I have done thousands of frames in the night sky with astrotracer and like it a lot. But this thread is about the new pixel shift technology that needs a precision of a few (one?) pixel to work. I am certain that astrotracer and pixel shift can not go together.
04-28-2015, 11:38 PM   #49
Site Supporter
K David's Avatar

Join Date: Oct 2012
Location: SF Bay Area
Photos: Gallery
Posts: 1,312
QuoteOriginally posted by Pete_XL Quote
Ian, you are right, the accuracy of the GPS is within an acceptable range and not the problem in my point of view. More important for the calculations of the rotation than the exact coordinates is the azimutal angle and the inklination of the optical axis of the lens. The crucial negative factor with O-GPS1 is the fluctuation of the azimutal angle that can be seen in the picture. And that is what I wanted to show with that example.
I found that, with my first shots, the O-GPS was basically not effective at all. So I re-did the precise calibration as carefully as I could. That made a HUGE difference and I was able to get shots to 3 minutes without trails.

04-29-2015, 12:27 AM   #50
Site Supporter
Pete_XL's Avatar

Join Date: Mar 2010
Location: Germany
Photos: Gallery
Posts: 207
QuoteOriginally posted by K David Quote
I found that, with my first shots, the O-GPS was basically not effective at all. So I re-did the precise calibration as carefully as I could. That made a HUGE difference and I was able to get shots to 3 minutes without trails.


I also did precise calibration often, carefully and repeatedly. Sometimes it helped and sometimes not. That is the variation-thing. Do the calibration, orientate the cam to a point in the sky and do a longer series of astrotracer shoots without changing orientation at all. Then look at the results e. g. with the freeware Geosetter. You see a variation in azimut angle (at least with my O-GPS1). This is where the problem lies when comming to long lenses and long exposure.


If I am right the variation of measured azimut should also appear if O-GPS1 is mounted, calibrated and she shoots are then made in interval mode. I will try that next days and give a report.


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".
04-29-2015, 12:40 AM   #51
Veteran Member
bossa's Avatar

Join Date: Jul 2011
Location: Adelaide
Posts: 4,546
Pixel Shifting with Astrophotography would be no different from the everyday application apart from the exposure time. It is effectively just in camera stacking and should work.
04-29-2015, 01:15 AM   #52
Pentaxian




Join Date: Apr 2007
Location: Sweden
Posts: 1,940
One problem with wide rectilinear lenses is that they distort the image in order to keep lines straight. The further away from the center the larger is the distortion. This means that stars that travels from the edge towards the center will appear to move faster at the edge, slow down in the center and then accelerate again as they move out of the frame.

And since the lens doesn’t move when using the astro tracer the stars will appear to change speed as they move across the sensor. The astro tracer can obviously not compensate for this and thus one ends up with prolonged stars when using wide rectilinear lenses and long exposures.
04-29-2015, 04:10 AM   #53
Site Supporter
Pete_XL's Avatar

Join Date: Mar 2010
Location: Germany
Photos: Gallery
Posts: 207
QuoteOriginally posted by bossa Quote
Pixel Shifting with Astrophotography would be no different from the everyday application apart from the exposure time. It is effectively just in camera stacking and should work.


As I understand it pixel shifting is absolutely not comparable with usual stacking, no matter if inside or outside of a camera. If it were just in camera stacking we then would not have a new technology because this is "old stuff" (look e.g. at HDR-functions in cameras).


In usual stacking the RGB values of Pixels in aligned photos are just recalculated with statistical methods e.g. applying median, sigma clipping, average and other functions at each "pixel coordinate" in the picture. The main effect of stacking in astrophotography is that by this process the erratic "noise" is statistically reduced and the static "signal" (stars etc.) is relatively amplified. Stacking does not result in more sharpness or resolution besides the effects of a better SNR (signal/noise ratio).


On the other hand pixel shift is intended to result in more resolution. For that pixel shift method is different. It is no "in camera stacking". The idea in pixel shift technology is to compensate the resolution decreasing fact that each pixel is only capable to catch only one color (R, G or B). This is done by shifting the individual pixels with decent sensor movement to neighbour positions and determin thereby the color value at points, where they normally could not be determined because of a static position of the one-color-pixels. This additional information is then put together to a picture containing the additional color information, which in fact means higher resolution.


Because this technique requires an absolutely static non moving photo motiv a stable tripod is necessary in normal photography with pixel shift. In astrophotograpy there is no static motif at all, every sky object is "moving". Astrophotos are always long exposures (besides sun, moon, bright planets to shoot). To have pixels aligned with a pixels-scale-error-tolerance in a long exposure you need professional tracking mounts with autoguiding at ist best. Comparing this with astrotracer is like comparing a VW-Beetle with a truck. You can drive the Beetle with a lot of fun but will never get more than a crate of beer into the luggage space. Leaving this picture I think the accuracy of the beetle-astrotracer should be not adequate to support pixelshift in astrophotography.

04-29-2015, 07:43 AM   #54
Loyal Site Supporter




Join Date: Apr 2010
Location: Tennessee
Posts: 5,794
QuoteOriginally posted by Pete_XL Quote
Because this technique requires an absolutely static non moving photo motiv a stable tripod is necessary in normal photography with pixel shift. In astrophotograpy there is no static motif at all, every sky object is "moving". Astrophotos are always long exposures (besides sun, moon, bright planets to shoot). To have pixels aligned with a pixels-scale-error-tolerance in a long exposure you need professional tracking mounts with autoguiding at ist best. Comparing this with astrotracer is like comparing a VW-Beetle with a truck. You can drive the Beetle with a lot of fun but will never get more than a crate of beer into the luggage space. Leaving this picture I think the accuracy of the beetle-astrotracer should be not adequate to support pixelshift in astrophotography.

GPS tells the camera its location, orientation, altitude, and the camera can already detect pitch. Yes, the stars are moving, but the SR should be able to account for the very slight movement and calculate for that when using the pixel shift. The camera could take exposures at 1/15th of a second for 30 seconds while using the SR to adjust for movement and stack/average the image exposures.


The technology is there to integrate GPS, Astro-tracker, and pixel shift technology. I think its a question of accuracy.
04-29-2015, 10:55 AM   #55
Site Supporter




Join Date: Dec 2011
Photos: Gallery
Posts: 82
I have been using the O-GPS1 since it first was available. After it is calibrated and mounted on a firm tripod, the camera body should set be parallel to the celestial equator. I believe the sensor follows the earths motion from east to west only in sidereal time. and not north and south, A ultra wide angle lens at 14mm takes larger apparent angle of the sky then a 300mm lens. On wide angle lens the stars on the four corners will seem to move faster then the stars focused at the center due to the earths curvature. This is less of a factor on a 300mm lens with and smaller apparent sky angle. When you look at a photo of star trails taken over a long period of time you can seen the apparent distance traveled near the poles is much smaller then at the equator.
When you input the focal length of the lens you are using the camera then knows the angle of the sky that is covered and then calculates the time rate needed to move the sensors over that part of the sky for the latitude you are located to capture pinpoint stars. The O-GPS1 is working hard. When you set up a telescope for astro photography your equatorial mount must be polar aligned before you can get pinpoint star, By pointing your camera parallel to celestial rather then an angle will help.
Hank
04-29-2015, 12:48 PM   #56
Pentaxian
philbaum's Avatar

Join Date: Oct 2007
Location: Port Townsend, Washington State, USA
Posts: 3,659
QuoteOriginally posted by Sagitta Quote
As far as I know, nothing.

I use my 10-20 with the tracker all the time. What you DO get at the 10mm end is trailing along the corners, since you're only tracking the stars in the center of the image, and as such the edges move a bit more than the center will due to the sheer amount of sky you're shooting. It has nothing to do with the weight of the lens unless you have a bad tripod mount and your lens sags over the course of the shot.

This is 29 seconds @ 16mm - if you pixel peep you'll notice that the stars trail in from each corner towards the center of the shot.



EDIT: Looking at it, I think this was a static shot (no tracking involved). If you were to use the tracker, the trailing could possibly be even more pronounced at the edges depending on what part of the sky you're shooting.
Good shots. The interesting thing in the first shot is that i see the worst trailing (not much of it but there) in the right side of the picture more than the left.

QuoteOriginally posted by Gimbal Quote
One problem with wide rectilinear lenses is that they distort the image in order to keep lines straight. The further away from the center the larger is the distortion. This means that stars that travels from the edge towards the center will appear to move faster at the edge, slow down in the center and then accelerate again as they move out of the frame.

And since the lens doesn’t move when using the astro tracer the stars will appear to change speed as they move across the sensor. The astro tracer can obviously not compensate for this and thus one ends up with prolonged stars when using wide rectilinear lenses and long exposures.
I read once that Ultra wide angle lenses had their kind of distortion, and fisheye lenses had their kinda distortion so it came down to deciding which one preferred.

I have the Sigma 10-20, but eventually got to dislike the rectilinear distortion at the wide end. Clouds often had this stringy diagonal look of them, towards each upper corner. I think Gimbal nailed it when describing the movement of the sensor relative to the fixed lens, and how that introduces different speeds of stars across the face of the sensor - depending on their location. The sensor movement can only compensate for one speed of movement so this explains differences in star appearances at the corners.

So best to keep exposure times low and especially be careful with ULWA lenses.
04-29-2015, 08:54 PM   #57
Pentaxian




Join Date: May 2009
Location: Ponoka Alberta Canada
Photos: Gallery
Posts: 545
QuoteOriginally posted by Pete_XL Quote
Ian, you are right, the accuracy of the GPS is within an acceptable range and not the problem in my point of view. More important for the calculations of the rotation than the exact coordinates is the azimutal angle and the inklination of the optical axis of the lens. The crucial negative factor with O-GPS1 is the fluctuation of the azimutal angle that can be seen in the picture. And that is what I wanted to show with that example.


O-GPS determins the azimutal angle by measuring the earth magnetic field. Obviously this not precise enough and/or varies in time. The inclination is determined by acellerator sensors in the camera and should also be quite unprecise in comparison with the accuracy that is normaly aspired in polar alignment when shooting astrophotos.



O-GPS1 is fine for easily tracking stars especially with short lenses. But if it comes to long lenses and/or long exposure times it reaches ist limits. I have done thousands of frames in the night sky with astrotracer and like it a lot. But this thread is about the new pixel shift technology that needs a precision of a few (one?) pixel to work. I am certain that astrotracer and pixel shift can not go together.
Are you sure that the determines Azimuth
With the coarse-acquisition (C/A) GPS it will very accurately calculate azimuth down to less than a minute often times to 10 seconds in the bearing so why would they not use this data Using the magnetic field to work out azimuth when the data is already calculated when GPS works out its location seems funny to me
I would take some test photos with a magnet around the camera setup and see if indeed they use the field to solve azimuth

I just tested with the K5 on that there is enough of a field on the camera body to disrupt a compass so I find using the field to determine azimuth very unlikely
04-29-2015, 11:43 PM   #58
Site Supporter
Pete_XL's Avatar

Join Date: Mar 2010
Location: Germany
Photos: Gallery
Posts: 207
QuoteOriginally posted by Ian Stuart Forsyth Quote
Are you sure that the determines Azimuth
With the coarse-acquisition (C/A) GPS it will very accurately calculate azimuth down to less than a minute often times to 10 seconds in the bearing so why would they not use this data Using the magnetic field to work out azimuth when the data is already calculated when GPS works out its location seems funny to me
I would take some test photos with a magnet around the camera setup and see if indeed they use the field to solve azimuth

I just tested with the K5 on that there is enough of a field on the camera body to disrupt a compass so I find using the field to determine azimuth very unlikely


To compensate earth's rotation by sensor movement the lens axis must be calculated against the rotation axis of the earth. Both are directional vectors.The lens axis is a position vector. How could a directional vector (azimut and inclination) be determined only by GPS? GPS delivers 3-D data only about the Position. To solve the necessary equations you need Information on where the lens is pointing to. To get the pointing angles (azimut and inclination) of the lens additional data is needed. As far as I know O-GPS1 determins azimut by inside magnetic compass and tilt/inclination by acceleration sensors in the camera. This is why a precise calibration of these sensors is necessary before every operation of the astrotracer.


For sure the measurement of magnetic field to get exact azimut (tolerance perhaps a few arcminutes) is a weak method in this case especially if the field is disrupted by adverse influences.
04-30-2015, 12:23 AM   #59
Pentaxian




Join Date: Apr 2007
Location: Sweden
Posts: 1,940
Right, a GPS can’t calculate azimuth if It’s not moving. So it has to use a magnetic sensor to get that information. I’ve done some tests with the compass (and this can be done even if the GPS can’t get a lock) by calibrating and checking the accuracy. Over and over again trying to find a method for calibrating that gives the best result. Without much success I might add.

Even if I calibrate two times in a row, exactly in the same way, the result varies. Sometimes it’s near perfect 360 degrees around, but usually the accuracy varies depending on direction.

It’s kind of strange that it works as well as is does though, considering there are strong magnets inside the camera.

So that problem might be even worse now with the K3II considering the fact that the now built in magnetic sensor is even closer to the magnets inside the camera compared with O-GPS that is mounted on top of the flash shoe.
04-30-2015, 04:58 AM   #60
Site Supporter
Pete_XL's Avatar

Join Date: Mar 2010
Location: Germany
Photos: Gallery
Posts: 207
Yes, Gimbal, that's what brings me again to the conclusion that Astrotracer is a fine, light and easy-to-use-tool for astrophotos with shorter lenses and suitable exposure times (what I use it for with a lot of fun!). A a built-in GPS with Astrotracer in K3 II will be a useful added value I would be pleased to own.


But when it comes to new pixel shift technology together with Astrotrace (what the diskussion as about) the limitations of the Astrotrace should be overrun by far.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
astrotracer, camera, dslr, exposure, filter, k-3, k-3 ii, k3, landscape, landscapes, lens, lot, movement, object, pentax k-3, pentax news, pentax rumors, photography, pixel, pixel shift, precision, resolution, ricoh, rotation, sensor, shift, shots, size, stars, technology, time
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Pixel shift practical applications Spodeworld Pentax K-3 20 04-25-2015 10:01 AM
K-3 II pixel shift resized. Wow factor. JimmyDranox Pentax K-3 21 04-25-2015 09:37 AM
Pixel shift and bracketing Spodeworld Pentax K-3 9 04-23-2015 08:56 PM
some questions for K-3 and K-5 owners carrrlangas Pentax K-3 11 06-27-2014 05:27 PM
Several questions about Pentax K-5 II and something else Casion Pentax K-5 17 07-20-2013 02:07 PM



All times are GMT -7. The time now is 03:36 PM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top