by alter-ego » Fri Mar 22, 2019 7:36 pm
Nitpicker wrote: ↑Fri Mar 22, 2019 6:52 am
neufer wrote: ↑Fri Mar 22, 2019 2:05 am
Nitpicker wrote: ↑Thu Mar 21, 2019 10:30 pm
The straight line of the sun path is by virtue of the direction the camera was pointed, rather than the equinox. Had the sun been less centred in the frame, the path would have been curved methinks.
With a normal (pseudo-pinhole/non-
fish eye) type camera lens
both the Milky Way & the sun's path at equinox appear as straight lines.
Not so sure about that. Even in the APOD, if you look closely, the trails just to the left (north) of the sun, at the top edge of frame, are curving ever so slightly to the right (south). I think this is because the lens axis is not pointed
perfectly eastward (which is not a criticism of the APOD, which is beautiful).
Edit: Of course, this APOD was taken a day or so before the equinox, so maybe that explains it. (Now I want to test to see if the celestial equator always appears straight in my images.)
Created by a normal camera lens, this image is essentially a stereographic projection in which all great circles (e.g. galactic equator, Earth's equator, and ecliptic)
can be imaged as straight lines. However, camera pointing does matter. If the camera is centered on the intersection of any two great circles, say the ideal horizon and equator, those great circles will appear as straight lines. This case is often encountered because we often want undistorted (flat) horizons. So, as it is here, this condition is had when looking due east/west on the horizon. If you point the camera away in any direction, the great circles will appear to curve, increasing the distortion the further away the camera is pointed. In the case of the ecliptic and galactic equator, they can be viewed as straight lines independent of the horizon
as long as the camera is centered at the intersection point.
So, the detail of the camera
not pointing due east will cause the equator (rising sun path) to curve a little. It should readily be visible in this image, but is not. However, it is possible to generate a "straight" equator this image at the cost of horizon and star-trail curvature distortions which is what I think is going on in this image. The real terrain geography is in fact sloped similarly to the image, and the star trails are curve close to the due east view. Consequently it is not obvious, nor trivial, to see the difference in the two stereographic projections. I'm assuming the photographer did not digitally modify the projection to generate the straight equator (rising sun path), although it is possible to do that.
[quote=Nitpicker post_id=290764 time=1553237536 user_id=141578]
[quote=neufer post_id=290760 time=1553220358 user_id=124483]
[quote=Nitpicker post_id=290759 time=1553207427 user_id=141578]
The straight line of the sun path is by virtue of the direction the camera was pointed, rather than the equinox. Had the sun been less centred in the frame, the path would have been curved methinks.[/quote]
With a normal (pseudo-pinhole/non-[url=https://apod.nasa.gov/apod/ap181109.html]fish eye[/url]) type camera lens
both the Milky Way & the sun's path at equinox appear as straight lines.
[/quote]
Not so sure about that. Even in the APOD, if you look closely, the trails just to the left (north) of the sun, at the top edge of frame, are curving ever so slightly to the right (south). I think this is because the lens axis is not pointed [i]perfectly [/i]eastward (which is not a criticism of the APOD, which is beautiful).
[i]Edit: Of course, this APOD was taken a day or so before the equinox, so maybe that explains it. (Now I want to test to see if the celestial equator always appears straight in my images.)[/i]
[/quote]
Created by a normal camera lens, this image is essentially a stereographic projection in which all great circles (e.g. galactic equator, Earth's equator, and ecliptic) [i]can[/i] be imaged as straight lines. However, camera pointing does matter. If the camera is centered on the intersection of any two great circles, say the ideal horizon and equator, those great circles will appear as straight lines. This case is often encountered because we often want undistorted (flat) horizons. So, as it is here, this condition is had when looking due east/west on the horizon. If you point the camera away in any direction, the great circles will appear to curve, increasing the distortion the further away the camera is pointed. In the case of the ecliptic and galactic equator, they can be viewed as straight lines independent of the horizon [u]as long as the camera is centered at the intersection point[/u].
So, the detail of the camera [u]not[/u] pointing due east will cause the equator (rising sun path) to curve a little. It should readily be visible in this image, but is not. However, it is possible to generate a "straight" equator this image at the cost of horizon and star-trail curvature distortions which is what I think is going on in this image. The real terrain geography is in fact sloped similarly to the image, and the star trails are curve close to the due east view. Consequently it is not obvious, nor trivial, to see the difference in the two stereographic projections. I'm assuming the photographer did not digitally modify the projection to generate the straight equator (rising sun path), although it is possible to do that.