Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Canon APS-C video astronomy with C5

  • Please log in to reply
22 replies to this topic

#1 Nicole Sharp

Nicole Sharp

    Apollo

  • -----
  • topic starter
  • Posts: 1403
  • Joined: 12 Jun 2018
  • Loc: Cumberland, Maryland, USA

Posted 30 November 2018 - 03:55 PM

Forking this from another thread, but can anyone tell me if I can record the full disc of Sol or Luna in 1080p video with a Canon APS-C camera on a C5 (1250/125 Schmidt-Cassegrain)?  The diameter of the baffle tube on the C5 is 27 mm, and the diagonal of the camera sensor is 26.8 mm (with an area of 22.3*14.9 mm^2).  1250 mm of focal length then should provide a field of view of 1.2 Lunar diameters (with one Lunar diameter being set as 34.1 arcminutes, the maximum angular diameter of Luna).  Photography with the Canon APS-C SL2 (200D) is 6000*4000 pixels, but videography is 1920*1080 pixels.  Would the field of view then be cropped to not fit the full disc of Luna at its maximum apparent size?

 

The C5 hasn't been shipped yet, so please let me know ASAP if I should cancel it if it was my intention to use it for full-disc Solar/Lunar video astronomy with a Canon APS-C camera.


Edited by Nicole Sharp, 30 November 2018 - 03:59 PM.


#2 Nicole Sharp

Nicole Sharp

    Apollo

  • -----
  • topic starter
  • Posts: 1403
  • Joined: 12 Jun 2018
  • Loc: Cumberland, Maryland, USA

Posted 30 November 2018 - 04:03 PM

Should probably move this to the DSLR forum; sorry.


Edited by Nicole Sharp, 30 November 2018 - 04:03 PM.


#3 Nicole Sharp

Nicole Sharp

    Apollo

  • -----
  • topic starter
  • Posts: 1403
  • Joined: 12 Jun 2018
  • Loc: Cumberland, Maryland, USA

Posted 30 November 2018 - 04:28 PM

I think I got something here.  If I did the math right, and the video mode of the Canon APS-C camera is cropping the usable area of the sensor, then the effective sensor size should be (22.3 mm)*([22.3 mm]/[1920/1080]) = (22.3 mm)*(12.5 mm), down from (22.3 mm)*(14.9 mm).  That would provide a vertical field of view of 1.01 Lunar diameters (+/-) 0.1 Lunar diameters.  So it should just barely work, but unable to be certain to due to only knowing the diameter of the baffle tube to one significant figure.  Simulating on Stellarium for the January 2019 total Lunar eclipse (when Luna will have an angular diameter of 99.6% of it's maximum possible angular diameter) as 22.3*12.5 mm^2 and 1920*1080 px, shows that the disc just barely fits.  Even ignoring the uncertainty, centering the disc of Luna in the camera's field of view to within 0.01 Lunar diameters (20 arcseconds) would be really difficult I think on a 5SE mount.  So should I return it?  My second choice after a C5 would probably be something like a 1000/203 Newtonian, but then I need a more expensive mount.


Edited by Nicole Sharp, 30 November 2018 - 04:35 PM.


#4 OleCuss

OleCuss

    Vanguard

  • *****
  • Posts: 2394
  • Joined: 22 Nov 2010

Posted 30 November 2018 - 05:02 PM

I do think that a C5 isn't going to make you happy for imaging.

 

However, if staying a little lighter than the 8" astrograph would help you with the mount situation, you might consider going with Orion's 6" astrograph?


Edited by OleCuss, 30 November 2018 - 05:03 PM.

  • Nicole Sharp likes this

#5 Kendahl

Kendahl

    Apollo

  • *****
  • Posts: 1416
  • Joined: 02 Feb 2013
  • Loc: Omaha, Nebraska

Posted 30 November 2018 - 05:13 PM

Focal length of a C5 is 1250 mm. Pixel size for an 18 megapixel Canon DSLR is 4.3 microns. Therefore, image scale is 206 x 4.3 / 1250 = 0.71 arc-secs/pixel. Width of the sensor is 5184 pixels which means the horizontal field of view is 61 arc-min. Aspect ratio of the video frame is 16x9. Therefore, height of the frame is 61 x ( 9 / 16 ) = 34 arc-min. Since the sun's diameter is about 32 arc-min, it will barely fit. Your tracking will have to be perfect to avoid cutting off an edge. Results for other Canon DSLRs with APSC sensors will be similar since the sensor are about the same size even though the pixel count varies.

 

A 0.63x reducer-corrector will increase the height of the field of view to 54 arc-min. It will also correct some optical defects inherent to Schmidt-Cassegain telescopes.

 

Have you considered taking a series of still photos and stringing them together with a video editing program? The remote control function of Canon's EOS Utility program will let you take a photo every five seconds. I've used it or Backyard EOS to make time lapse videos of events like lunar occultations and eclipses. My first videos were made with Windows Moviemaker. It no longer comes with Windows and had the limitation that the fastest frame rate for a video made up of still photos was only eight per second. Fifteen per second is a good minimum. I now use VideoPad from NCH Software.


  • Nicole Sharp likes this

#6 Nicole Sharp

Nicole Sharp

    Apollo

  • -----
  • topic starter
  • Posts: 1403
  • Joined: 12 Jun 2018
  • Loc: Cumberland, Maryland, USA

Posted 30 November 2018 - 05:14 PM

I do think that a C5 isn't going to make you happy for imaging.

 

However, if staying a little lighter than the 8" astrograph would help you with the mount situation, you might consider going with Orion's 6" astrograph?

You read my brain.  That was actually my third choice.  But a 2X Barlow at 650/150 might still be too much focal length.  The field of view is proportional to only the focal length and the camera sensor size, so if 1250 mm is too tight a fit, then an effective 1300 mm wouldn't work.  A field of view of 34.1 arcminutes in a Canon APS-C camera corresponds to a focal length of 1502 mm.  If I am doing the math right, then the field of view for videography instead of photography should be cropped by 0.844X, corresponding to a focal length of 1267 mm---which I don't think I did that right actually, since the computation above showed that 34.1 arcminutes should fit into the field of view at 1250 mm within 0.01*34.1 arcminutes.  I think that around 1000 mm would probably be ideal for full-disc Lunar/Solar videography with a Canon APS-C camera.


Edited by Nicole Sharp, 30 November 2018 - 05:19 PM.


#7 headlight_fluid

headlight_fluid

    Explorer 1

  • -----
  • Posts: 50
  • Joined: 20 Dec 2013
  • Loc: Atlanta, GA, US

Posted 30 November 2018 - 05:19 PM

I don't believe video mode uses a cropped area of the sensor, instead it does something like binning.

 

To test this out for sure, take a still photo with the camera and then take video of the same thing with the same lens and same position. You'll see the field of view is exactly the same. If they were cropping the sensor then the video would be a smaller field of view than the still photo.

 

Edit: I may be totally wrong about this! Apparently some (all?) DSLRs do crop the sensor for video. 

 

https://pixelvalleys...rame-sensor-bad

 

https://www.eoshd.co...and-cropped-4k/

 

I can say from personal experience that the video from my ASP-C camera (Sony Nex-6, now known as a6000 I think) has the exact same field of view as still shots.


Edited by headlight_fluid, 30 November 2018 - 05:27 PM.


#8 Nicole Sharp

Nicole Sharp

    Apollo

  • -----
  • topic starter
  • Posts: 1403
  • Joined: 12 Jun 2018
  • Loc: Cumberland, Maryland, USA

Posted 30 November 2018 - 05:25 PM

Focal length of a C5 is 1250 mm. Pixel size for an 18 megapixel Canon DSLR is 4.3 microns. Therefore, image scale is 206 x 4.3 / 1250 = 0.71 arc-secs/pixel. Width of the sensor is 5184 pixels which means the horizontal field of view is 61 arc-min. Aspect ratio of the video frame is 16x9. Therefore, height of the frame is 61 x ( 9 / 16 ) = 34 arc-min. Since the sun's diameter is about 32 arc-min, it will barely fit. Your tracking will have to be perfect to avoid cutting off an edge. Results for other Canon DSLRs with APSC sensors will be similar since the sensor are about the same size even though the pixel count varies.

 

A 0.63x reducer-corrector will increase the height of the field of view to 54 arc-min. It will also correct some optical defects inherent to Schmidt-Cassegain telescopes.

 

Have you considered taking a series of still photos and stringing them together with a video editing program? The remote control function of Canon's EOS Utility program will let you take a photo every five seconds. I've used it or Backyard EOS to make time lapse videos of events like lunar occultations and eclipses. My first videos were made with Windows Moviemaker. It no longer comes with Windows and had the limitation that the fastest frame rate for a video made up of still photos was only eight per second. Fifteen per second is a good minimum. I now use VideoPad from NCH Software.

 

I did think of using continuous shooting to create a time-lapse video.  But I would like to have the option of doing both photos and videos without needing a near-perfect alignment and near-perfect tracking to fit the disc within the field of view.  I went ahead and canceled the order.  It hasn't even been shipped yet, so hopefully won't be any problem.  I think that maybe I should just order the DSLR camera first, and try it out with the telescope I have now, before upgrading to a GOTO telescope.  I already have a 1000/114 telescope, so if for example I wanted to upgrade to the Celestron Advanced VX 8" Newtonian (1000/200), the focal length and the field of view would actually be exactly the same, just with more resolution from the increased aperture (and the catadioptric design of the Bird-Jones Newtonian should hopefully provide enough backfocus for a DSLR unlike other 4.5" Newtonians).  But I think an 8" Newtonian might be too big and heavy for me.  I really liked the light weight and portability of the 5SE.


Edited by Nicole Sharp, 30 November 2018 - 05:30 PM.


#9 Nicole Sharp

Nicole Sharp

    Apollo

  • -----
  • topic starter
  • Posts: 1403
  • Joined: 12 Jun 2018
  • Loc: Cumberland, Maryland, USA

Posted 30 November 2018 - 05:34 PM

I don't believe video mode uses a cropped area of the sensor, instead it does something like binning.

 

To test this out for sure, take a still photo with the camera and then take video of the same thing with the same lens and same position. You'll see the field of view is exactly the same. If they were cropping the sensor then the video would be a smaller field of view than the still photo.

 

Edit: I may be totally wrong about this! Apparently some (all?) DSLRs do crop the sensor for video. 

 

https://pixelvalleys...rame-sensor-bad

 

https://www.eoshd.co...and-cropped-4k/

 

I can say from personal experience that the video from my ASP-C camera (Sony Nex-6, now known as a6000 I think) has the exact same field of view as still shots.

Not sure what binning is, but the geometry still doesn't sound right.  The relative height of the video file is shorter than that of a still photo.  You have to be losing some field of view somewhere, otherwise there would be distortion.


Edited by Nicole Sharp, 30 November 2018 - 05:37 PM.


#10 Nicole Sharp

Nicole Sharp

    Apollo

  • -----
  • topic starter
  • Posts: 1403
  • Joined: 12 Jun 2018
  • Loc: Cumberland, Maryland, USA

Posted 30 November 2018 - 05:48 PM

Focal length of a C5 is 1250 mm. Pixel size for an 18 megapixel Canon DSLR is 4.3 microns. Therefore, image scale is 206 x 4.3 / 1250 = 0.71 arc-secs/pixel. Width of the sensor is 5184 pixels which means the horizontal field of view is 61 arc-min. Aspect ratio of the video frame is 16x9. Therefore, height of the frame is 61 x ( 9 / 16 ) = 34 arc-min. Since the sun's diameter is about 32 arc-min, it will barely fit. Your tracking will have to be perfect to avoid cutting off an edge. Results for other Canon DSLRs with APSC sensors will be similar since the sensor are about the same size even though the pixel count varies.

 

A 0.63x reducer-corrector will increase the height of the field of view to 54 arc-min. It will also correct some optical defects inherent to Schmidt-Cassegain telescopes.

 

Have you considered taking a series of still photos and stringing them together with a video editing program? The remote control function of Canon's EOS Utility program will let you take a photo every five seconds. I've used it or Backyard EOS to make time lapse videos of events like lunar occultations and eclipses. My first videos were made with Windows Moviemaker. It no longer comes with Windows and had the limitation that the fastest frame rate for a video made up of still photos was only eight per second. Fifteen per second is a good minimum. I now use VideoPad from NCH Software.

Unfortunately, reducing a C5 with a Canon APS-C camera (27 mm) actually makes things worse, since the 27-mm image circle is reduced by 0.63X.  Same problem with pretty much any other SCT.  Even the 42-mm baffle of the EdgeHD 8 reduced by 0.7X provides a field of view of only 1.1 Lunar diameters for photography, and less than that for videography.  Using a smaller camera sensor could work out with a reduced Cassegrainian OTA, but I want to be using a self-contained recording system, without needing a computer.


Edited by Nicole Sharp, 30 November 2018 - 05:52 PM.


#11 Kendahl

Kendahl

    Apollo

  • *****
  • Posts: 1416
  • Joined: 02 Feb 2013
  • Loc: Omaha, Nebraska

Posted 30 November 2018 - 06:23 PM

I don't believe video mode uses a cropped area of the sensor, instead it does something like binning.

Canon DSLRs use almost the full width of the sensor for videos. They generate the 16x9 aspect ratio for the video by cropping the height of the frame.

 

Binning means combining several pixels into one. For example, 2x2 binning means taking two adjacent pixels from a row and two from the same columns of the next row and combining all four into one pixel. 3x3 binning means combining nine pixels into one. The "binning" that Canon uses to make videos with an 18 megapixel camera is 2.7x2.7 which must mean doing something funky with partial pixels. If the camera sensor dimensions are 6000x4000, the ratio is 3.13.


  • headlight_fluid likes this

#12 james7ca

james7ca

    Fly Me to the Moon

  • *****
  • Posts: 7333
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 30 November 2018 - 07:52 PM

I’m pretty sure a C5 will not produce a sharp image over a full APS-C field. Having said that I know that my C5 takes really nice lunar images with my smaller cameras and I’ve done full disk mosaics of the crescent moon that have looked very good (link below).

 

https://www.cloudyni...178c/?p=8530915


Edited by james7ca, 30 November 2018 - 07:58 PM.


#13 james7ca

james7ca

    Fly Me to the Moon

  • *****
  • Posts: 7333
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 30 November 2018 - 08:06 PM

Also, a compressed and resampled video capture isn’t going to look that good.

#14 Kendahl

Kendahl

    Apollo

  • *****
  • Posts: 1416
  • Joined: 02 Feb 2013
  • Loc: Omaha, Nebraska

Posted 30 November 2018 - 08:47 PM

I’m pretty sure a C5 will not produce a sharp image over a full APS-C field.

Would a 0.63x reducer-corrector, combined with a 2x Barlow or focal length extender to recover focal length, fix this? Problems with the outer portions of the field are why Celestron developed the Edge HD.



#15 Kendahl

Kendahl

    Apollo

  • *****
  • Posts: 1416
  • Joined: 02 Feb 2013
  • Loc: Omaha, Nebraska

Posted 30 November 2018 - 09:00 PM

I'm not convinced that a C5 with a reducer-corrector won't work for you. However, it would be best to check with an authoritative source like tech support at a vendor or manufacturer. I've dealt with Astronomics, Orion, OPT, Agena and Highpoint as well as Celestron. No experience with Meade.

 

Since you have cancelled your order for a C5 and don't yet own a DSLR, you are starting from scratch. Your requirements are to fit the disc of the sun or moon comfortably into the frame of a 1080p video in a stand alone system that consists of a telescope, tracking mount and a DSLR. Rather than buy something and try to find something else that will work with it, I suggest you think the whole system through before making a commitment to any single component.

 

My experience is with a refractor (102 mm aperture, 783 mm focal length) and an 18 megapixel Canon T3i (aka 600D) DSLR. Field of view for snapshots is 98x65 arc-min. This shrinks to 94x53 arc-min in a 1080p video. The frame is plenty large for your application. Telescope and camera weigh about 15 pounds which can be handled comfortably by a mount with a 30 pound rating. (Unless the mount comes from a premium manufacturer like Astrophysics, discount their load rating by 50% for photography.) Stick to a triplet ED refractor to avoid color fringes around bright objects. The principal disadvantage with refractors is that, for their aperture, they are long, heavy and expensive compared to Schmidt-Cassegrains.

 

A Newtonian astrograph with a focal length no longer than 1000 mm would work. It needs to be an astrograph rather than a scope designed for visual observing. A DSLR's sensor is buried deep within the body rather than being out front like an eyepiece. That puts the sensor too far out to achieve focus with a visual Newtonian. An astrograph takes this into account. Suppliers include an extension to use an eyepiece with an astrograph. Newtonians need to be collimated frequently. This gets difficult as the focal ratio decreases. It's easy with an f/8 and significantly harder with an f/4. Newtonians are much longer than Schmidt-Cassegrains. That makes them heavier for a given aperture and much more difficult for a mount to keep stable. They are more vulnerable to wind. Unless the tube is very rigid, it can flex enough to compromise collimation. I tried to use a 10" Newtonian (30 pounds and 4 feet long) on my iOptron iEQ45 mount. Although the iEQ45 is rated for a 45 pound load, the 10" scope overloaded it badly enough to damage the gears. It's fine with my refractor and guide scope and should work with a C8, too, but that's the upper limit.

 

You will need a tracking mount. For long term success and flexibility, my recommendation is an Orion Sirius GoTo or better. Explore Scientific's Bresser is significantly cheaper but has yet to build a reputation. (There used to be a version for $500 but ES upgraded the electronics and raised the price.) Celestron's Advanced VX has a spotty reputation. Some are good. Some aren't. The learning curve for a German equatorial mount is steeper than for an altitude-azimuth mount. For one thing, it needs to be polar aligned to track well. The advantage to a GEM is that it can be used for long exposure photography, too, while an alt-az mount cannot.


  • Nicole Sharp likes this

#16 Nicole Sharp

Nicole Sharp

    Apollo

  • -----
  • topic starter
  • Posts: 1403
  • Joined: 12 Jun 2018
  • Loc: Cumberland, Maryland, USA

Posted 01 December 2018 - 02:07 AM

I'm not convinced that a C5 with a reducer-corrector won't work for you. However, it would be best to check with an authoritative source like tech support at a vendor or manufacturer. I've dealt with Astronomics, Orion, OPT, Agena and Highpoint as well as Celestron. No experience with Meade.

 

Since you have cancelled your order for a C5 and don't yet own a DSLR, you are starting from scratch. Your requirements are to fit the disc of the sun or moon comfortably into the frame of a 1080p video in a stand alone system that consists of a telescope, tracking mount and a DSLR. Rather than buy something and try to find something else that will work with it, I suggest you think the whole system through before making a commitment to any single component.

 

My experience is with a refractor (102 mm aperture, 783 mm focal length) and an 18 megapixel Canon T3i (aka 600D) DSLR. Field of view for snapshots is 98x65 arc-min. This shrinks to 94x53 arc-min in a 1080p video. The frame is plenty large for your application. Telescope and camera weigh about 15 pounds which can be handled comfortably by a mount with a 30 pound rating. (Unless the mount comes from a premium manufacturer like Astrophysics, discount their load rating by 50% for photography.) Stick to a triplet ED refractor to avoid color fringes around bright objects. The principal disadvantage with refractors is that, for their aperture, they are long, heavy and expensive compared to Schmidt-Cassegrains.

 

A Newtonian astrograph with a focal length no longer than 1000 mm would work. It needs to be an astrograph rather than a scope designed for visual observing. A DSLR's sensor is buried deep within the body rather than being out front like an eyepiece. That puts the sensor too far out to achieve focus with a visual Newtonian. An astrograph takes this into account. Suppliers include an extension to use an eyepiece with an astrograph. Newtonians need to be collimated frequently. This gets difficult as the focal ratio decreases. It's easy with an f/8 and significantly harder with an f/4. Newtonians are much longer than Schmidt-Cassegrains. That makes them heavier for a given aperture and much more difficult for a mount to keep stable. They are more vulnerable to wind. Unless the tube is very rigid, it can flex enough to compromise collimation. I tried to use a 10" Newtonian (30 pounds and 4 feet long) on my iOptron iEQ45 mount. Although the iEQ45 is rated for a 45 pound load, the 10" scope overloaded it badly enough to damage the gears. It's fine with my refractor and guide scope and should work with a C8, too, but that's the upper limit.

 

You will need a tracking mount. For long term success and flexibility, my recommendation is an Orion Sirius GoTo or better. Explore Scientific's Bresser is significantly cheaper but has yet to build a reputation. (There used to be a version for $500 but ES upgraded the electronics and raised the price.) Celestron's Advanced VX has a spotty reputation. Some are good. Some aren't. The learning curve for a German equatorial mount is steeper than for an altitude-azimuth mount. For one thing, it needs to be polar aligned to track well. The advantage to a GEM is that it can be used for long exposure photography, too, while an alt-az mount cannot.

 

I was looking at refractors too.  Sol and Luna are both bright enough you can get away with smaller apertures with enough focal length.  Something like this would be pretty good I think for manual tracking:

 

https://www.telescop.../335/p/9024.uts

 

Specs are 910/90, with a chromatic aberration index of 2.85 (>2.25, ~3).  But the length of the tube would give it a really large moment of inertia if I ever upgraded to a GOTO mount.  Also, not sure how much deviation there is in the orbits of Sol and Luna from the equatorial alignment of a manual GEM over the 6 hours needed for an eclipse or transit.  My guess is the deviation might be smaller than the polar-alignment error, especially if polar-aligning in the daytime.  But I don't really want to sit with a telescope to manually track for 6 hours in the freezing cold.  So I would have to resort to still photography or get a motor drive.

 

Even a 360/60 refractor I think could be Barlowed to 3X maybe (effective focal length of 1080 mm) with the reduced resolution of 1080p.  I would think also that Barlowing an achromatic refractor would further reduce the chromatic aberration by increasing the focal ratio.  So maybe a Barlowed short-tube refractor is a better option?  I actually did order a 360/60 achromatic refractor to use as a wide-field deep-sky OTA on the 5SE mount.  So maybe I should order the DSLR and try it with the 360/60 refractor at 3X and the 1000/114 catadioptric Newtonian first before trying to get a GOTO mount or another OTA.  Right now, Celestron has the SE and Evolution mounts on sale, but I think the AVX will go on sale in December.  My top choice for any GOTO mount is the Sky-Watcher AZ-EQ5 (a.k.a. Orion Sirius Pro), but it is a lot more expensive.


Edited by Nicole Sharp, 01 December 2018 - 02:51 AM.


#17 Nicole Sharp

Nicole Sharp

    Apollo

  • -----
  • topic starter
  • Posts: 1403
  • Joined: 12 Jun 2018
  • Loc: Cumberland, Maryland, USA

Posted 01 December 2018 - 02:40 AM

Would a 0.63x reducer-corrector, combined with a 2x Barlow or focal length extender to recover focal length, fix this? Problems with the outer portions of the field are why Celestron developed the Edge HD.

That would make an effective focal length of 1575 mm, so no.  A 1.5X Barlow and a 0.63X reducer would make 1181 mm though, so maybe, but still sounds like a bad idea to be tele-reducing and tele-extending at the same time.

 

And the image circle would be 25.5 mm on the 27-mm sensor (I think).  At 1181 mm, the vignetting might be outside of the disc, but still could be close, and could be an issue when filming Lunar occultations (is the star disappearing because it is being vignetted or because it is being occulted?).  Plus I would prefer to get Barlows of prime-number magnifications if possible.


Edited by Nicole Sharp, 01 December 2018 - 02:54 AM.


#18 Nicole Sharp

Nicole Sharp

    Apollo

  • -----
  • topic starter
  • Posts: 1403
  • Joined: 12 Jun 2018
  • Loc: Cumberland, Maryland, USA

Posted 01 December 2018 - 05:52 AM

I think I did do the numbers correctly above.  To fit a vertical field of view of 34.1 arcminutes or more into an effective sensor size of 22.3*12.6 mm^2 (cropped from a full sensor size of 22.3*14.9 mm^2), a focal length of 1267 mm or less is needed.  So a focal length of 1250 mm would provide a vertical field of view consisting of 98.6% of an angular diameter of 34.1 arcminutes, meaning that the maximum angular diameter of the disc of Luna would need to be centered in the field of view to within 0.0101 Lunar diameters (28.5 arcseconds).  To get a field of view of at least 1.25 Lunar diameters (which would be a lot easier to work with), a focal length of 1014 mm or less would be needed.  So I think a focal length between 600 mm (field of view of 2 Lunar diameters) and 1000 mm (field of view of 1.25 Lunar diameters) could be optimum to leave a practical amount of space around the disc of Luna (800 mm would be 1.5 Lunar diameters).  A field of view of 1.9 Lunar diameters in a 650/150 Newtonian might work out well to reduce off-axis coma, so that there would be a margin of 0.47 Lunar diameters above and below the disc, though that would provide less detail than a field of view of 1.3 Lunar diameters in a 1000/203 Newtonian (with a margin of 0.13 Lunar diameters above and below the disc).


Edited by Nicole Sharp, 01 December 2018 - 06:34 AM.


#19 james7ca

james7ca

    Fly Me to the Moon

  • *****
  • Posts: 7333
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 01 December 2018 - 07:03 PM

I've got to caution again that capturing full disk lunar or solar images using the HD video output of a DSLR isn't going to produce very impressive images. That video is going to be resampled and compressed before it is even output from the camera and in any case a one mega pixel image of the full disk of the sun or moon isn't going to provide much detail. That said, if you are only after low resolution images of the full disk then you might be satisfied with the results, so some of this will certainly be up to the eye of the beholder.

 

It is, however, possible to do some pretty good full-disk lunar and solar imaging using STILL capture with an APS-C camera. At low ISO settings you can even get nice images with a single frame and by stacking just a few still frames you can get even higher resolution and less noise.

 

Here are links to two images (one of the gibbous moon and the other of a solar eclipse) that I did with APS-C cameras using STILL captures:

 

Gibbous moon (On CN with a link to a higher resolution on Flickr, 5" refractor):  https://www.cloudyni...-4#entry6719422

 

Solar eclipse (on Flickr, 80mm refractor):   https://flic.kr/p/pv8pBA


Edited by james7ca, 01 December 2018 - 07:09 PM.


#20 Nicole Sharp

Nicole Sharp

    Apollo

  • -----
  • topic starter
  • Posts: 1403
  • Joined: 12 Jun 2018
  • Loc: Cumberland, Maryland, USA

Posted 01 December 2018 - 07:38 PM

I've got to caution again that capturing full disk lunar or solar images using the HD video output of a DSLR isn't going to produce very impressive images. That video is going to be resampled and compressed before it is even output from the camera and in any case a one mega pixel image of the full disk of the sun or moon isn't going to provide much detail. That said, if you are only after low resolution images of the full disk then you might be satisfied with the results, so some of this will certainly be up to the eye of the beholder.

 

It is, however, possible to do some pretty good full-disk lunar and solar imaging using STILL capture with an APS-C camera. At low ISO settings you can even get nice images with a single frame and by stacking just a few still frames you can get even higher resolution and less noise.

 

Here are links to two images (one of the gibbous moon and the other of a solar eclipse) that I did with APS-C cameras using STILL captures:

 

Gibbous moon (On CN with a link to a higher resolution on Flickr, 5" refractor):  https://www.cloudyni...-4#entry6719422

 

Solar eclipse (on Flickr, 80mm refractor):   https://flic.kr/p/pv8pBA

Yeah, been running numbers, and it would be much less impressive for the 2019 Mercury transit than it will be for the 2019 total Lunar eclipse.  I got the following for the size of Mercury in a Canon APS-C sensor at 1920*1080:

 

2 px @ 360 mm
2 px @ 420 mm
3 px @ 700 mm
3 px @ 714 mm
3 px @ 720 mm (1623 km/px)
4 px @ 840 mm (1392 km/px)

 

Still think I made the right decision in canceling the 5SE.  I was getting it more for the OTA than for the mount.  I think I will be better off with a 2X Barlow on a small refractor for the 2019 total Lunar eclipse, but I still might want a longer focal length for the 2019 Mercury transit and planetary observing in general.

 

Doing continuous shooting and then time-lapsing the video would not improve resolution at all though.  E.g. taking one photograph per second for 60 minutes can be time-lapsed at 60 FPS into 1 minute of video at a very high resolution of 6000*4000 pixels.  But if I want to share that video online (or even if I want to share any still photographs online), it will have to be compressed down to 1920*1080 or smaller anyway, and I will lose all of the extra resolution.  It would be faster and easier to just film at 1080p and then remove the extra frames if I wanted to time-lapse.  If I want planets with their small angular diameters to be framed well without significant loss of resolution though, I really need to get a camera with a smaller sensor, or a camera that supports region-of-interest filming.

 

Regarding sampling for Canon APS-C, I got the following:

 

(minimum theoretical resolution at red, green, and violet wavelengths per aperture)

50 mm.  3.52" / 2.53" / 2.01"
60 mm.  2.94" / 2.11" / 1.68"
70 mm.  2.52" / 1.8" / 1.44"
72 mm.  2.45" / 1.75" / 1.4"
102 mm.  1.73" / 1.24" / 0.987"

 

(resolution per pixel per focal length)

360 mm: 6.65"/px > 3.52" (360/50, 360/60)
420 mm: 5.70"/px > 2.45" (420/72)
700 mm: 3.42"/px > 2.94"  (700/60, 700/70)
714 mm: 3.36"/px > 1.73" (714/102)
720 mm: 3.33"/px > 2.94" (2X 360/60)
840 mm: 2.85"/px > 2.45" (2X 420/72)

 

It seems that the 1080p resolution is undersampled, even with a 2X Barlow, so that increasing the aperture would not appear to increase the resolution.  Therefore, the best option seems to be to get the longest focal length to get as much of the disc into the field of view (within an acceptable margin).

 

I capped out my OTA search to effective focal lengths between 634 mm (margin of 0.5 Lunar diameters) and 854 mm (margin of 0.25 Lunar diameters), and also to only OTAs of less than 15 pounds in weight.  That eliminates the Newtonian astrographs, and puts my two top choices as a 360/60 refractor or a 420/72 refractor for grab-and-go full-disc Solar/Lunar videos with a Canon APS-C.  I already ordered the Meade Adventure Scope 60 (360/60 achromat) to use as a wide-field deep-sky OTA on the 5SE mount.  So I can keep that as my grab-and-go for quick full-disc Solar/Lunar videos (such as ISS transits, Lunar eclipses, and Lunar occultations), and then focus on getting something with a longer focal length for planetary imaging.


Edited by Nicole Sharp, 01 December 2018 - 07:55 PM.


#21 james7ca

james7ca

    Fly Me to the Moon

  • *****
  • Posts: 7333
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 01 December 2018 - 08:39 PM

I'm not sure exactly what you are trying to achieve, so take the following with a "grain of salt."

 

A Mercury transit or a lunar eclipse are both fairly long events and I don't think you'd need or even want to capture those using a simple video capture. Just snap a single still image every minute or two and you'll end up with a higher resolution image that can be saved in much less space than a 30fps HD video stream. Then, take those stills and do a simple animation of the entire event. If you want an even higher time resolution then you can take a still frame every 15 seconds (or whatever).

 

Also, a lunar eclipse can get pretty dark and require a fairly long exposure which might not be that convenient to capture using video (unless you can capture each video frame with an exposure that might be equal to a large fraction of a second, meaning a video frame rate of 1 or 2 fps. 

 

As for the output resolution (an HD video) it's always better to resample downward in size and that may even produce a better result than if you captured an HD original. Beside, you may eventually want to make a video version in 4K resolution and if you capture in full-resolution still mode that would also be an option.



#22 OleCuss

OleCuss

    Vanguard

  • *****
  • Posts: 2394
  • Joined: 22 Nov 2010

Posted 01 December 2018 - 09:13 PM

OK, I think I missed something earlier.

 

Imaging a solar eclipse is a priority?  If so I'd be looking at a couple of things:

 

1.  If it is your first total solar eclipse (as it would be mine) I've repeatedly read that I shouldn't bother imaging it but rather go and experience it.

2.  If you are going to image the total solar eclipse I'd shift gears on the focal length you first anticipated by a lot.  IIRC the corona is likely to be imaged out to around 5 solar diameters?  So if you are tight on the solar disk you are going to be getting about the least impressive images I can imagine from my perspective.

3.  If one wants to capture the corona Nikon recommends no more than 900mm focal length with their APS-C sensors and if I were a novice solar eclipse imager I'm sure I'd be going considerably shorter than 900mm - when I was looking at the possibility of doing the 2017 eclipse I was looking at 600mm or less.  But then, the 600mm or less focal length priority I had was based in part on the idea that I wanted to experience the eclipse and I was not going to get too tight so that I could avoid being obsessed with any potential tracking problems (I was going to use a camera lens with a SkyTracker).

 

But my priorities needn't have any relation at all to yours.

 

I've no interest in doing imaging of the solar disc.  But I like testing my instruments for false color on the Moon because it tends to show up well there.  I've concerns that a fairly fast achromat might not give me a nice image on the Sun or the Moon.  I'm quite sure that with that achromat I'd not be worrying overly much about my sampling as it would likely be just about the least of my worries.

 

If you have the DSLR there is a decent chance you have a camera lens which may be better for the relatively wide FOV stuff than would be that 60mm achromat?

 

To go on further about when I was looking at imaging the 2017 total solar eclipse?  I finally decided that if I were going to do it I'd just use the Nikon P900.  Not a great lens and not a great sensor - and it'll only save in JPEG.  But it would be easy to use for a few quick and easy snapshots so that I'd not have to haul a bunch of gear around and I'd not be obsessing about the equipment instead of experiencing the event.  If what I got wasn't great I wasn't going to stress about it.  In the end I didn't get to go and did other stuff which just may have been more rewarding.


Edited by OleCuss, 01 December 2018 - 09:15 PM.


#23 Kendahl

Kendahl

    Apollo

  • *****
  • Posts: 1416
  • Joined: 02 Feb 2013
  • Loc: Omaha, Nebraska

Posted 02 December 2018 - 08:09 PM

I've done time lapse videos of long duration events like eclipses, transits and occultations.

  • Basic procedure is to take snapshots at a fixed interval of several seconds and string them together into a video using a video editing program. Video frame dimensions need not match 1080p.
  • Fifteen frames per second is fast enough. I like to make 60 second videos. That means 900 or so photos over the duration of the event.
  • For transits and occultations, select a combination of aperture, ISO and shutter speed that gives a good picture of the sun or moon and stick with it through the entire event.
  • Brightness varies too much through an eclipse to use the same settings throughout. You need a camera control program or device that enables you to adjust settings over a period of several hours. If it can run fast enough, bracket each exposure.
  • There are good programs for capturing solar eclipses. One of them is Eclipse Orchestrator.
  • How well the sun or moon stays steady in the center of the frame will depend on how well your mount tracks. If you are a programmer, it's not hard to write software to correct minor tracking errors by shifting images.
  • If capturing the event is a high priority, be prepared to travel as much as a thousand miles in any direction to get out from under cloud cover.

Edited by Kendahl, 02 December 2018 - 08:18 PM.



CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics