26 replies to this topic

[banjo1000]

I am planning to use my ASI6200MC with FireCapture to shoot the total solar eclipse this April in Australia. Has anyone calculated the Gain - ISO equivalence?  If I have to do it myself, I would expose a daytime patch of sky with the same exposure for the ASI6200MC and a DSLR, to give a 50% histogram, and then cross check with a full solar disk through AstroSolar film or the moon.

 

I would be interested to hear experiences from anyone who has used a full or crop frame frame astro camera to shoot a total solar eclipse.

 

For a total eclipse, the exposures have to be programmed in advance.  No chance to tweak theexposures in real time based on the histogram, and live histograms aren't so useful anyway if shooting the outer corona because the inner corona will saturate the histogram.

[ch-viladrich]

Hello,

 

You don't need this

 

Here are some references :

- the exposure time for the inner corona next to the solar limb is about the same as for full Moon,

- at one solar radius from the solar limb, the corona is 1000x weaker than at the solar limb,

- the exposure time for the corona at five solar radii from solar limb is about the same as the earthshine.

 

So, just use the Moon  to determine the right exposure time with your equipement for the eclipse. And don't hesitate to use extra exposure times.

[ch-viladrich]

Another reference is this one by Druckmüller with the ASI 1600 from 1/1000 s to 1 s:

 

http://www.zam.fme.v...47mm/0-info.htm

 

Starting from this, you just have to play with the pixel size and sampling to find out the required exposure times from your ASI 6200. Quantum efficiency plays a little bit too, but you can forget about it.

 

Here are some eclipse images I took with a Basler 1920-155 (IMX174) in 2016

http://astrosurf.com...ernate2016.html

 

And just set the gain of the camera to the minimum value since we need to use the maximum dynamic range of the camera.

 

And keep gamma setting to 1.

[banjo1000]

Christian, Thanks for the helpful posts and for sharing the link to your gorgeous 2016 inner corona eclipse images. The description for the 2016 limb images mentions a thermal focus shift with the temperature drop. Did you refocus visually on the partially eclipse sun a few minutes before totality? Is a  

 

Your advise is to use a gamma setting of 1.0, yet 0.6 was used in the 2016 shots.  Was that intentional?  I have always used a gamma of one for planetary imaging.  The only time I deviate sometimes is to see the diffraction pattern of  a bahtinov mask.

 

Do you have any advice for determining the wRed and wGreen settings? This wasn't an issue for mono Basler IMX174, but it is for a color camera. I have experimented with an ASI462MC and a PlayerOne Neptune C-II using the Moon as a neutral white reference before shooting planets. Interestingly, I see some variation with altitude above the horizon.  I expect the values I found are site dependent, and probably vary with humidity, but I have not tested this. For the eclipse, I will only arrive in Exmouth a day or two before the eclipse.  I am thinking that my best bet would be to find a star with a Sun-like spectrum at the altitude of the Sun during totality, and to use that as a "neutral white" and to adjust wRed and wBlue so that the the Red and Blue histograms show the same maxima as Green.

 

Can you expound on the choice for the lowest gain setting to maximize the dynamic range in 2016? It seems that the goal was to capture fine structure in inner corona, but the dynamic range hardly varies much in your images, save for the black disk and the overexposed limb. In retrospect, woould you make the same choice? Might you increase the gain and run a shorter exposure to capture more frames, and perhaps capture the east limb too? It looks like you shifted the scope manually during the eclipse. That must have been nerve wracking!

 

Three separate exposures were used for the different limbs. How and why was the decision main to adjust the exposures on the fly?

 

Was a dark subtraction used, or just a flat? How and when did you shoot the flats?

 

Did you ever find an improved method to align your frames?  Druckmüller talks of a phase correlation. I understand the meaning of the words, but I don't understand what means in this context. When I processed with my 2017 TSE images, I tried to align on a prominence on the limb.

I plan to have 3 full frame cameras (2 DSLR and the ASI6200) shooting with three different focal lengths, and a Neptune C-II with a DSLR lens to catch the widest angle view so I can try to assemble an image worthy of Druckmüller. Druckmüller does not state the gain for his ASI1600-MM Pro, but it would make sense for him to use a low value to optimize dynamic range as he shoots out to 5 solar radii. Do you know the gain value Druckmüller used?

 

Maximizing dynamic range is important for me as well, but I am wondering whether I would be better off  using Gain=100. Please have a look at the ASI6200 specifications.  at Gain>=100, the camera modifies its readout electronic. Thanks to a much lower  readout noise, the dynamic range jumps back to 13.5 stops, with about 1/3 the exposure time.  Gaining 10 db (~3x) means I can record about 3 times as many frames in a fixed amount of time. Does that seem like a good compromise?

 

Do you understand how Druckmüller and others produce TSE images with a perfectly round Lunar disk, perhaps illuminated with Earthshine plunked in the middle of a corona? The Lunar disk is  bigger than the Solar disk.  Are they obscuring prominence details on the east and west limbs for the aesthetic appeal of a round Moon?  When I try to retain detail on each limb in a composite, the Moon is not round, because the Moon was in a different position near 2nd and 3rd contacts. My long exposures to capture Earthshine did not have clean edges due to encroachment/glow from the overexposed inner corona.  Is there a silver bullet to solve this problem?

 

Thanks,
Joe

Attached Thumbnails

• 

[ch-viladrich]

Hello Joe,

 

1)  Regarding the focus drift : I have a focuser with fine graduations. So, during the partial phase, I was able to monitor the change of focus. I made a "final" focus just a few minutes before totality, including the anticipatation of the focus shift between that moment and totality. It turned out pretty well. Needless to say, it is way better to have an instrument with low focus shift with temperature. It seems that photograph lenses have a focus very dependant on temperature drift. Nor completely sure of that. To be confirmed.

 

2) Using a 0.6 gamma was intentional. The initial idea was to facilitate the job of the software used for registration of images. Then, I went down to gamma 0.5 in 2017. It was indeed my worst idea since the last 20 years..It turned out pretty bad with strong artefacts.

http://astrosurf.com...sa/usa2017.html

 

So, don't do this mistake again and just use gamma = 1. Then, after acquisition, you can change gamma as needed with PIPP software.

 

3) Regarding RGB balance. I would use a Kodak color chart. If you don't have one, you can use a piece of white paper. But the "white" color and the "colors" of the color chart depends on the illuminant. So, it is a bit tricky at the end of the day. 

The Full Moon in a cristal clear sky and high in the sky is probably a good way to do it.

Using a G2 star is quite interesting too, though I found it a bit disappointing for planetary imaging.

In any case, it is good to start with the best possible tuning.

DSLR don't have this problem ...

 

4) The question of exposure vs dynamic (or gain) is certainly a good one for video cameras.

A first option is to saturate the USB3 connection (155 fps for the IMX174). My feeling is that even if you have more images this would not compensate the loss of dynamic range. This is a feeling, not a demonstration ;-)

 

The dynamic range on these images is quite high in fact. This one (with strong artefacts due to 0.5 gamma value) shows a 7.8 mag star :

http://astrosurf.com...00-17k-crop.jpg

 

and remember the luminosity of the corona drops by 1000x one radius from the solar limb. So you really need a huge dynamic range.

 

The question is more : what is the maximum exposure time for HR imaging ? For planetary / lunar / solar imaging, we are more in the 5 ms to 20 ms range, but we don't need dynamic range. And AutoStakker uses a number of registration points to compensate stretching due to turbulence.

For solar eclipse imaging, we can only choose one (or two) points of registration, because details of the corona have a too low contrast. This reduces the benefit of using very short exposure times.

 

At Jackson 2017, I went from 10 to 20 ms. Next time, I will probably stick to 20 - 25 ms. But other options are possible.

[ch-viladrich]

3) Moving the field of view : I had a motorized equatorial mount, and I moved manually from one field to the other. It was OK.

 

BTW, AIA images were used before the eclipse to identify the locations of the more active parts of the corona.

 

Noteworthy, the Rainbow RST 135E mount can be programmed to imagealong  the 360° of the solar limb with excellent accuracy. Videos are available on their site. It seems to be a very nice mount for solar eclipses !

 

 

4) In 2016 (and to a much lower extend in 2017), exposure time was adjusted on the fly because of changing transparency. This was really annoying !!

 

 

5) I took offset and flat images. For flat images, just put a piece of white paper over the aperture of the telescope. It acts like a diffuser. The source of light is the Sun.

 

 

6) ""Did you ever find an improved method to align your frames?  Druckmüller talks of a phase correlation. I understand the meaning of the words, but I don't understand what means in this context. When I processed with my 2017 TSE images, I tried to align on a prominence on the limb.""

 

This is the BIG question :-)

 

It make the difference between a good and a blurred image.

 

- In 1998, I registred the images manually using the Moon, then offsetting the image in X  and Y to compensate the relative motion of the Sun with respect to the Moon. It took a while to do it, for only 20 images or so. It is no longer practicable with video files :

http://astrosurf.com...upe/process.htm

 

- In 2006, I could register CCD images using a bright star. I have not really check the field for 2024 eclipse, but I think there is no bright star nearby ?

 

- In 2016 and 2017, I used AutoStakkert with only two points of registration on the corona. It was not easy at all to find out the right settings. If prominences are there and not saturated, it would be much easier.

 

Miroslav Druckmüller and Nicolas Lefaudeux are the only guys I know you had successfully developped a software to register the solar corona. They use the Fourrier transform. But just knowing that is not very helpfull ;-)

 

One idea could be to register the frames using the lunar limb (you have to find the center of a circular arc, or may be just use ImPPG). Then you apply an offset in X Y to take in account the relative motion between Moon and Sun. Sounds feasible ...

[ch-viladrich]

7) "Do you know the gain value Druckmüller used?"

 

=> I am not sure I have the info on my computer. You have to check their paper (with Shaddia Habbal). Maybe they gave the information there. I guess they set the gain to the minimum value.

 

 

8) ""Maximizing dynamic range is important for me as well, but I am wondering whether I would be better off  using Gain=100. Please have a look at the ASI6200 specifications.  at Gain>=100, the camera modifies its readout electronic. Thanks to a much lower  readout noise, the dynamic range jumps back to 13.5 stops, with about 1/3 the exposure time.  Gaining 10 db (~3x) means I can record about 3 times as many frames in a fixed amount of time. Does that seem like a good compromise?""

 

=> I would use a gain of 100 too !

[ch-viladrich]

9) "" Is there a silver bullet to solve this problem?""

 

Yes :-)   The trick is to have a stack registered on the Moon disk. Then, because the Moon is not large enough, you resize the Moon. Not very scientific, just for the beauty of the final image.

 

If you don't have the earthshine, you just use an elliptic mask instead of the Moon. Just like this ;-)

http://astrosurf.com...006_masqang.jpg

 

10/ Another point, the corona is visible on the "opposite" limb of contact C2 about 1 min before contact C2. So, if you have a narrow field, you can start imaging the corona before C2 ! But for this, you need a telescope with no straylight.

It is a good idea to check for stray light using the Full Moon. Just massively overexpose the image and check whether you have ghost images in the field of view.

 

 

11) And Nicolas Lefaudeux coroan images :

https://hdr-astropho...-solar-eclipse/

 

Hope this helps !

[Francois314]

We are working on processsing solar eclipse images, I wrote to Nicholas and Druckmuller but they did not respond. Thankfully some of Durckmuller collaborators are very nice and gave me enough info to start coding the algorithm.

The sub pixel alignement using FFT is solved, we are working on local filter to enhance pictures.

If you are interested, join the group: https://www.cloudyni...ing-achf-fnrgf/

Edited by Francois314, 21 August 2023 - 09:34 PM.

[PI_CO100]

Hi Joe,

I stumbled upon this thread while researching TSE settings for the ASI6200MC, a camera I'm excited to get in a few weeks. I'd be particularly keen to learn from your experience, particularly exposure times for the Baily Beads, Diamond Ring, and totality. May I ask you how did things go for you in Australia, and are there any specific settings you found most effective? If you were to do it again, what settings would you use, would you make any changes?

Your expertise in this area would be very helpful, and I'd be really grateful for any insights you could share. Many thanks,

Piero

[banjo1000]

Piero,

Here's a link to my best shot. Details are given with the photo.

https://www.facebook...150121553598457

 

This is a heavily processed image of a series that I combined and stretched heavily. The following is closer to what the raw exposures were like.

 

https://www.facebook...59802289522476/

 

Celestron 8 with a Meade 0.63x reducer (f/6.3 combined), ASI6200MC G=100 2.0ms. This was perhaps a touch too long for prominences. It's a benchmark though that you can use with say, Xavier Jubier's exposure calculator in Solar Eclipse Maestro, to scale to your target for 2024. That's what I am planning to do. 

 

I plan to use the ASI6200MC in 2024 with a C14 and a Hyperstar at F/2. I should be able to go out to almost 6x4 solar radii with much shorter exposures than with my refractors. I am not brave enough to take my solar filter off at F/2 before the start of totality. I will pull the filters off my refractors at -30 secs, gawk at the start of totallity, then de-filter my C14.

 

Joe

[PI_CO100]

Many thanks Joe!

 

This is super-helpful. The pictures you shared are really good, very sharp; the detail shown from the inner corona in the stacked image is up there with anything I have seen. From what you are saying it looks you have used an constant exposure of 2ms all way through the eclipse which we can see works very well for the inner corona. So, using that as benchmark, I can scale up or down for the different phases after adjusting for a slightly different focal lenght.

 

I will be really interested to see your images with the C14 at f2 - it looks like a really a strong setup and your prudence in regard to waiting for totality before removing the filters is really understandable.

 

I will be travelling with a f7.6 4-inch refractor to Mexico and my general plan, given the eclipse will be more than 4 minutes, would be to try to capturing the different areas of the corona so shorter exposures at C2 and C3 but longer in the middle. So perhaps 3 or 4 sets of different exposure times. As I do not have the camera yet, there is a lot of learning, testing and stress for me in the next 96 days...

 

Piero  

[banjo1000]

Piero,

I am going off memory that I will be able to confirm in a few days. I am presenting to the Northern Colorado Astronomical Society (nocoastro.org) tomorrow evening about my research at CERN, and prepping takes precedence, even over Cloudy Nights. There will be a Zoom livestream starting at 6:45 Mountain Time.

 

I didn’t shoot the same exposure throughout. I will confirm, but I used 3 exposure times, clustered in groups of 4 exposures each, and I cycled through this pattern. I used FireCapture, the tool I am most familiar with. I didn’t have a way to to synch with GPS time like one has with Solar Eclipse Maestro and a DSLR. With SEM, if the camera gets bogged down reading out, exposures are skipped. By contrast, FireCapture pauses until there’s room in its buffer. I just cycled the 3x4 sequences. The shorter sequences had better color saturation for prominences, but the longer exposure that I shared had more coronal signal. I composed the image from the best pair of quadruplets of the longest exposure near C2 and C3. If I had to do it over, I would have used my longest exposure time as my shortest. I had to stretch heavily to dig the coronal detail out of the bottom of the bit well.

 

If you figure out a way to sync a ASI6200 readout with a time signal, I would be very interested. I use a pattern similar to the one you described with SEM, a refractor, and a DSLR. For this eclipse, My thinking is to use my C14/Hyperstar for outer corona longer exposures and my 127mm refractor for diamond ring/beads/inner corona/earth shine since I can sync it to GPS time.

 Joe 

[PI_CO100]

Hi Joe,

 

Many thanks again, yes, any more insight into your results would be great  . I will look into your question of synchronising exposures with GPS time when I get the camera (it is on order and should be with me in a couple of weeks), particulary if that can be achieved using Sharpcap scripts. If that is of interest I'd be happy to share my script after is tested with you for comments.

Good luck with your talk today!

Piero

[banjo1000]

Just to be clear, FC can record a time stamp stamp on each frame, but I don’t know of a way to program it to shoot at a particular time.

[Domer]

I'm also preparing for the April 2024 eclipse.  My plan is to use an Esprit 100 w/focal reducer (f/3.6, 358mm) feeding an ASI2600MC camera (APS-C sensor), all mounted on an iOptron CEM40 mount.  Also attached will be a Canon 6D Mark II camera (full frame sensor) and a Sigma 150-600mm lens (f/6.3, 600mm).

 

I have Eclipse Orchestrator Pro and have programmed it to control the DSLR from C1 through C4 so I'm all set with the DSLR.  Now I'm trying to generate a similar sequence in NINA to control the ASI2600MC and have the same concerns as Joe (the OP) regarding exposure times.

 

The partial eclipse exposures are easy.  I simply attached the solar filters to the camera lens and telescope and played with the exposure times until the histograms looked good with no pixels saturated.  But the C2-C3 exposures are more challenging.  I dare not apply the same exposure ratio to the unfiltered exposures because there is no guarantee that the two solar filters have exactly the same attenuation (both are ND5, but I get the feeling that's a pretty coarse measure).

 

So I replaced the filters with a white cloth and pointed to a spot in the sky that was about 10 degrees from the sun on a bright, clear day (avoiding the strong blue bias that comes from pointing to the blue sky).  Starting with the DSLR, I adjusted the exposure to get a good histogram and settled on 0.0025 sec.  Then I did the same with the ASI2600MC and a comparable histogram was achieved with an exposure time of 0.00025 sec.  That's a whopping 10x faster!

 

Here are the stats of those two images from PixInsight (DSLR, then ASI2600MC):

 

 

Part of that is due to the focal ratio - 6.3 on the DSLR vs 3.58 on the astrocam, so the astrocam optical train would be 3x faster just due to the focal ratio.  But the remaining 3.33x is puzzling me.

 

The DSLR sensor's pixel size is 5.75 um and it has a 14-bit ADC.

The ASI2600MC pixel size is 3.76 um and it has a 16-bit ADC.

 

If I just blindly assume that my ASI2600MC exposures for C2-C3 should all be 1/10th of the DSLR, I'm concerned that they will be too short and I might underexpose the diamond ring, Baily's Beads, the chromosphere, and prominances.  The Baily's Beads and Chromosphere will only last about 5 seconds along the centerline so I may not have enough time to do exposure bracketing (2 images max).

 

Does anyone have any direct experience with this or know a better way to compute the exposure ratio for the C2-C3 phases?

 

 

 

[CreatorsHand]

I'm also preparing for the April 2024 eclipse.  My plan is to use an Esprit 100 w/focal reducer (f/3.6, 358mm) feeding an ASI2600MC camera (APS-C sensor), all mounted on an iOptron CEM40 mount.  Also attached will be a Canon 6D Mark II camera (full frame sensor) and a Sigma 150-600mm lens (f/6.3, 600mm).

 

I have Eclipse Orchestrator Pro and have programmed it to control the DSLR from C1 through C4 so I'm all set with the DSLR.  Now I'm trying to generate a similar sequence in NINA to control the ASI2600MC and have the same concerns as Joe (the OP) regarding exposure times.

 

The partial eclipse exposures are easy.  I simply attached the solar filters to the camera lens and telescope and played with the exposure times until the histograms looked good with no pixels saturated.  But the C2-C3 exposures are more challenging.  I dare not apply the same exposure ratio to the unfiltered exposures because there is no guarantee that the two solar filters have exactly the same attenuation (both are ND5, but I get the feeling that's a pretty coarse measure).

 

 

Does anyone have any direct experience with this or know a better way to compute the exposure ratio for the C2-C3 phases?

Maybe; what type/brand of filters are you using? Gordon (foxwoodastronomy) has put out an app "Solar Eclipse Timer" and has a book on photographing the eclipse, and he bases his teaching method based on his experience with Thousand Oaks Optical glass filters, which they don't sell anymore. Based on his experience with those glass filters, the exposure of the full Sun disc that gives a decent exposure of the middle and some limb darkening on the Sun with give a proper exposure for the inner corona without a filter during totality. The problem with that correlation is, unless you know how your filters relate to the Thousand Oaks Optical glass filters, or happen to own one, it is difficult to use that relationship. I have done initial testing of four filters (initial because we've had very little Sun, but I hope to do more the next sunny day): 1)Thousand Oaks Optical glass filter (I found a used one), 2)Spectrum Telescope glass filter, 3)Thousand Oaks SolarLite filter (polymer-purchased before 2017 eclipse) and 4)Thousand Oaks black polymer filter (I purchased the film and made my own in 2016). Testing was done on my 7DII using a 500mm f/4L and 1.4x extender (inadvertently included) at f/5.6. Based on those preliminary tests, relative to the Thousand Oaks glass filter (1/400 sec), the Spectrun glass filter (1/160 sec) needs about 1 1/3 stops more light and the Thousand Oaks Optical SolarLite film (1/50 sec) needs about 3 stops more light to get an equivalent exposure. My next test will be without the 1.4x extender and at f/8, which is what I shot the 2017 eclipse at, and will give me the chance to correlate my filter results this year with the filter results and corona images I captured in 2017. With that information, you may be able to capture a full disc image of the Sun at a proper exposure and be able to roughly correlate what exposure you need for the inner corona by comparing your exposure settings to what I get. Since there is some leeway with capturing the corona, and you can add an additional exposure or two to cover variances in equipment, it may be adequate to estimate what you should shoot at. One thing I am not sure how to correlate between a DSLR and an astro camera is the QE; for most Canon DSLRs (others may vary), the QE seems to be about half that of the cooled astro cameras (+/- 40% vs. +/- 80%). How that correlates between the exposure required for a (Canon) DSLR and a dedicate astro camera, I don't know, but you should be able to correlate well between an astro camera with a filter and the astro camera without a filter during totality if you know how your filter compares to a Thousand Oaks Optical glass filter. I hope that all makes sense. For reference, this is the image I captured yesterday using the above referenced equipment at 1/400 sec, f/5.6, ISO 100 using the Thousand Oaks Optical glass filter. I don't know how the 1.4x extender affects the exposure, which is why I need to retest.

 

Paul

 

 

Thousand Oaks Optical glass filter

[Domer]

Maybe; what type/brand of filters are you using? Gordon (foxwoodastronomy) has put out an app "Solar Eclipse Timer" and has a book on photographing the eclipse, and he bases his teaching method based on his experience with Thousand Oaks Optical glass filters, which they don't sell anymore. Based on his experience with those glass filters, the exposure of the full Sun disc that gives a decent exposure of the middle and some limb darkening on the Sun with give a proper exposure for the inner corona without a filter during totality. The problem with that correlation is, unless you know how your filters relate to the Thousand Oaks Optical glass filters, or happen to own one, it is difficult to use that relationship. I have done initial testing of four filters (initial because we've had very little Sun, but I hope to do more the next sunny day): 1)Thousand Oaks Optical glass filter (I found a used one), 2)Spectrum Telescope glass filter, 3)Thousand Oaks SolarLite filter (polymer-purchased before 2017 eclipse) and 4)Thousand Oaks black polymer filter (I purchased the film and made my own in 2016). Testing was done on my 7DII using a 500mm f/4L and 1.4x extender (inadvertently included) at f/5.6. Based on those preliminary tests, relative to the Thousand Oaks glass filter (1/400 sec), the Spectrun glass filter (1/160 sec) needs about 1 1/3 stops more light and the Thousand Oaks Optical SolarLite film (1/50 sec) needs about 3 stops more light to get an equivalent exposure. My next test will be without the 1.4x extender and at f/8, which is what I shot the 2017 eclipse at, and will give me the chance to correlate my filter results this year with the filter results and corona images I captured in 2017. With that information, you may be able to capture a full disc image of the Sun at a proper exposure and be able to roughly correlate what exposure you need for the inner corona by comparing your exposure settings to what I get. Since there is some leeway with capturing the corona, and you can add an additional exposure or two to cover variances in equipment, it may be adequate to estimate what you should shoot at. One thing I am not sure how to correlate between a DSLR and an astro camera is the QE; for most Canon DSLRs (others may vary), the QE seems to be about half that of the cooled astro cameras (+/- 40% vs. +/- 80%). How that correlates between the exposure required for a (Canon) DSLR and a dedicate astro camera, I don't know, but you should be able to correlate well between an astro camera with a filter and the astro camera without a filter during totality if you know how your filter compares to a Thousand Oaks Optical glass filter. I hope that all makes sense. For reference, this is the image I captured yesterday using the above referenced equipment at 1/400 sec, f/5.6, ISO 100 using the Thousand Oaks Optical glass filter. I don't know how the 1.4x extender affects the exposure, which is why I need to retest.

 

Paul

 

 

Thousand Oaks Optical glass filter

T O Opt Glass-0484G.jpg

Hi Paul,

 

Thanks for the reply. 

 

The telescope filter is made with the Thousand Oaks SolarLite material (supposedly 16 stops) and DSLR filter is made with optical glass (supposedly 16.5 stops).  I've read many of Gordon's posts on the topic of filters and understand his reasoning, but I'm sticking with what I have at this late date.  I'm getting crisp, clear images with both filters, but as Gordon has pointed out, the SolarLite film does have a red bias to it.  I'm OK with that as colors can be adjusted in post.

 

I was also thinking the QE difference in the sensors is a contributing factor.  The ASI2600MC has a QE >80% while the 6D Mark II is only 52%.  So in theory, the ASI2600MC should be capturing 60% more photons, all other things being equal.  But that would only account for part of the discrepancy.  If the focal ratio accounts for 3x and the QE accounts for another 1.6x, that leave roughly a 2x actor still to be explained.

 

I was wondering if the difference in sensor pixel size was the missing piece but the DSLR pixels have 2.33 times the area of the ASI2600MC so if anything that should make the DSLR faster, not slower.

 

Another test I could run (once the current atmospheric river passes us by) is to do an exposure comparison with both optical trains using the telescope (SolarLite) filter.  I can't use the smaller glass filter on the telescope, but I can secure the larger telescope filter on the camera lens.  That would give me an apples-to-apples comparison.  However, I don't expect those results would differ much from what I did using the white cloth and pointing to the bright sky (which is what yielded the 10x factor). It's worth running the test in any case.  If the results are similar, that adds credibility to conclusion.

[Jquitadamo]

Hi,

Has anyone here been able to compare gain on 2600/6200/533 chip to DSLR ISO yet using daytime patch of sky and Sharpcap/Firecap histogram? I read somewhere that gain 0 or 1 is similar to ISO 6 or 12, but that has NOT been my experience when testing with ND 5.0 solar film, however I don't have a DSLR to use as a true head-to-head comparison.

Thanks,
Jason

[R Botero]

Hi,

Has anyone here been able to compare gain on 2600/6200/533 chip to DSLR ISO yet using daytime patch of sky and Sharpcap/Firecap histogram? I read somewhere that gain 0 or 1 is similar to ISO 6 or 12, but that has NOT been my experience when testing with ND 5.0 solar film, however I don't have a DSLR to use as a true head-to-head comparison.

Thanks,
Jason

Jason

 

If you are trying to figure out exposure brackets, why not start with you ND5.0 filter, aim to capture the full disc and note your settings (use Gain 0) for the exposure necessary for your system (aim for 1/2 to 2/3 histogram).  Then go to Xavier Jubiert's exposure page - for example - and enter the conditions you are expecting for the eclipse in terms of Sun altitude and the f/number of your imaging system.   By playing with the ISO as variable, you can match what ISO number replicates the exposure you needed for the partial phase of the eclipse using an ND5.0 filter.  

 

For my 491mm EFL scope and ASI2600MC, I need 1.60-2.00ms using a Baader ND5.0 filter.   I worked out my other exposure settings from that.

 

Roberto

Edited by R Botero, 18 March 2024 - 08:02 AM.

[Jquitadamo]

Jason

 

If you are trying to figure out exposure brackets, why not start with you ND5.0 filter, aim to capture the full disc and note your settings (use Gain 0) for the exposure necessary for your system (aim for 1/2 to 2/3 histogram).  Then go to Xavier Jubiert's exposure page - for example - and enter the conditions you are expecting for the eclipse in terms of Sun altitude and the f/number of your imaging system.   By playing with the ISO as variable, you can match what ISO number replicates the exposure you needed for the partial phase of the eclipse using an ND5.0 filter.  

 

For my 491mm EFL scope and ASI2600MC, I need 1.60-2.00ms using a Baader ND5.0 filter.   I worked out my other exposure settings from that.

 

Roberto

 

Hi Roberto,

 

That's EXACTLY what I did a few days ago, and essentially why I'm asking this question. 

 

I got very similar results as you (1.0-1.5ms @f/4.7 using baader film). The equivalent ISO (~50-75) was much higher than I was expecting. I read somewhere that gain 0 for 2600 should be closer to ISO 6-12, but source was pretty vague. Either way, I think a problem with this method is that it assumes ND5 film is near-equivalent across brands, but it appears this is not the case, especially for Baader film.

 

I read/heard from multiple sources that Baader film is "faster" (lets through more light) than non-Baader film or glass. If this is true, which appears to be based on my research, cross referencing exposure times for DR and BB, using unobstructed sun and baader ND5 as the reference, may be significantly off.

 

These findings were enough to convince me to pull the trigger on a glass filter, at least to do a comparison test. Just trying to tighten up the estimated exposure times for those phases since bracketing during BB is a limited option.

 

Thanks,

Jason

Edited by Jquitadamo, 18 March 2024 - 08:23 PM.

[Domer]

I read/heard from multiple sources that Baader film is "faster" (lets through more light) than non-Baader film or glass. If this is true, which appears to be based on my research, cross referencing exposure times for DR and BB, using unobstructed sun and baader ND5 as the reference, may be significantly off.

I can tell you from personal experience and a lot of testing that the Baader film is faster than the Thousand Oaks film.  Here is a histogram of an image of the sun taken with my Canon 6D Mark II at 1/500th second using a Thousand Oaks "SolarLite" filter:

 

 

And taken the same day with the same equipment but using a Baader filter:

 

 

Not only does the Baader filter pass a lot more light, but the RGB channels are much more even, giving a "whiter" result.  So no, not all ND5 filters are created equal.

[R Botero]

Hi Roberto,

 

That's EXACTLY what I did a few days ago, and essentially why I'm asking this question. 

 

I got very similar results as you (1.0-1.5ms @f/4.7 using baader film). The equivalent ISO (~50-75) was much higher than I was expecting. I read somewhere that gain 0 for 2600 should be closer to ISO 6-12, but source was pretty vague. Either way, I think a problem with this method is that it assumes ND5 film is near-equivalent across brands, but it appears this is not the case, especially for Baader film.

 

I read/heard from multiple sources that Baader film is "faster" (lets through more light) than non-Baader film or glass. If this is true, which appears to be based on my research, cross referencing exposure times for DR and BB, using unobstructed sun and baader ND5 as the reference, may be significantly off.

 

These findings were enough to convince me to pull the trigger on a glass filter, at least to do a comparison test. Just trying to tighten up the estimated exposure times for those phases since bracketing during BB is a limited option.

 

Thanks,

Jason

Jason

 

Does it really matter what the "speed" of the solar filter is if you can obtain your own set of exposures working back from Xavier's page?  What I mean is that Xavier is only providing a guideline.  You are trying to figure out what the equivalent ISO is for your CMOS camera by matching what a good partial eclipse picture should look like.  The solar filter is only an attenuator, you are only interested in a similar result from an equivalent DSLR, not what filter they used.  The variable is not the filter, it's the ISO.  Once you - broadly - match the ISO (you will get an exact match), you bracket around the suggested exposures ...and hope for the best.

 

Roberto

[Jquitadamo]

Hi Roberto,

 

I think it matters, but I could be wrong. I'll attempt to explain using fake numbers...

 

Let's assume Baader ND5 film requires 1 ms exposure with my setup, which works out to be ISO 50. That would mean I should shoot BB at 0.125 ms using Xavier's calculator.

 

Now let's assume Spectrum ND5 glass requires 8 ms exposure, which works out to be ISO 6. That would mean I should shoot BB at 1 ms using Xavier's calculator.

 

That's an 8x difference (3 stops) in exposure, depending on what the real ISO equivalent is for the 2600. 

 

My experimental results using Baader ND5 tells me that equivalent ISO is ~50-70, but I've also read anecdotally that these cameras are closer to ISO 6-12 at gain 0. 

 

I'm just trying to get a better estimate for the true 2600 equivalent ISO at gain 0 so I can shoot BB without bracketing.

 

I purchased a Spectrum glass filter and will report back my findings within a week.

 

Thanks,

Jason

Edited by Jquitadamo, 19 March 2024 - 12:09 PM.

[R Botero]

I think the lower ISO you are referencing is from Baader’s website here.  I suspect this is an estimate derived from camera measurements and will likely be model dependent. I note none of the cameras Baader tested had the ASI2600 sensor. It may be easier to test one of those models, for example the 533 with an ND5 filter and see if the ISO matches Xavier’s exposure. 
 

Roberto