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

Monochrome vs Color camera integration time

astrophotography imaging
  • Please log in to reply
27 replies to this topic

#1 Nemesis_318

Nemesis_318

    Explorer 1

  • *****
  • topic starter
  • Posts: 62
  • Joined: 19 Jan 2008
  • Loc: Michigan

Posted 18 July 2020 - 11:13 AM

I am debating on whether to purchase a monochrome or color camera and weighing the pros and cons of each.  An important consideration is exposure time as I have very little free time for astrophotography.    

 

So my question is, if one were to shoot monochrome with LRGB filters,   1 hour for each filter for a total of 4 hours of exposure time - is this equivalent to 1 hour of color camera exposure or 4 hours of color camera exposure or neither?

 

Thanks

 



#2 jdupton

jdupton

    Soyuz

  • *****
  • Posts: 3,959
  • Joined: 21 Nov 2010
  • Loc: Central Texas, USA

Posted 18 July 2020 - 11:29 AM

Nemesis_318,

 

 

1 hour for each filter for a total of 4 hours of exposure time - is this equivalent to 1 hour of color camera exposure or 4 hours of color camera exposure or neither?

 

   The answer is generally "neither." Mono cameras shooting LRGB will outperform an OSC in terms of signal (captured photons) per unit time. The amount by which mono outperforms increases rapidly the more "L" you shoot.

 

   The particular comparison you gave highlights the argument that is often mis-applied. The claims that Mono is more than 4x faster than OSC often stems from allowing 4 hours of imaging time with the Mono (1 hour per filter for a total of four filters = 4 hours) beats an OSC being exposed for 1 hour. Of course it does! My personal Mono camera will outperform every other Mono camera if you let me shoot for 4 hours and I allow you to shoot for only 1 hour with either a mono or OSC. tongue2.gif

 

   In the end, for equal amounts of time under the sky, the Mono with RGB filters will be on par with an OSC with the same total integration time. The OSC loses out on any competition as soon as you start adding "L" exposure time for the mono. The more L you add to the RGB mix, the bigger the win.

 

 

John


Edited by jdupton, 18 July 2020 - 11:33 AM.

  • 44maurer and OldManSky like this

#3 endless-sky

endless-sky

    Apollo

  • -----
  • Posts: 1,014
  • Joined: 24 May 2020
  • Loc: Padova, Italy

Posted 18 July 2020 - 11:51 AM

There's a very interesting thread in this forum that covers in detail the advantages of a mono camera vs an OSC camera.

 

This: https://www.cloudyni...numbers-please/

 

In synthesis, it depends. But, as said by John, where the mono outperforms the OSC is by imaging with an L filter. Other things to consider: mono doesn't have the Bayer matrix cutting down the efficiency of the camera on each channel.

 

I used to think that shooting with mono would take longer than shooting with an OSC. After reading that thread, I know that when I will upgrade from my DSLR to a dedicated camera, I most likely will buy a monochromatic one. There's just no contest between the two.


  • jdupton likes this

#4 sg6

sg6

    Voyager 1

  • *****
  • Posts: 10,184
  • Joined: 14 Feb 2010
  • Loc: Norfolk, UK.

Posted 18 July 2020 - 11:51 AM

Again Neither.

20 Mega Pixel Mono is 20 Mega Pixels in the filter "color".

20 Mega Pixel Color is 10 Mega Pixels Green, 5 Mega Pixels Red and 5 Mega Pixels Blue.

 

Makes a direct comparison sort of irrelevant.



#5 jdupton

jdupton

    Soyuz

  • *****
  • Posts: 3,959
  • Joined: 21 Nov 2010
  • Loc: Central Texas, USA

Posted 18 July 2020 - 12:05 PM

Nemesis_318,

 

   I also forgot to add that the differences in exposure time between mono and OSC depends heavily on the particular cameras you are comparing and the particular filter set you use for the mono. The thread that @endlessky referenced has a tool that allows you to make such comparisons pretty easily.

 

 

endlessky,

 

 

Other things to consider: mono doesn't have the Bayer matrix cutting down the efficiency of the camera on each channel.

   Very true. However, the mono is uses an external filter to cut down the efficiency  as you capture each individual color channel. Remember that a mono camera's external filter is always blocking roughly 2/3 of the light coming from the sky. smile.gif (The Red filter, for example, blocks all wavelengths from the green and blue portions of the spectrum by necessity.)

 

 

John



#6 Wintermute

Wintermute

    Vostok 1

  • *****
  • Posts: 180
  • Joined: 06 Feb 2010
  • Loc: Ramona, CA

Posted 18 July 2020 - 02:24 PM

If you have very little time for the hobby I'd strongly recommend a color camera. Monochrome often involves shooting a target over multiple nights, with color even if clouds roll in or you need to break down your equipment early you can process the data you captured and get a workable image. Processing monochrome also takes longer; you'll need to integrate multiple channels of data and correct these individually before they are combined. Anyways just my $0.02, I have both color and monochrome cameras and see the benefits of both. 



#7 Nemesis_318

Nemesis_318

    Explorer 1

  • *****
  • topic starter
  • Posts: 62
  • Joined: 19 Jan 2008
  • Loc: Michigan

Posted 18 July 2020 - 03:27 PM

Thanks for all the replies and information.  Reading through this information as well as the link above there are many comments saying monochrome is the winner.  would a layperson be able to tell the difference between a 'better' photo taken with monochrome and filters as compared to a color camera - or is this a slight difference only noticeable by experienced astrophotographers ?

 

I live in Michigan so we have long stretches of cloudy weather for days at a time - even weeks at a time in the winter.  Color seems like it might be the better choice for me here due to the weather and limited time available for the hobby.  I'm looking at a ZWO ASI294MC Pro cooled camera with possibly an Optolong L-enhance filter.   My skies are Bortle 5ish and I have an Orion 80ED.  Would I get good results with this camera in Bortle 5?

 

 

One other question.  I have been shooting with a Canon 7D unmodified.  I also use this camera for bird photography and work so I never had it modified for astrophotography.    My understanding is the modded cameras have the IR cut filter removed.   If it's advantageous to remove the IR cut filter, why do manufacturers make IR cut filters for astrophotography?



#8 Peregrinatum

Peregrinatum

    Surveyor 1

  • *****
  • Posts: 1,820
  • Joined: 27 Dec 2018
  • Loc: South Central Valley, Ca

Posted 18 July 2020 - 03:49 PM

RGB filters will give higher SNR than a OSC for the same total integration time...

 

if time is limited, get a mono camera and use RGB filters and an synthetic Luminance



#9 klaussius

klaussius

    Apollo

  • *****
  • Posts: 1,449
  • Joined: 02 Apr 2019
  • Loc: Buenos Aires

Posted 18 July 2020 - 03:51 PM

A key factor you haven't mentioned is your location. How much light pollution are you subjected to? Can you regularly get to a dark site?

 

An OSC can work very well in dark sites, but with a lot of light pollution you'll want what mono can give you: fast data gathering with L, efficient and high-quality narrow band.

 

I don't think time is an issue. I used to think that, but I'm forced to spend days on targets with a DSLR just the same, because the DSLR is slow at gathering data. So days with a DSLR, days with a mono sensor, if I'm going to spend days in any case, I prefer the mono sensor.

 

Also, if you are under heavy light pollution, you should really go with narrowband. Every time I take my H-alpha filter I'm pleasantly surprised at how much easier capture is with it. It has its challenges, but the data comes out so clean in comparison with OSC data, and is a pleasure to work with.

 

If you live at or near a reasonably dark site, however, things may shift a bit. OSC at a dark site can quickly produce amazing results. Mono will just as well, but if you have a dark site nearby you have to travel to, OSC may have strong pros in simplicity, especially if you have to move your gear around often.



#10 WadeH237

WadeH237

    Cosmos

  • *****
  • Posts: 8,034
  • Joined: 24 Feb 2007
  • Loc: Ellensburg, WA

Posted 18 July 2020 - 05:04 PM

would a layperson be able to tell the difference between a 'better' photo taken with monochrome and filters as compared to a color camera

The skill of the person processing the data is far more important in what the final result looks like, than "mono vs. OSC".

 

The layperson would find the image done with better processing to be the better image.  If you have the same person process two sets of data, one from a mono camera with filters, and one with a OSC camera, a layperson would likely be unable to tell the difference in the final result.

 

In terms of getting the S/N in the shortest time, the mono camera will probably win, depending on how much time overhead is involved with changing filters, etc.  But the difference for comparing simple RGB images between them is going to be minor.  The mono camera starts to make lots more sense when you do LRGB or narrow band imaging.  The two big advantages of the OSC camera over mono/filters, is cost.  A OSC camera often costs the same as a mono camera with the same sensor.  But for the mono camera, you need to add the cost of a filter wheel and filters (and for many of the current CMOS cameras, a good set of filters can easily cost more than the camera).


  • jdupton likes this

#11 calypsob

calypsob

    Fly Me to the Moon

  • *****
  • Posts: 7,193
  • Joined: 20 Apr 2013
  • Loc: Virginia

Posted 18 July 2020 - 05:18 PM

Nemesis_318,

 

 

   The answer is generally "neither." Mono cameras shooting LRGB will outperform an OSC in terms of signal (captured photons) per unit time. The amount by which mono outperforms increases rapidly the more "L" you shoot.

 

   The particular comparison you gave highlights the argument that is often mis-applied. The claims that Mono is more than 4x faster than OSC often stems from allowing 4 hours of imaging time with the Mono (1 hour per filter for a total of four filters = 4 hours) beats an OSC being exposed for 1 hour. Of course it does! My personal Mono camera will outperform every other Mono camera if you let me shoot for 4 hours and I allow you to shoot for only 1 hour with either a mono or OSC. tongue2.gif

 

   In the end, for equal amounts of time under the sky, the Mono with RGB filters will be on par with an OSC with the same total integration time. The OSC loses out on any competition as soon as you start adding "L" exposure time for the mono. The more L you add to the RGB mix, the bigger the win.

 

 

John

But you also have to remember that if you go to a dark site and it clouds up after 2 hours, with OSC you have 2 hours of data, with LRGB mono you have whatever filters you got through before 2 hours are up.  That has always been my biggest hesitation to go full mono. 


  • Wintermute likes this

#12 Stelios

Stelios

    Voyager 1

  • *****
  • Posts: 11,123
  • Joined: 04 Oct 2003
  • Loc: West Hills, CA

Posted 18 July 2020 - 05:36 PM

You have an Orion ED80, an F/7.5 scope. If speed of acquisition is your goal, you'd be better off going to an F/6 scope with a reducer bringing it to F/4.8--this will get signal 2.5 times faster. You may also want to consider getting a reflector around F/4, though that can be an adventure if you've never used one.

 

At Bortle 5 you should be OK with an OSC in general. Mono offers you unmatched flexibility as well as speed, but mono is expensive. If you do go the mono way, the best deal is the ZWO ASI1600MM-Pro kit with EFW8 and filters. 



#13 Alex McConahay

Alex McConahay

    Voyager 1

  • *****
  • Posts: 11,006
  • Joined: 11 Aug 2008
  • Loc: Moreno Valley, CA

Posted 18 July 2020 - 07:50 PM

In a situation where you do not have lots of time to image, and you just want to take some images, a OSC camera has advantages. You can run fifteen five minute exposures in a session lasting an hour and a half, and go home, and have something to process. You might have something with LRGB, but you certainly would have something with OSC. 

 

Most of us would, however, prefer longer sessions, and the longer the session, the more LRGB will pull ahead. 

 

I went for a hike today. Took some snaps with my phone camera. Nice pics. I did not take my full frame DSLR, Tripod, and set of lenses. I got good memories of Mt. San Gorgonio, and the desert, and Mt. San Jacinto in the distance. I got nothing that I could blow up to 20 x 30, with deep colors, tack sharp details, great contrast. But I got some shots. 

 

Mono/LRGB just has a higher top end. 

 

(Now, don't get me wrong. I did not say you CANNOT take great pics with a OSC astro camera, or a modern cell phone. I said that if your goal is to get the best, you will want to use the higher end equipment. If you could do that  pic with a OSC or cell phone, you could do that and more with the better equipment. ) 

 

Alex 



#14 RogueZero

RogueZero

    Lift Off

  • *****
  • Posts: 21
  • Joined: 09 Mar 2020
  • Loc: East Bay, CA

Posted 18 July 2020 - 10:36 PM

Shouldn't mono also have greater resolution than OSC? RGB Mono captures color signal on each pixel while OSC spreads it across multiple pixels. 



#15 klaussius

klaussius

    Apollo

  • *****
  • Posts: 1,449
  • Joined: 02 Apr 2019
  • Loc: Buenos Aires

Posted 18 July 2020 - 11:00 PM

Shouldn't mono also have greater resolution than OSC? RGB Mono captures color signal on each pixel while OSC spreads it across multiple pixels. 

The lack of a CFA makes data cleaner and easier to process, but you can still extract a lot of resolution from an OSC through bayer drizzle and image deconvolution. If you have enough integration time and good enough dithering to make that viable.



#16 Alex McConahay

Alex McConahay

    Voyager 1

  • *****
  • Posts: 11,006
  • Joined: 11 Aug 2008
  • Loc: Moreno Valley, CA

Posted 19 July 2020 - 01:22 AM

The resolution is the same in one respect. Assuming identical sensors, with the only difference being the bayer matrix. Each pixel is the same size, which is the first determinant of resolution.

In another respect, it is different.

Each pixel reports a value. But in a mono sensor the value is from that pixel only. For a color, the value in some respects is derived from the other pixels in the area, and thus may be somewhat compromised.

Alex

#17 endless-sky

endless-sky

    Apollo

  • -----
  • Posts: 1,014
  • Joined: 24 May 2020
  • Loc: Padova, Italy

Posted 19 July 2020 - 01:45 AM

The lack of a CFA makes data cleaner and easier to process, but you can still extract a lot of resolution from an OSC through bayer drizzle and image deconvolution. If you have enough integration time and good enough dithering to make that viable.

But if you give the dithering/drizzle and deconvolution advantage to the OSC, what's stopping you to do the same to the mono? If the goal is get the optimal data, then it's only fair we give the optimal pre and post-processing to both camera types, and then compare.

 

I guess that what I am trying to say is: ALL things equal, except for some very rare circumstances described on the thread I linked, the monochromatic version of the same camera will most of the times produce the better results than the OSC version. And, let's keep this in mind, since the misconception is that capturing data with a mono camera takes longer (until I read that thread, I used to think that, too), with the same total integration or less.



#18 darkcloud

darkcloud

    Vostok 1

  • *****
  • Posts: 131
  • Joined: 04 Feb 2006
  • Loc: Central NC

Posted 19 July 2020 - 09:00 AM

"If it's advantageous to remove the IR cut filter, why do manufacturers make IR cut filters for astrophotography?"

 

Depends on what you are interested in photographing.  If you are interested in visual, the then IR cut filter is needed to keep the IR from overwhelming the visual.  If you are interested in photographing in the IR, then you don't use the filter.



#19 Alex McConahay

Alex McConahay

    Voyager 1

  • *****
  • Posts: 11,006
  • Joined: 11 Aug 2008
  • Loc: Moreno Valley, CA

Posted 19 July 2020 - 10:47 AM

>>>>>"If it's advantageous to remove the IR cut filter, why do manufacturers make IR cut filters for astrophotography?"

 

Are the "IR cut" filters the same? IR extends for a ways. Perhaps you want some IR, but not too much. 

 

Alex



#20 endless-sky

endless-sky

    Apollo

  • -----
  • Posts: 1,014
  • Joined: 24 May 2020
  • Loc: Padova, Italy

Posted 19 July 2020 - 11:29 AM

One other question.  I have been shooting with a Canon 7D unmodified.  I also use this camera for bird photography and work so I never had it modified for astrophotography.    My understanding is the modded cameras have the IR cut filter removed.   If it's advantageous to remove the IR cut filter, why do manufacturers make IR cut filters for astrophotography?

The stock filter(s) that usually come with the DSLRs actually do more than just block the IR. Camera sensors are much more sensitive and to a broader range of the light spectrum than the human eye. DSLRs are typically used for every day, terrestrial photographs, so the manufacturers put a filter that blocks much of the red spectrum, since the human eye is more sensitive to the green part of the spectrum. This way the photographs produced by the cameras better resemble what the human eye is used to perceive.

 

Unfortunately, these filters block (or cut by a huge percentage) a very important part of the spectrum: the H-alpha, which is very common on many nebulae that the astrophotographers wish to photograph. That's why for astrophotographic purposes the stock filters are removed (not for the IR part).

 

On the other hand, completely removing the filter allows for other parts of the spectrum (other than the wanted H-alpha) that are not wanted or even detrimental for astrophotography: the UV and the IR (as I previously said, the sensors are natively very sensitive to a broad range of the light spectrum). These parts of the spectrum, if allowed to pass through and get captured by the sensor, cause the stars to be bloated or to have halos around them (think achromatic refractors), since not all incoming light gets focused by a lens in the same focal plane. UV and IR being the worst, as they are on the far left of the blue and the far right of the red, respectively. In order to complete the astromodification of the camera a UV/IR cut filter is placed back in front of the sensor to block these unwanted effects.

 

TLDR: for astrophotography H-alpha is good, UV/IR are not*

 

*unless you are interested in ONLY the UV or the IR part of the spectrum, in which case only one of these parts would be captured (using a UV pass only or a IR pass only filter), essentially negating the focusing problem



#21 klaussius

klaussius

    Apollo

  • *****
  • Posts: 1,449
  • Joined: 02 Apr 2019
  • Loc: Buenos Aires

Posted 19 July 2020 - 12:57 PM

But if you give the dithering/drizzle and deconvolution advantage to the OSC, what's stopping you to do the same to the mono? If the goal is get the optimal data, then it's only fair we give the optimal pre and post-processing to both camera types, and then compare.

 

I guess that what I am trying to say is: ALL things equal, except for some very rare circumstances described on the thread I linked, the monochromatic version of the same camera will most of the times produce the better results than the OSC version. And, let's keep this in mind, since the misconception is that capturing data with a mono camera takes longer (until I read that thread, I used to think that, too), with the same total integration or less.

Bayer drizzle can still upsample and get the same improvements of mono drizzle, if you've got enough integration time and are undersampled.

 

What's different with bayer drizzle is that it sorts out pixels according to the CFA. It's sorting out pixels according to which filter was in fron of them during integration, separating the CFA into 3 mono images. Each sub has 25% R pixels, 25% B pixels and 50% G pixels, but in the end you end up covering the whole image with R G and B.

 

So it's equivalent to RGB mono, with the caveat that the CFA isn't such a good filter in comparison with interference RGB filters, so that's where you lose some resolution/information. But it's otherwise very close to what you can get with mono.

 

The big difference is in the quality of the underlying data itself. The CFA interferes optically with the signal. You need a matrix operation to properly separate the colors because the CFA bands overlap considerably, and that induces color noise and artifacts around saturated pixels. Most CFAs have an OLPF that blurs the image to avoid moire. It can be overcome with deconvolution but doing so amplifies noise, requiring more integration time to compensate. Etc.

 

Then, if you try to do NB with an OSC, you're double-filtering light, reducing QE. It's not impossible, I do it with my DSLR and enjoy it, but it's inefficient. If you aim to do NB and can choose, a mono is a better choice. If constrained by other factors, an OSC will do the job.


  • jdupton likes this

#22 TelescopeGreg

TelescopeGreg

    Soyuz

  • -----
  • Posts: 3,957
  • Joined: 16 Jul 2018
  • Loc: Auburn, California, USA

Posted 19 July 2020 - 02:01 PM

I think one can look at it in relative terms.  (At least, that's part of my plan's justification...)

 

My DSLR, unmodified, has a QE of about 50%, and an CFA that pretty much makes Ha invisible.  The plan is to swap it for an ASI2600MC Pro.  QE is 80%, and it's got good sensitivity to Ha.  And, it's cooled.  If I were to put a NB Ha in front, I'd still come out ahead.

 

Can I do better?  Sure.  But I think this is a big enough step forward in many dimensions to be worth doing.  Going further than that, into full-blown Mono imaging, is more than I want to take on right now.



#23 jdupton

jdupton

    Soyuz

  • *****
  • Posts: 3,959
  • Joined: 21 Nov 2010
  • Loc: Central Texas, USA

Posted 19 July 2020 - 02:25 PM

klaussius,

 

So it's equivalent to RGB mono, with the caveat that the CFA isn't such a good filter in comparison with interference RGB filters, so that's where you lose some resolution/information. But it's otherwise very close to what you can get with mono.

 

The big difference is in the quality of the underlying data itself. The CFA interferes optically with the signal. You need a matrix operation to properly separate the colors because the CFA bands overlap considerably, and that induces color noise and artifacts around saturated pixels. Most CFAs have an OLPF that blurs the image to avoid moire. It can be overcome with deconvolution but doing so amplifies noise, requiring more integration time to compensate. Etc.

   This portion of your post confuses me.

 

   First, can you outline why CFA filters are not as good as interference filters? In what ways are they inferior? How does the mere presence of the CFA filters cause a problem? I do realize that many think CFA filters are very inefficient but that is mostly because of incorrect reading of the sensor plots vendors give us. See the following post for what I think is the right way to read the data we are given concerning CFA filter responses.

 

https://www.cloudynights.com/topic/715122-osc-in-redwhite-zone-lost-cause/?view=findpost&p=10303177

 

   The usual deBayer interpolation we often blindly use does cause a loss of color resolution. While color resolution suffers, luminance / intensity resolution is affected to a much lesser amount. In other words, imaging two stars close together will still show them resolved but their colors could be smeared due to normal deBayer color interpolation. However, it is the deBayer interpolation used to recover the color that is causing this loss of color resolution not the CFA filters themselves. When you use CFA Drizzle, there is no longer the same type of "smearing" interpolation going on.

 

   Regarding the statement "The CFA interferes optically with the signal.": can you explain to me how the presence of a CFA filter over one pixel introduces optical interference with adjacent pixels? And why would the same thing not happen on a pixel without a filter? At the pixel level, I cannot envision a process by which one pixel affects another just because of a filter covering it. Any detrimental effects would only seem to happen when you go to recover the color information for this pixel using its neighbors, to the best of my understanding. Using CFA Drizzle, we avoid that dependence on the neighbors to get a pixel's color information.

 

   Along the same lines of though, why would an overlap of filter responses cause loss of information. In my way of looking at it, the overlap enhances the information recovered. It actually aids the process by helping with the color fidelity of the data. I think of this as a case where "what CFA can lose in color resolution, it gains in color fidelity."

 

   An example of this might be to look at a particular emission line in a target. You can choose something like Oiii or Ha. With non-overlapping interference filters you capture the Ha line with the red filter only. With the CFA filters, you capture the same Ha line with a combination of red , green (~11%), and blue (~4%) filters. Now reconstruct the color via Drizzle for both sets of images. The interference filtered image only has data in the red channel. When displayed, the line will look whatever color of pure red your monitor can (or chooses to) produce. However, the CFA filtered image will present a color that is closer to how the eye sees the Ha emission line with a proper mixture of colors. The fidelity of the Ha reproduction is better because of the overlap of filter responses which more closely resemble the eye's response.

 

   There was another thread debating the benefits and issues associated with filter band overlap a few months ago. This topic is often just as hotly debated as many others here on CloudyNights.

 

https://www.cloudynights.com/topic/591704-overlapping-rgb-filters-and-colour-accuracy/

 

   Some day I would like to find the time to run an experiment as outlined in these other threads. I have both Mono and OSC cameras. I would like to take a bright light source and shine it through a prism. Then, the resulting rainbow would be photographed with both my Mono camera and AstroDon LRGB (i-Series) filters and again with my ASI294MC-Pro OSC camera. I strongly suspect that the rainbow simply cannot be reproduced cleanly (or pleasantly) with the Mono and filters because there just isn't enough overlap in the interference filter band-passes. The result would likely look very much "color posterized" compared to the result from the OSC camera. Without a good bit of PixelMath blending, non-overlapping interference filters cannot produce smooth tonal ranges. This just doesn't matter if you are going to be creating synthetic pallet images from the interference filtered images, however.

 

 

John



#24 klaussius

klaussius

    Apollo

  • *****
  • Posts: 1,449
  • Joined: 02 Apr 2019
  • Loc: Buenos Aires

Posted 19 July 2020 - 02:26 PM

I think one can look at it in relative terms.  (At least, that's part of my plan's justification...)

 

My DSLR, unmodified, has a QE of about 50%, and an CFA that pretty much makes Ha invisible.  The plan is to swap it for an ASI2600MC Pro.  QE is 80%, and it's got good sensitivity to Ha.  And, it's cooled.  If I were to put a NB Ha in front, I'd still come out ahead.

 

Can I do better?  Sure.  But I think this is a big enough step forward in many dimensions to be worth doing.  Going further than that, into full-blown Mono imaging, is more than I want to take on right now.

Exactly.

 

And also there aren't many options in mono sensors that match the 2600MC specs (in that price range). You'd be forced to go all the way to full-frame for quite a higher price, not to mention filters, CFW, a scope that can illuminate a full frame sensor, etc.

 

So it bears repeating, mono has quality as an advantage, OSC has price.



#25 klaussius

klaussius

    Apollo

  • *****
  • Posts: 1,449
  • Joined: 02 Apr 2019
  • Loc: Buenos Aires

Posted 19 July 2020 - 03:16 PM

klaussius,

 

   This portion of your post confuses me.

 

   First, can you outline why CFA filters are not as good as interference filters? In what ways are they inferior? How does the mere presence of the CFA filters cause a problem? I do realize that many think CFA filters are very inefficient but that is mostly because of incorrect reading of the sensor plots vendors give us. See the following post for what I think is the right way to read the data we are given concerning CFA filter responses.

That may be true, but to recover saturation you have to apply a transformation to the signal that amplifies color noise. That's not a minor hindrance, and it effectively reduces SNR.

 

While you may be collecting lots of photons, not classifying them accurately is a form of error. As with deconvolution, recovering from that mixup amplifies noise as a tradeoff.

 

 

       Regarding the statement "The CFA interferes optically with the signal.": can you explain to me how the presence of a CFA filter over one pixel introduces optical interference with adjacent pixels? And why would the same thing not happen on a pixel without a filter? At the pixel level, I cannot envision a process by which one pixel affects another just because of a filter covering it. Any detrimental effects would only seem to happen when you go to recover the color information for this pixel using its neighbors, to the best of my understanding. Using CFA Drizzle, we avoid that dependence on the neighbors to get a pixel's color information.

The CFA and the OLPF are a combo. The OLPF defocuses the image ever so slightly, reducing resolution. On purpose, to avoid color aliasing in the CFA. It can also cause other aberrations, like CA, although most high-end sensors I've seen are free from these defects.

 

 

Along the same lines of though, why would an overlap of filter responses cause loss of information. In my way of looking at it, the overlap enhances the information recovered. It actually aids the process by helping with the color fidelity of the data. I think of this as a case where "what CFA can lose in color resolution, it gains in color fidelity."

That's probably a contentious point. As I already mentioned, the matrix transform amplifies noise, and that decreases SNR. That's a fact I battle with every day. Does that completely undo the increased SNR from overlapping bands? Hard to tell.

 

Also, whether you want the overlap in astro imaging is a debatable thing. While it is true that given a monochromatic light source, overlapping filters retain spectral information, sources contributing a single pixel are rarely monochromatic. In order to recover that information you have to make assumptions about the source.

 

If you're going to make assumptions, you can also make useful assumptions with mono. Emission nebulae are narrowband, and usually emit Ha, Hb, SII, OIII. Interpreting nonoverlapping RGB into those bands isn't hard, even splitting Ha from SII may be possible, since Ha and Hb are usually correlated. Reflection nebulae and stars are broadband, which reflect/emit in the black body spectrum. Given RGB ratios it's not hard to extrapolate the full spectrum. In fact that's what PCC does.

 

My point is, that trick is not exclusive to OSC. It may just take a different form in mono.

 

 

An example of this might be to look at a particular emission line in a target. You can choose something like Oiii or Ha. With non-overlapping interference filters you capture the Ha line with the red filter only. With the CFA filters, you capture the same Ha line with a combination of red , green (~11%), and blue (~4%) filters. Now reconstruct the color via Drizzle for both sets of images. The interference filtered image only has data in the red channel. When displayed, the line will look whatever color of pure red your monitor can (or chooses to) produce. However, the CFA filtered image will present a color that is closer to how the eye sees the Ha emission line with a proper mixture of colors. The fidelity of the Ha reproduction is better because of the overlap of filter responses which more closely resemble the eye's response.

When I do Ha, I usually colorize the Ha channel to synthesize Hb. Similarly, if you have pure Ha data you can easily mix it in RGB to match your monitor gamut. Beware that depends a lot on your monitor and may not produce the same effect in other monitors.

 

I find that process a lot more understandable and flexible than depending on the overlapping responses of the CFA to do it on its own. That's of course an opinion.

 

 

Some day I would like to find the time to run an experiment as outlined in these other threads. I have both Mono and OSC cameras. I would like to take a bright light source and shine it through a prism. Then, the resulting rainbow would be photographed with both my Mono camera and AstroDon LRGB (i-Series) filters and again with my ASI294MC-Pro OSC camera. I strongly suspect that the rainbow simply cannot be reproduced cleanly (or pleasantly) with the Mono and filters because there just isn't enough overlap in the interference filter band-passes. The result would likely look very much "color posterized" compared to the result from the OSC camera. Without a good bit of PixelMath blending, non-overlapping interference filters cannot produce smooth tonal ranges. This just doesn't matter if you are going to be creating synthetic pallet images from the interference filtered images, however.

That very well may be true. I expect you'd be right there.

 

But we're not photographing rainbows are we?




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





Also tagged with one or more of these keywords: astrophotography, imaging



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics