Difficulty in imaging the IFN

[cfinn]

Seeing some amazing images of the IFN (integrated flux nebula) in the deep sky imaging section of the forum, I thought it would be interesting to run some calculations to see just how difficult this is to do. The IFN is a component of the interstellar medium at high galactic latitudes that is illuminated by the integrated light of our galaxy. It is incredibly faint, with the brightest sections having a surface brightness of just 24.5 magnitudes per square arc second at visual wavelengths.

Let's suppose that we are in a Bortle 5 area, with a sky brightness of 20 magnitudes per square arc second. This is a fairly typical scenario, and the best that many people can realistically hope for. In this scenario, the IFN is more than four orders of magnitude fainter than the sky background! Many now image with CMOS sensors with small pixels, so let's suppose that those pixels are 3.8 microns in size. We'll assume we are using a mono sensor with 80% quantum efficiency on average, and a luminance filter with a bandwidth of 300nm. We'll also assume we are using sub-exposures long enough to swamp the read noise of the sensor, so the dominant source of noise is photon shot noise from the sky.

Under the above scenario, the chart below shows the total integration time in hours required to capture targets of varying surface brightness with a signal to noise ratio of 5 using telescopes of different focal ratios. The dashed horizontal line indicates the surface brightness of the brightest portions of the IFN, with more being revealed at fainter magnitudes. At f7, we require a total integration time > 13 hours just to capture the brightest portions with a SNR of 5! The benefit of faster focal ratios is very evident here, which explains why many of the best images I have seen come from fast systems such as RASAs and Newtonian astrographs. The other way to improve the situation will be with larger pixel sizes (or binning) or seeking out darker skies.

Charles

 

[tomato]

Thanks for this, it helps in explaining how @gorann gets those amazing IFN images. I have a RASA8 and a QHY268c, but alas not the same class of sky, or indeed his dedication to the task.

[gorann]

Very clever and interesting Charles! Yes, I guess it is the combination of my Bortle 2-3 (SQM 21.3 - 21.6) and f/2 RASA that makes it relatively easy for me. In my old images of M81 & 82 with 5-6" f/7 refractors the IFN was barely detectable even here.

12 hours with refractors:

13 hours with the RASA8:

Edited May 17, 2021 by gorann

[ollypenrice]

The RASA seems to be the instrument of choice, here! 

Olly

[tomato]

Yes, the RASA is delivering some superb images from the Hole Observatory, but I think the sky quality helps an awful lot too. I imaged the squid and bat nebula using the same set up as Goran  for the same integration time and only the brighter parts of the squid were visible on my image, nothing like the result that he obtained. 
 

However, I’ll give the M81/M82 region a go at some point when the RASA goes back on the mount.

[cfinn]

17 hours ago, gorann said:

Very clever and interesting Charles! Yes, I guess it is the combination of my Bortle 2-3 (SQM 21.3 - 21.6) and f/2 RASA that makes it relatively easy for me. In my old images of M81 & 82 with 5-6" f/7 refractors the IFN was barely detectable even here.

12 hours with refractors:

13 hours with the RASA8:

Quite a striking difference!

[cfinn]

8 hours ago, ollypenrice said:

The RASA seems to be the instrument of choice, here! 

Olly

The more I think about it, if you want to go this deep for not an extortionate amount of money then a RASA/HyperStar or a fast Newtonian are your best/only options. You could use a small refractor with a focal reducer, but then I don't think you can get to f4 or below without problems, plus you start to under sample and lose detail. You could also trade focal ratio for big aperture and large pixels (e.g. CDK, RC), but then you have a very long focal length and would need a very large corrected field and a large sensor to maintain a decent image scale, so things start getting expensive quickly. Also the bulk would mean you couldn't travel to dark skies if you live (like most of us) in a light polluted area.

Despite being a refractor fan, I would probably conclude that the apex of value, performance and versatility for average seeing and variable weather is probably an 8-10 inch, f4 corrected Newtonian Astrograph, such as this. With modern CMOS sensors this gives you an ideal sampling rate of 1"/pixel, you get good sky coverage with APS-C sized sensors and good SNR on faint targets in reasonable amount of time. You would need four 5" f7 apochromatic refractors imaging simultaneously to match it! The downside of course are those diffraction spikes and Newtonians are a bit cumbersome. They don't cut the clean, slender lines of a nice refractor and they are slightly less easy to use. So I guess if money was no object, I would spend £20k for a Riccardi Honders astrograph such as this!

Edited May 17, 2021 by cfinn

[ollypenrice]

3 minutes ago, cfinn said:

The more I think about it, if you want to go this deep for not an extortionate amount of money then a RASA/HyperStar or a fast Newtonian are your best/only options. You could use a small refractor with a focal reducer, but then I don't think you can get to f4 or below without problems, plus you start to under sample and lose detail. You could also trade focal ratio for big aperture and large pixels (e.g. CDK, RC), but then you have a very long focal length and would need a very large corrected field and a large sensor to maintain a decent image scale, so things start getting expensive quickly. Also the bulk would mean you couldn't travel to dark skies if you live (like most of us) in a light polluted area.

Despite being a refractor fan, I would probably conclude that the apex of value, performance and versatility for average seeing and variable weather is probably an 8-10 inch, f4 corrected Newtonian Astrograph, such as this. With modern CMOS sensors this gives you an ideal sampling rate of 1"/pixel, you get good sky coverage with APS-C sized sensors and good SNR on faint targets in reasonable amount of time. You would need four 5" f7 apochromatic refractors imaging simultaneously to match it! The downside of course are those diffraction spikes and Newtonians are a bit cumbersome. They don't cut the clean, slender lines of a nice refractor and they are slightly less easy to use. So I guess if money was no object, I would spend £20k for a Riccardi Honders astrograph such as this!

I would choose a RASA 8 because there are a couple on here which we know work. Simple as that.

Olly

 

[ultranova]

13 minutes ago, cfinn said:

The more I think about it, if you want to go this deep for not an extortionate amount of money then a RASA/HyperStar or a fast Newtonian are your best/only options. You could use a small refractor with a focal reducer, but then I don't think you can get to f4 or below without problems, plus you start to under sample and lose detail. You could also trade focal ratio for big aperture and large pixels (e.g. CDK, RC), but then you have a very long focal length and would need a very large corrected field and a large sensor to maintain a decent image scale, so things start getting expensive quickly. Also the bulk would mean you couldn't travel to dark skies if you live (like most of us) in a light polluted area.

Despite being a refractor fan, I would probably conclude that the apex of value, performance and versatility for average seeing and variable weather is probably an 8-10 inch, f4 corrected Newtonian Astrograph, such as this. With modern CMOS sensors this gives you an ideal sampling rate of 1"/pixel, you get good sky coverage with APS-C sized sensors and good SNR on faint targets in reasonable amount of time. You would need four 5" f7 apochromatic refractors imaging simultaneously to match it! The downside of course are those diffraction spikes and Newtonians are a bit cumbersome. They don't cut the clean, slender lines of a nice refractor and they are slightly less easy to use. So I guess if money was no object, I would spend £20k for a Riccardi Honders astrograph such as this!

Interesting topic  for sure, my only add to this is focal length,

the two scopes you have suggested are very nice but with two to  nearly 3 times the focal length,

to see the IMF in the better context, you want the Rasa Like Olly said, extremely  fast and a relatively short focal length

off hand I think about 400mm if I remember correctly, .

other wise to do a deep mosaic of the IFN it would probably take a year, , well it would in rainy England where I am lol

Paul

 

[cfinn]

5 minutes ago, ultranova said:

Interesting topic  for sure, my only add to this is focal length,

the two scopes you have suggested are very nice but with two to  nearly 3 times the focal length,

to see the IMF in the better context, you want the Rasa Like Olly said, extremely  fast and a relatively short focal length

off hand I think about 400mm if I remember correctly, .

other wise to do a deep mosaic of the IFN it would probably take a year, , well it would in rainy England where I am lol

Paul

 

Absolutely, I would agree with that. If capturing the IFN is your primary goal then the RASA definitely seems to be the best choice.

[CraigT82]

I've seen some excellent IFN images from the 135mm Samyang camera lens shooting at f/2 ish... cheap way to get going.

 

Edited May 18, 2021 by CraigT82
Samyang not Samsung!

[ollypenrice]

For comparison with Goran's image, this is mine from an F5 Tak FSQ and full frame CCD camera in 20 hours, from a site probably a tad darker than his. (SQM peaks at 22.) It simply isn't as good. As well as being fainter, the IFN shows much less localized structure. (The galaxies were enhanced by long focal length data, hence the Ha in the Cigar. The wider image is LRGB.)  I was pleased with the image at the time because, back then, it was unusual to see colour in the recently-discovered IFN and finding it was our main objective.  However, the greater aperture of the RASA and the high CMOS sensitivity have moved us forward.

Olly

[michael.h.f.wilkinson]

Very interesting discussion. I am getting a Starizona Night Owl 0.4x reducer for my C8, and whilst no RASA, it is a big improvement over what I can currently achieve. Its image circle also fits the ASI183MM I have neatly. I will certainly see whether I can image this region, preferably from the Lauwersoog Dark Sky Park. The RASA is still definitely on the wish list, of course. I might also give this region a wide-field shot with the Sigma 50-100 mm F/1.8 zoom, just to see what happens

[gorann]

Alternatives to the RASA 8 are fast and short FL camera lenses like the Samyang 135 f/2 or similar. Peter Shah @peter shahhas recently posted som very nice IFN images on Astrobin, like this one:

https://www.astrobin.com/bjwawk/

Peter uses a Sigma 105 f/1.4.

I have a Samyang 135 but have not tried it on IFN yet. My personal problem with those very short focal lengths is that they often do not handle small stars well and starfields may melt into a grayish mass rather than discrete stars like with the RASA. I also have a Canon 300 mm f/4 and its does show rather discrete stars, so I started wondering if there is an ideal focal length that both give a very wide field and can resolve stars resonably well. Maybe 200 mm would do that so I started searching for a lens on ebay and found the old Canon FD 200mm f/2.8. Our collegue Metsavainio in Finland bought one of those lenses in 2008, gave it excellent remarks (round stars all over and no chromatic aberration) and took some great images with it:

https://astroanarchy.blogspot.com/2008/02/first-light-canon-fd-200mm-f28-lens.html

The good thing is that they cost less that 200 pounds on ebay (mainly Japanese sellers) and, because loss of astrodarkness has sent me into the yearly AP-buying mode, I have one on the way now (it only took it a week to reach Sweden and I may get it delivered today).

The FD lenses were made for the old non-EOS Canon cameras (those with a roll of film inside) and the adapters sold for FD to EOS do not allow it to reach infinity focus (unless you buy an expensive adapter that contains some lens system), but that should not be a problem if attaching a CMOS astro-camera that has a much shorter sensor distance than a DSLR. I will need two adapters, one from the FD mount to EOS mount (very cheap), and one from EOS to T2.  I have been unable to find any direct FD to T2 adapter.

It only weighs 735 g (som maybe 1.5 kg with camera) and my plan is to have it as a piggy back on a RASA 8 to show what I am missing outside the RASA's FOV. Just have to figure out the best way to mount it.

 

Edited May 18, 2021 by gorann

[vlaiv]

I find it quite interesting that IFN is only magnitude 24.5.

Are you sure of that?

We regularly capture magnitude ~25 in fainter parts of bright galaxies without too much of a problem.

For example, here is M51 surface brightness:

You can see the image and reached magnitude in that part - it goes down to mag23 without much depth in the image itself and no mention of tidal tail

I made similar chart of M51:

It shows that much of tidal tail is mag25 (and similarly - where line ends in above image - we have mag23 band).

In any case, reaching mag24.5.

By the way - above diagram of M51 was made in sqm18.5 skies with 2 hours of exposure on F/8 telescope and ASI1600 (with use of IDAS LPS P2 filter). It did take some clever manipulation - like extensive x4 binning and morphological transforms of the image to get it relatively smooth like that.

 

[cfinn]

37 minutes ago, vlaiv said:

I find it quite interesting that IFN is only magnitude 24.5.

Are you sure of that?

We regularly capture magnitude ~25 in fainter parts of bright galaxies without too much of a problem.

For example, here is M51 surface brightness:

You can see the image and reached magnitude in that part - it goes down to mag23 without much depth in the image itself and no mention of tidal tail

I made similar chart of M51:

It shows that much of tidal tail is mag25 (and similarly - where line ends in above image - we have mag23 band).

In any case, reaching mag24.5.

By the way - above diagram of M51 was made in sqm18.5 skies with 2 hours of exposure on F/8 telescope and ASI1600 (with use of IDAS LPS P2 filter). It did take some clever manipulation - like extensive x4 binning and morphological transforms of the image to get it relatively smooth like that.

 

Here is the discovery paper in 1976 by Sandage et al. In the abstract it states the brightest portions are ~25 mag per square arc second in the V band. Note that these are the brightest portions only. Much of it extends to much fainter surface brightness. I am not completely surprised by this, otherwise there would not be reports of detecting the IFN visually.

It is also a question of signal to noise ratio. My calculations above are for a SNR of 5, which is quite small. I suspect many detect the IFN but then lose the signal in processing what with background calibration, removing light pollution gradients and so on. To get great images like Goran's will require much higher SNR, which is where faster f ratio combined with darker skies really helps you. Binning will of course compensate for slow f ratio, which you mention you have done above. The argument for the RASA is clearly that you can cover a much larger area of sky than you could with an f8 RC with large pixels or aggressive binning.

One other thing to say about my calculations are that they are for a mono sensor with a luminance filter. For an LRGB image you would need to multiply the total integration time by 2 (assuming same integration time per channel) or for RGB/OSC cameras multiply by 3!

Edited May 18, 2021 by cfinn

[cfinn]

4 hours ago, ollypenrice said:

For comparison with Goran's image, this is mine from an F5 Tak FSQ and full frame CCD camera in 20 hours, from a site probably a tad darker than his. (SQM peaks at 22.) It simply isn't as good. As well as being fainter, the IFN shows much less localized structure. (The galaxies were enhanced by long focal length data, hence the Ha in the Cigar. The wider image is LRGB.)  I was pleased with the image at the time because, back then, it was unusual to see colour in the recently-discovered IFN and finding it was our main objective.  However, the greater aperture of the RASA and the high CMOS sensitivity have moved us forward.

Olly

Lovely image!

[vlaiv]

48 minutes ago, cfinn said:

Here is the discovery paper in 1976 by Sandage et al. In the abstract it states the brightest portions are ~25 mag per square arc second in the V band. Note that these are the brightest portions only. Much of it extends to much fainter surface brightness. I am not completely surprised by this, otherwise there would not be reports of detecting the IFN visually.

It is also a question of signal to noise ratio. My calculations above are for a SNR of 5, which is quite small. I suspect many detect the IFN but then lose the signal in processing what with background calibration, removing light pollution gradients and so on. To get great images like Goran's will require much higher SNR, which is where faster f ratio combined with darker skies really helps you. Binning will of course compensate for slow f ratio, which you mention you have done above. The argument for the RASA is clearly that you can cover a much larger area of sky than you could with an f8 RC with large pixels or aggressive binning.

One other thing to say about my calculations are that they are for a mono sensor with a luminance filter. For an LRGB image you would need to multiply the total integration time by 2 (assuming same integration time per channel) or for RGB/OSC cameras multiply by 3!

I don't doubt your calculations at all. I'm quite aware how detrimental sky brightness is on capturing faint detail.

I was just surprised with brightness of IFN being mag24.5. I would have expected more images to capture IFN at that brightness given that many people manage to capture mag24-25 parts of other objects in their images.

 

[cfinn]

6 minutes ago, vlaiv said:

I don't doubt your calculations at all. I'm quite aware how detrimental sky brightness is on capturing faint detail.

I was just surprised with brightness of IFN being mag24.5. I would have expected more images to capture IFN at that brightness given that many people manage to capture mag24-25 parts of other objects in their images.

 

I quite agree. My suspicion is that it gets lost in processing all too often, due to the fact that it is usually very spread out and could be mistaken for background in background calibration.

[gorann]

This image just came out on Astrobin. A RASA8 can catch IFN even in Bortle 9 Los Angeles!

But you need stamina to do it.....

https://www.astrobin.com/oop1fh/

[tomato]

Wow, 49 hrs... Respect.

LA as an imaging site for capturing IFN, who would have thought it?

[cfinn]

1 hour ago, tomato said:

Wow, 49 hrs... Respect.

LA as an imaging site for capturing IFN, who would have thought it?

Just ran another calculation to back this up (as if further proof were needed!). The chart below shows the integration time necessary to reach increasing levels of signal to noise ratio on a target having surface brightness of 25 mag per square arcsecond in Los Angeles (Bortle 8/9). This is with a RASA at f/2 and specs to match the camera in the description (ZWO ASI 294MC Pro). The dashed vertical line shows the integration time of 49.2 hours, which suggests a SNR of roughly 5 should have been reached on IFN of that surface brightness. If you want SNR of 10 you would have to go for 150 hours!

Having said all of this, it's pretty mind blowing any of this is possible because at this point we are talking about imaging something that is 10 million times fainter than the sky background. I think this should bolster owners of the RASA that much is possible with such a fast instrument, even from the middle of the city!

[gorann]

3 hours ago, tomato said:

Wow, 49 hrs... Respect.

LA as an imaging site for capturing IFN, who would have thought it?

Actually, he seems to have spent 100 hours on it and had to throw away half of it - we need guys like him!

EDIT: And he replied to me on Astrobin: Markice has now joined SGL. Hopefully we will se more from him.

Edited May 19, 2021 by gorann

[Guest]

10 hours ago, gorann said:

Actually, he seems to have spent 100 hours on it and had to throw away half of it - we need guys like him!

EDIT: And he replied to me on Astrobin: Markice has now joined SGL. Hopefully we will se more from him.

Thanks for the shout, Göran! Yes, I had to throw away lots of subs because my AVX was not able to handle some of the accessories that I added to the mount. With the CGEMII I am now getting around .35-.45  total RMS consistently and even as low as .2. Unfortunately, I was only able to use this mount for a few nights at the end of my run @ IFN.

[Guest]

11 hours ago, cfinn said:

Just ran another calculation to back this up (as if further proof were needed!). The chart below shows the integration time necessary to reach increasing levels of signal to noise ratio on a target having surface brightness of 25 mag per square arcsecond in Los Angeles (Bortle 8/9). This is with a RASA at f/2 and specs to match the camera in the description (ZWO ASI 294MC Pro). The dashed vertical line shows the integration time of 49.2 hours, which suggests a SNR of roughly 5 should have been reached on IFN of that surface brightness. If you want SNR of 10 you would have to go for 150 hours!

Having said all of this, it's pretty mind blowing any of this is possible because at this point we are talking about imaging something that is 10 million times fainter than the sky background. I think this should bolster owners of the RASA that much is possible with such a fast instrument, even from the middle of the city!

Really interesting! I've now moved onto M63, attempting to pull out the tidal streams from the same location. I don't know if I have enough time left to finish the project before I run into some obstructions, but at 6 hours I already see a bit of structure. Another variable that may be worth looking into for imaging from this location is the UV/IR cut rather than a light pollution filter combined with a mono camera. I began the IFN project using the Optolong L-Pro and at 10 hours, there was a bit of IFN present as well as much cleaner data. However, there was clearly a lot of good signal cut out when compared to the Baader UV/IR cut at 10 hours. The 294 Mono + the UV/IR cut filter made such a huge difference. The one shot color data was actually pretty nasty for M81/M82, even with a LP filter, so I am reluctant to switch to UV/IR with the color cam (unless I use short subs to keep out LP) but I will have a look at it for some smaller projects.

Astrobin makes it pretty easy to miss, but I use a 2 camera setup (294MM and MC) because it made it a bit easier to not have to switch between individual filters for the mono. I'm sure that the 90% QE also helped a lot with the LUM data

Edited May 20, 2021 by LeftyAstro