What is the target faintness limit for city skies?

Good morning, team.

I’d like to start a discussion about the faintest objects that can be captured under heavily light-polluted skies, such as those in Bortle 8/9 zones.

My primary imaging location is in West London, where the sky brightness ranges from approximately 18.26 (according to LightPollutionMap estimates) to 18.6 (based on my own measurements). To combat light pollution, I use narrowband imaging. Initially, I worked with an OSC camera paired with 6nm Ha/OIII filters but have since upgraded to 3nm Ha/OIII/SII and LRGB filters with a mono camera.

Using the OSC camera and an L-Pro filter, I’ve had moderate success imaging objects like M51, M81 and M82.

With the mono camera and 3nm Ha/OIII filters, I managed to capture Sh2-129 and Ou4 with nearly 24 hours of OIII integration. I was pleasantly surprised to detect some dust and faint structures using LRGB filters near the Seahorse Nebula as well. While the final image incorporates supplemental LRGB data captured under Bortle 4 skies, I’ve also experimented with processing Bortle 8/9 data alone, with some interesting results.

Given these conditions (assuming the more optimistic estimate of 18.6 for sky brightness), do you think it’s feasible to attempt imaging the Integrated Flux Nebula (IFN) near Polaris with LRGB filters and mono camera?

Also, what’s the faintest object you’ve successfully captured from similarly bright city skies?

Looking forward to hearing your thoughts and experiences!

P.S.: I know, I know… I should bite the bullet and travel to darker skies to get better signal with shorter total integration time.

--
Konstantin

My recipe in my B7 skies (most of the times anyways) is to go fast (and I mean real fast, f/4 and below) and go long. Traveling to a dark location is nonsense and I'm having none of that. Rich guys with a lot of disposable income can set up on a remote site but I ain't rich so my urban/sub-urban skies will have to do, presently. Main issue here in the UK is having enough time in any given season given the awful and inconstant weather we have, so my "deep" projects may last years.

andrea tasselli:
My recipe in my B7 skies (most of the times anyways) is to go fast (and I mean real fast, f/4 and below) and go long. Traveling to a dark location is nonsense and I'm having none of that. Rich guys with a lot of disposable income can set up on a remote site but I ain't rich so my urban/sub-urban skies will have to do, presently. Main issue here in the UK is having enough time in any given season given the awful and inconstant weather we have, so my "deep" projects may last years.

I completely agree with you on the limits that British weather sets upon our imaging time span. I do multi-year projects as well. Sometimes, a buck can save the time (like moving from 533 sensor to 2600 and getting much wider FOV), but it general, the target sets well before it's enough of integration (never enough though).

What was the faintest target you approached and succeeded at? What's you sky quality on LightPollutionMap or ClearOutside?

I shoot most of the time from my "backyard site" which is not my backyard at all, but my daughters' elementary school soccer fields, about 3/4 mile away from my house.
Once the parking lot lights go out there, it's fairly dark. Light Pollution Map lists it has being right on the edge of a Bortle 5 and Bortle 4, so we'll call it a 4.5

.
I've been pleasantly surprised by the targets that I am able to image in these conditions.
I would say most very dim objects are within reach at these light levels (I ususally shoot with a f4.8 refractor, and sometimes with a Edge 8 HD). The deciding factor for dim objects at this site tends to be the big, bossy moon, honestly.
The last time I was out without clouds, I attempted to start imaging IC 2118 The Witchhead Nebula. I think it might have worked, but I was dealing with a 65% illuminated moon, and the subs were….less than encouraging.
I do have 2 other sites I go to that get progressively darker. So I don't tend to try and shoot the dimmest objects at my "backyard" Bortle 4.5 site. But it sure would be convenient to do so , so I am going to continue to try.
This is a great topic BTW.
It's always interesting to hear what kind of LP levels other people live in and how they cope with it.
Cheers and CS!
Aaron

Konstantin Dzuin:
I completely agree with you on the limits that British weather sets upon our imaging time span. I do multi-year projects as well. Sometimes, a buck can save the time (like moving from 533 sensor to 2600 and getting much wider FOV), but it general, the target sets well before it's enough of integration (never enough though).

What was the faintest target you approached and succeeded at? What's you sky quality on LightPollutionMap or ClearOutside?

So far the faintest I went for is SH2-91 although SH2-73 is close second. I'm also doing Ou4 with Sh2-129 but might not get enough exposures this time, we shall see. And I'm on the IFN around the M81-M82 complex as being to the north is in the more forgiving B6 sector of my sky.

Last time I measured it, it was 19.0-19.2 with very clear skies but I'd reckon I'm lower than that on average and all my SE to SW is B7 because of the industrial estates to my south.

Until recently I imaged under fair skies of B5/6 right at 19.6. I tried for IFN around M81/M82 and despite a ton of effort it was not very successful. After about 16 hrs of Lum, I could overstretch and see the IFN but it was hard to make out what was IFN, what was a gradient and what was just bad calibration. I also made technique mistakes like using small pixel mode instead of Bin2 on my camera and my processing was less advanced. Even now, under B3/4 skies, dust and IFN are hard. I can get them but the processing is next level and my skills aren't (yet). That said I had a lot of success imaging the Iris nebula and the dust around it. That was with an OSC camera and I just pointed it and let it run for a few nights.

My conclusion is that my processing is likely lacking and that puts a big dampener on what you can get out of the data on top of the challenge of capturing it.

I live in the SF Bay Area and recently measured B7 SQM 18.15 in my backyard. I’ve attempted Ou4 but with an F7 and 7nm system for about 75 hours. I couldn’t cleanly bring out the nebula. However, my system is pretty slow and I’m using OSC. I might reattempt it next year with just longer exposures and see how that goes. I see you already managed to do pretty well with it!

I also have 21 hours on M81/82 with a quadband filter and there’s IFN in the overstretch so I imagine with a faster system, longer integrations, and about 40-80 hours you should be able to pull out the IFN a little.

My latest image was of CTB1. This was fainter than I was expecting and likely the faintest one that I’ve attempted aside from Ou4. This one had to be saved by processing. I think I had to bake the image to save it in the end.

So far, I’ve been mainly using extra integration time and processing to bring out the fainter objects. It’s a nice challenge to start out here and see what I can image.

Joel Lee:
I’ve attempted Ou4 but with an F7 and 7nm system for about 75 hours. I couldn’t cleanly bring out the nebula. However, my system is pretty slow and I’m using OSC. I might reattempt it next year with just longer exposures and see how that goes. I see you already managed to do pretty well with it!

Joel,

Firstly, thank you for appreciating my work on Ou4.

Secondly, I’d like to highlight, that I had switched to 600” exposures for OIII using 3nm filter on ASI2600MM and RedCat 51 f/4.9. The rationale behind that is the improved overall techniques around gradient extraction leave us with real battle against read noise and random shot noise, where much of sky glow is theoretically removed.

From what I saw in the internet, the signal from Ou4 is such a weak one, that getting anything above that remaining noise is challenging in the span of 300” exposures.

Keeping in mind, that read noise is static (quantity wise) in a sub exposure, the shot noise will increase as well as signal from Ou4. The only hope for me, is that useful signal is growing faster than the remainder of shot noise (which should be true). In this case, 600” exposures would give me some kind of extra ADUs for the Ou4 as such.

Unfortunately, I realised that I should do this experiment way too late when the culmination for the target was not in the darkness time, and I won’t get any more useful signal this year.

And as a rule of thumb for me, try acquiring the useful information from dim objects in bright skies, when they are the darkest. Moonless, cloudless (UK weather, you listening?), highest in the sky (zenith ideally).

I’ll follow up this project next year, once the target is back up high, and the astronomical darkness is back as well.

Joe Linington:
Until recently I imaged under fair skies of B5/6 right at 19.6. I tried for IFN around M81/M82 and despite a ton of effort it was not very successful. After about 16 hrs of Lum, I could overstretch and see the IFN but it was hard to make out what was IFN, what was a gradient and what was just bad calibration. I also made technique mistakes like using small pixel mode instead of Bin2 on my camera and my processing was less advanced. Even now, under B3/4 skies, dust and IFN are hard. I can get them but the processing is next level and my skills aren't (yet). That said I had a lot of success imaging the Iris nebula and the dust around it. That was with an OSC camera and I just pointed it and let it run for a few nights.

My conclusion is that my processing is likely lacking and that puts a big dampener on what you can get out of the data on top of the challenge of capturing it.

Joe,

thanks for your reply, this topic is very interesting for me to discuss.

I am evaluating the possibility in general of imaging IFN near Polaris at my backyard location. One of the options is to use fast lens (85mm at f/1.4-2.0 let’s say) with full frame camera and enormous pixels (Sony mirrorless at 5.93um) to get overall idea of possibilities.

At the same time, if it is any kind of success, I’d better do that in mono with better pixel scale of around 1.55-3.11 arcsec/px.

But if that isn’t a good enough idea, I’d better focus on something else in the sky. Number of clear nights is quite limited to say the least.

Konstantin Dzuin:
Joe Linington:
Until recently I imaged under fair skies of B5/6 right at 19.6. I tried for IFN around M81/M82 and despite a ton of effort it was not very successful. After about 16 hrs of Lum, I could overstretch and see the IFN but it was hard to make out what was IFN, what was a gradient and what was just bad calibration. I also made technique mistakes like using small pixel mode instead of Bin2 on my camera and my processing was less advanced. Even now, under B3/4 skies, dust and IFN are hard. I can get them but the processing is next level and my skills aren't (yet). That said I had a lot of success imaging the Iris nebula and the dust around it. That was with an OSC camera and I just pointed it and let it run for a few nights.

My conclusion is that my processing is likely lacking and that puts a big dampener on what you can get out of the data on top of the challenge of capturing it.

Joe,

thanks for your reply, this topic is very interesting for me to discuss.

I am evaluating the possibility in general of imaging IFN near Polaris at my backyard location. One of the options is to use fast lens (85mm at f/1.4-2.0 let’s say) with full frame camera and enormous pixels (Sony mirrorless at 5.93um) to get overall idea of possibilities.

At the same time, if it is any kind of success, I’d better do that in mono with better pixel scale of around 1.55-3.11 arcsec/px.

But if that isn’t a good enough idea, I’d better focus on something else in the sky. Number of clear nights is quite limited to say the least.

I have tried the Polaris IFN with a Sony A7R and a Tamron 90mm f/2.8 Macro but there was a lot of blue bloat and I had calibration issues. Point blank, that was back when I had no clue so my results aren't useful. I may give it a go again in Galaxy season. I did learn how to guide on Polaris using a guidescope offset 45* from the imaging train. So that was useful. I have tried the Sony 85mm f/1.8 lens and it was a nightmare. I couldn't get it to calibrate because it always has some image manipulation turned on in camera that I couldn't disable. I have had much better luck with manual lenses that don't talk to the camera.

The answer isn't longer subs and with modern CMOS sensors read noise is a non-issue unless you are doing extremely short subs, a few seconds or less. Shot noise is the enemy and the only answer beside filters is more light. So, the answer is a larger aperture while maintaining the same focal ratio or faster. At 60 or 70 hours you are really up against the wall of diminishing returns. Are you really going to go to 140 hours or 280? What could you get with narrowband filters in that amount of time? I can understand the challenge but if imaging time is severely restricted by weather, I would want to use that time in a more productive way.

Tony Gondola:
The answer isn't longer subs and with modern CMOS sensors read noise is a non-issue unless you are doing extremely short subs, a few seconds or less. Shot noise is the enemy and the only answer beside filters is more light. So, the answer is a larger aperture while maintaining the same focal ratio or faster. At 60 or 70 hours you are really up against the wall of diminishing returns. Are you really going to go to 140 hours or 280? What could you get with narrowband filters in that amount of time? I can understand the challenge but if imaging time is severely restricted by weather, I would want to use that time in a more productive way.

Tony,

Thank you for your input regarding the noise and its influence on the sub exposure data. Speaking numbers, if he have fixed setup (the same aperture, the same focal ratio, the same sensor), integration of 10 hours worth of 300" sub exposures vs 10 hours worth of 600" sub exposures, do they have any difference?

In a single sub exposure, I can register a signal of approximately 4 ADU (16 bit space) per 300" sub exposure. I wasn't able to assess properly 600" sub exposures due to lower altitude of the target, but my guess is that I should be able to achieve more than 2x increase in useful signal (after gradient extraction), comparing to 300" sub exposures.

Since the read noise for my sensor is around 1.5 e-/RMS, which translates to roughly 6.25 ADU RMS in 16 bit space, I need that longer than 300" exposures with my current setup.

Please correct me, if I understand things incorrectly.

This is how I understand it:

https://www.youtube.com/watch?v=3RH93UvP358

start at 12:45

My standard sub length with a 585 sensor, 6nm dual band, under bortle 8 skies is 15 sec. and I'm getting great results. It's also important to remember that resolution decreases with longer subs, I'm talking in terms of measured FHWM. With short integrations you can with proper culling, do better than your average seeing would suggest.

Joe Linington:
I couldn't get it to calibrate because it always has some image manipulation turned on in camera that I couldn't disable. I have had much better luck with manual lenses that don't talk to the camera.

I feel your pain. I was chasing a tail with this ring artefacts from native Sony lenses. I'm not sure there is any workaround for this. Even though you fully disable any kind of noise reduction, vignette correction, you name it, the device bakes this pattern according to F-stop selected on the lens.

I would like to check with other users over the internet, if it is possible to duct-tape one of the contacts maybe?

You ask if we attempted certain things, but no option to say whether they were a success or not.
I am in Bortle 8 SE London, and attempted the Squid, but hardly any of it showed.

I gave up trying to do broadband targets a long time ago and only image narrowband from home now.

Carastro:
You ask if we attempted certain things, but no option to say whether they were a success or not.
I am in Bortle 8 SE London, and attempted the Squid, but hardly any of it showed.

I gave up trying to do broadband targets a long time ago and only image narrowband from home now.

Carole,

Thank you for your reply and for sharing your experience. I apologize for not including an ‘unsuccessful’ option in the poll—it’s a great suggestion, and I’ll definitely keep it in mind for future discussions.

Hearing about both successes and challenges in this area can be incredibly valuable. It’s not only motivational for those who didn’t achieve their desired results initially but also practical for identifying whether certain targets are simply unfeasible under bright skies, even with advanced gradient removal techniques.

For example, from my Bortle 8/9 location in West London, I’ve started to see the brighter parts of the Squid Nebula (Ou4) with roughly 4 hours of exposure using a 3nm OIII filter. At around 10 hours, I could detect the general outline, and with 20 hours, I was able to stretch the image enough to reveal some of its 3D structures and details. It’s been a challenging but rewarding target for narrowband imaging.

Sh2-129 and Ou4

I appreciate you sharing your approach of focusing exclusively on narrowband from home—it makes so much sense, given the challenges with broadband targets. Looking forward to hearing more thoughts and experiences from others!

Best regards,
Konstantin

You did well to get the Squid from your location.

I was really surprised at how much surrounding dust I could pull out from 36 hours of integration on the Cocoon Nebula, Bortle 7/8. I had never imaging I would be able to pick up such details, however, I am also using an L-Quad Enhance LPF.

IC5146 (Cocoon Nebula) December 2024

Fabian Butkovich:
I was really surprised at how much surrounding dust I could pull out from 36 hours of integration on the Cocoon Nebula, Bortle 7/8. I had never imaging I would be able to pick up such details, however, I am also using an L-Quad Enhance LPF.

IC5146 (Cocoon Nebula) December 2024

Fabian,

Well done! That’s an impressive achievement, especially under Bortle 7/8 skies. Pulling out the surrounding dust from 36 hours of integration on the Cocoon Nebula is no small feat—I imagine the details must be incredible.

Using the L-Quad Enhance LPF seems to have been a great choice for isolating faint structures. I was looking forward to seeing the image but couldn’t locate it earlier, so thank you for posting it here!

Best regards,
Konstantin

@Fabian,

What’s the minimum integration time you aim for to reveal faint details? Your M45 image is absolutely stunning—33 hours is an incredible effort!

Looking forward to hearing more about your approach.

Best regards,
Konstantin

Konstantin Dzuin:
@Fabian,

What’s the minimum integration time you aim for to reveal faint details? Your M45 image is absolutely stunning—33 hours is an incredible effort!

Looking forward to hearing more about your approach.

Best regards,
Konstantin

Thank you! While there do exist scientific methods to determine this number, I just keep acquiring and stacking periodically till I land at a point where I am pleased with the result. I will say galaxies are much more forgiving in my experience when imaging from light polluted skies versus nebula, especially reflection type.

I usually try and target at least 15 hours for Galaxies and at least 25+ hours for Nebula, obviously this varies by Bortle scale and target type.

I do the vast majority of imaging from my backyard now-a-days, which is about bortle 7 (closer to 8 if humidity or dust is in the air). The most important thing I do now is go long… the longer the total integration the better, regardless of what focal ration I image at… be it widefield narrowband stuff with a little scope or galaxies with my C11 or RC10.

I tried to image these from my city skies