Deep Sky Imaging from Bortle 8 - 9?

LAX must be tough, I feel your pain I am in a Bortle 7/8 myself. Getting halfway decent images from these zones is not impossible and can be done. Usually to get my images to look anywhere near "acceptable" I aim for 10s of hours of integration, depending on the target. I use Optolong L-enhance for emission nebula which I have found really helps bring out the fainter red parts that are near invisible in broadband light using a standard UV/IR cut filter. 

Really look forward to trying an L-pro filter someday as according to https://astrophotographylens.com/blogs/astro/optolong-l-pro-vs-l-enhance-vs-l-extreme it can be used for broadband targets while maintaining natural RGB colored stars, something that is lost with the L-enhance. 

One trick you can try which I sometimes utilize, if you process your images in photoshop, is to reduce the saturation of the yellow or orange channel in camera raw or using the hue and saturation tool, this can help cut out those tones produced by light pollution. 

I have also recently started using a mask and selecting only the shadows in my images, and then desaturating the parts of my image I know are the black sky, this essentially is a form of gradient removal, however, I still will use this step even if I have used conventional gradient removal tools previously in my processing workflow. 

I always use as many dark calibration frames as I have light frames. 


Fabian

I am in a Bortle 8-9 zone just outside of Washington, DC and near its two airports.

I have a modest setup (a 90mm refractor and an ASI183M Pro camera on an AM5 mount). I'm always looking for the silver bullet filter with a perfect result (including the L-Pro, L-eNhance, L-eXtreme, UHS NBZ, Antlia Triband, and Antlia Quadband), but I haven't found that filter yet. 

I have also tried my luck with simple broadband imaging with just a UVIR cut filter, and I've had good success under the circumstances.

I've found that the best approach is to 1) be as selective as you can about the nights you image, 2) be as selective as you can about the data you choose to stack, 3) try to capture targets when they are highest in the sky, and 4) increase integration time for longer than you otherwise might think you should.

Here are some images with what I consider to be good broadband results with modest gear under Bortle 8 skies:

The Hidden Galaxy Is Aptly Named! (IC 342)

The Embryo Nebula (NGC 1333)

A Last-Minute Detour (M78)

The Slow Evolution Of M74

None of these will win Astrobin awards, but I am very pleased with the quality I've been able to achieve in the face of light pollution.

Hope this helps!

CS, Matt

Ralph Ford:
I have seen some remarkable images from locations in or near urban areas.  I live about 3 miles south of LAX, limiting magnitude about 3 on a good night.  I've tried some different filters, but do not like the results.  I use a QHY268C and a Stellarvue SVX125D.

If anyone is willing to share their "secrets" I am all ears.  Thanks!

Having used a variety of LP filters for years, and having processed countless sets of images from Bortle8-9 skies that used LP filters, I am not a real fan. Here is the dirty little secret about LP filters. They eliminate half the spectrum, or more. The visible spectrum, that is, which is about 330nm wide. So you are going to get about 155nm of the spectrum with an LP filter. They usually tend to nuke the greens, which has a significant impact on the quality of stars and diversity of star color. They aren't quite good enough at passing nebula, unless you get a dual-band NB filter, so nebula are ok but not great. You will usually not pick up much reflection nebula, in large part because so many reflections rely on green and yellow light, which is usually eliminated with LP filters. The list goes on. 

With a mono camera, you can do LRGB imaging. My strong opinion is that, under light pollution, using a normal L filter is actually a better way to overcome the additional noise you get with LP, than an LP filter. Why? Basic rule of signals...signal grows faster than noise. An L filter passes twice the signal as an LP filter. It is simply a matter of time, before the L filter will integrate a signal much stronger than an LP filter. You then remove the gradient, and while you will still have additional noise from light pollution, your SNR will be higher with the L filter than with an LP filter. You can then get a decent amount of RGB data, and make do.

When it comes to light polluted zones, if you explicitly just want to make broadband images, then this is my recommendation: Mono+LRGB.

If you want good results under light polluted skies, IMO, you really need narrow band. After I figured out that I disliked LP filters, many many years ago, I switched to NB (unless I was imaging at a dark site) and that gave me so many wonderful targets to image and so many ways to process the results. With NB data, it is easy to get high quality data, with minimal impact from the light pollution, and you can create great results. With an OSC camera, you can get dual or multi-band narrow band filters.

You can in fact do some pretty nice NB imaging with an OSC camera. However, I recommend a different process than you might normally have heard of. Since you will largely be using 25% of the sensor for Ha, and maybe 25% or a bit more (depends on how much green light your sensor actually picks up, OIII has two bands close to each other, one of them will often be picked up a bit more by green pixels, although at a low sensitivity rate) with OIII. Since you are going to be spatial resolution limited here IMPLICITLY, there is little need to keep the resolution 100% or do debayering or any of that. What you should do instead, is demosaic the image using Super Resolution, which will simply put the red pixels in the red channel, green pixels in the green channel, and blue pixels in the blue channel, at half the sensor native resolution. You will then have 100% fill factor blue, green and red channels, or OIII, OIII (2) and Ha channels, if you will.

You can then process as if you had independently captured your narrow band channels. Do SHO blends, HOO blends, try FORAXX, etc. Only real drawback here is, you just won't acquire the same amount of signal as a mono camera will.

To that end, if you REALLY want to make the most of your polluted skies...I honestly do not believe you could do better, than to sell the OSC and buy a mono camera. Then buy at least an Ha and OIII filter, but I'd also recommend SII (and if you like planetary nebula, an NII filter could be useful as well.) If you are unable to use a dark sky location, and are stuck imaging with very high LP (Bortle 9 is brutal!!) then there is really no better way to get good quality signal than mono+NB. I live right at the border of Bortle 8 and 9. My skies rarely get better than 18mag/sq", sometimes they get much worse. This is what mono+NB has allowed me to do:



NGC7822

This is a lot of signal, but, it just shows how clean your signals can be, with very narrow band filters and mono, no matter how polluted your skies are. Some other examples:



Pickering's Triangle



Rosette

Clean, high contrast, higher SNR data is easier to process, less hassle (especially these days with things like GraXPert or NXT), etc. OSC is fine, if you have access to dark skies. OSC becomes more and more of a challenge the deeper into polluted skies you go, and at Bortle 9... Even with a dual-band NB filter, the OSC is going to handicap you a bit. You might start there, see if you can get the hang of NB processing as I mentioned above using Super Resolution to demosaic, which will give you much cleaner channels to work with than standard debayering. If you like that, and like NB image processing, I strongly believe moving to a mono camera with real (and narrow, like 3nm, maybe 5nm) narrow band filters will allow you to make the absolute most of your really terrible skies. ;)

hi

I am in a Bortle 9  or more (  SQM 16.85 ) Miami city .for me only narrowband filter 3nm and the OIII is like mission impossible 

Dan Kearl:
I live in near Downtown Portland Oregon on a 5000 sq. ft. lot.  Bortle 8.
I took this last week with 95% Moon.  200 minutes (20-10 min, subs.) just over 3 hours.
Radian Triad filter, 2600mc camera, Z81 scope.
I am sure LA is worse but there is no reason not to shoot with OSC cameras in Bortle 7-9.
Its not IOTD material but if its where you live, it works.

Just checking...it looks like you used a dual-band narrow band filter here? If so, I totally agree, OSC is a fine option with such a filter. Where I think OSC becomes a huge and often frustrating challenge, is when either LP filters (not narrow band), no lp filters and just bare sensor or maybe a cutoff filter is used. OSC can be used, but, its really tough to create an image as nice as what you have shared here. 

Dual-band (or multi-band) narrow band filters, though, make OSC a fine option for imaging under light polluted skies.

I live in Mexico City, Bortle 8-9 with heavy street lights surounding my setup. I use a RASA8 with either an ASI294MC or a 2600MC. Always use light pollution filters and narrowband. The IDAS NBZ UHS has worked out great, but still have gradients in my lights which have to be removed with the flats. Also use gradient elimination tools, such as the ones in PIxinsight or Graxpert. Also use a dew shield to prevent lights from creating artifacts.

Chuck Schreiner:
Is this something you can do in pixinsight? I've been wandering around the internet trying to see how this is done and come up empty-handed.

I live in the same area as Ralph, and he & I often travel to a dark location. But that really limits things. I use a ASI2600MC and am *always* tempted by mono but so far have resisted. I have used l-extreme locally but ended up with a lot of halos which are a pain to deal with.

In PI this is achieved via SplitCFA. You loose 50% of your potential resolution and should really only be done after ImageCalibration and CosmeticCorrection have been performed on the original CFA files. I disagree that this is the best way forward unless you are significantly oversampled (by a factor of 1.5x at least). The best way forward is really CFA drizzle if you want to preserve you original image resolution.

Chuck Schreiner:

Jon Rista:
You can in fact do some pretty nice NB imaging with an OSC camera. However, I recommend a different process than you might normally have heard of. Since you will largely be using 25% of the sensor for Ha, and maybe 25% or a bit more (depends on how much green light your sensor actually picks up, OIII has two bands close to each other, one of them will often be picked up a bit more by green pixels, although at a low sensitivity rate) with OIII. Since you are going to be spatial resolution limited here IMPLICITLY, there is little need to keep the resolution 100% or do debayering or any of that. What you should do instead, is demosaic the image using Super Resolution, which will simply put the red pixels in the red channel, green pixels in the green channel, and blue pixels in the blue channel, at half the sensor native resolution. You will then have 100% fill factor blue, green and red channels, or OIII, OIII (2) and Ha channels, if you will.

Is this something you can do in pixinsight? I've been wandering around the internet trying to see how this is done and come up empty-handed.

I live in the same area as Ralph, and he & I often travel to a dark location. But that really limits things. I use a ASI2600MC and am *always* tempted by mono but so far have resisted. I have used l-extreme locally but ended up with a lot of halos which are a pain to deal with.

Absolutely. I have used PI for about 10 years. I've always done this in PI. There should be a demosaicing option for super pixel:

https://pixinsight.com/doc/tools/Debayer/Debayer.html

Since there is so much empty data with R and B channels (3/4 pixels are going to be devoid of said colors), you really don't have the spatial resolution there anyway. An interpolation algorithm can spread color information around, but I have always thought that produces an unnatural noise profile and smears signal in ways that make normally debayered OSC images (say with VNG) harder to process. SuperPixel is a great way to get really good quality data from OSC (along with bayer drizzling, but you need a lot of subs for drizzling to become optimal.)

Fabian Butkovich:
Really look forward to trying an L-pro filter someday as according to https://astrophotographylens.com/blogs/astro/optolong-l-pro-vs-l-enhance-vs-l-extreme it can be used for broadband targets while maintaining natural RGB colored stars, something that is lost with the L-enhance.

I've been using the L-Pro for a couple years now, first as a clip-in filter for my Canon RP, and now with my 585MC. Can definitely confirm that it maintains the natural RGB stars, and is pretty good filtering out LP from my Bortle 8 backyard facing downtown Montreal. I pretty much use it as a stock filter for any broadband imaging I'm doing. 

I have less experience using the L-Enhance, which I only just added to the arsenal. That was, admittedly, mostly due to it completely filtering out stars when using my RP (and thus not being able to achieve focus). Now that I'm shooting with the 585, I intend to utilize it more.

Rajat Kumar:
One thing which I have noticed is that OIII signal often takes much more time to start showing up in your image in a meaningful way. Often, the Ha would overwhelm the OIII in OSC and with shorter integrations it would be difficult to process or even differentiate between the two.

Rajat, that's really good info to know. The one time I've shot with my L-Enhance (on M42), I saw next to no OIII after an hour of imaging, but a whole lot of Ha. How much integration are we talking to start seeing the OIII come out more noticeably?

I've always been curious why its called the L-Enhance filter. Its really not a luminance filter of any kind…its a dual-bandpass narrow band filter. A name like NB-Enhance might have been more informative…and indicative of the potential consequences when used on an object like M42.

andrea tasselli:
Not on M42. This is best done with no NB filter. The ratio varies with target and M42 isn't the most suitable one.

Thanks, Andrea. Good to know there's not really a point with M42. Just means more BB time on target!