ASI294mm Pro Bin 1x1 or Bin 2x2 - Forums

Hello fellow astronomy enthusiasts!

I've been using my ASI294mm Pro in combination with my William Optics GT71 for a couple of months now.
So far I've been using the bin 2x2 mode, which results in pretty substantial undersampling. Now it's not a huge issue, as it will only
become visible when zooming in on the stars, but it still bothers me.

Now I'm aware the ASI294mm Pro has a Bin1 mode, which will result in perfect resolution for me, however there's a couple of drawbacks…
Lower dynamic range & less full well capacity. Now my question is quite simple, but I'm afraid the anwser is not:

What bin mode will give me the best results with my 336mm FL?

From a pure sampling perspective, you're correct. The native 2.315µm pixels are a far better match for your reduced GT71. Now, the "best results" is completely subjective

. I've used both the "default" and the "unlocked" modes of my 294MM with my GT81. I've also 2x drizzled my "default mode" data.

Here's an example of an image using the "unlocked" (bin 1) mode:

North America and Pelican in SHO

Now, here's an example of one using the "default" (bin 2) mode:

The Lion in SHO

Hi Lars!

I think you should use 294 mm in 1x1 binning on 336 mm FL. In good seeing conditions, the sampling will be just right. However, you can always donwsample the images to 2x2 binning in post processing, if necessary.

D

I agree with Jonny Bravo.

I personally go with Bin1 for Ha and Bin2 for OIII and SII with my 294MM PRO and Redcat 51.
When shooting LRGB i shoot Bin1 for everything to get the tightest stars (however if you're better at using PixInsight than me you can probably also go for Bin2 for colors here as well).

Another drawback with Bin1 are the somewhat large image-files which are almost 100mb each.

I'm still experimenting with the best combination for me, and I can recommend that you do the same

Jonny Bravo:
From a pure sampling perspective, you're correct. The native 2.315µm pixels are a far better match for your reduced GT71. Now, the "best results" is completely subjective . I've used both the "default" and the "unlocked" modes of my 294MM with my GT81. I've also 2x drizzled my "default mode" data.

Here's an example of an image using the "unlocked" (bin 1) mode:

North America and Pelican in SHO

Now, here's an example of one using the "default" (bin 2) mode:

The Lion in SHO

*** Thank you so much for your reply! Both results look great, personally I do feel like the bin1 looks a bit sharper. Do you agree? What do you typically go with now, having tried both? Do you experience the lower dynamic range or full well capacity of bin1 as a major drawback? ***

David Nozadze:
Hi Lars!

I think you should use 294 mm in 1x1 binning on 336 mm FL. In good seeing conditions, the sampling will be just right. However, you can always donwsample the images to 2x2 binning in post processing, if necessary.

D

Hi David! Thanks for taking the time to reply. What's your opinion on the lower full well capacity and lower dynamic range when using bin1? Will it have a noticable impact on the quality of my images or is the sampling more important in this case?

I agree with Jonny Bravo.

I personally go with Bin1 for Ha and Bin2 for OIII and SII with my 294MM PRO and Redcat 51.
When shooting LRGB i shoot Bin1 for everything to get the tightest stars (however if you're better at using PixInsight than me you can probably also go for Bin2 for colors here as well).

Another drawback with Bin1 are the somewhat large image-files which are almost 100mb each.

I'm still experimenting with the best combination for me, and I can recommend that you do the same

Thank you so much for taking the time to reply. What is the reason for shooting bin2 for OIII and SII? Is this to save some space on your computer or is there any real purpose for it?

I wish I could experiment a lot but clear skies are a rare occasion in the Netherlands and every second of clear sky I get I would like to gather data

Lars Boneschansker:
David Nozadze:
Hi Lars!

I think you should use 294 mm in 1x1 binning on 336 mm FL. In good seeing conditions, the sampling will be just right. However, you can always donwsample the images to 2x2 binning in post processing, if necessary.

D

Hi David! Thanks for taking the time to reply. What's your opinion on the lower full well capacity and lower dynamic range when using bin1? Will it have a noticable impact on the quality of my images or is the sampling more important in this case?

Lars

I have multiple set ups that I use the 294MM from 135 mm rokinon, 280 mm askar 400, 400 mm RASA8. I will use the bin 2x2 for RGB and O3 and S2. In orrder to increase resolution and getting ideal sampling I use the 1x1 mode for Lum and Ha. While lower dynamic range is a slight hit, it is dividing the 1 square into 4 equal squares bring smaller detail and variation. Also, this same reason allows the full well depth to be a mostly equivalent point as the full 1x1 for each pixel at 1x1 would need to be multiplied by 4 to get to the 2x2 well depth.

In my experience, the 1x1 can really bring out a lot of detail as seen in my m33 comparison

RGB + Ha bin 2x2

Lum 1x1 + RGB + Ha bin 2x2

Here is an example of NGC 604 within M33

Hope that helps!
-Brandon

@Brandon Tackett Thank you so much for your so detailed reply! That truly is a stunning difference! What is the reason for not using 1x1 for every channel, but mixing it with 2x2 for certain channels?

Lars Boneschansker:
What's your opinion on the lower full well capacity and lower dynamic range when using bin1? Will it have a noticable impact on the quality of my images or is the sampling more important in this case?

The difference in full well capacity is almost exclusively due to the fact that the pixels are smaller, so can hold less charge, versus when you bin them, you are adding the charge in 4 pixels which gives you higher FWC. So there is very little actual degradation of FWC.

In LRGB imaging, I have used the 294 in Bin 1 for Lum and Bin 2 for RGB which is sort of best of both worlds. The luminance is sampled higher to gather detail, color sampled lower for space. Especially with something like BlurX, high signal areas that are oversampled will show excellent recovery of detail in deconvolution.

Using combinations of bin1 and bin2 takes advantage of the way our eyes work. We see changes in detail from brightness far better than we see changes in detail from chrominance. It's what makes LRGB imaging work so well. Basically, we capture the detail with luminance, and then paint it with the RGB. Thus, the idea is to get as much detail as you can in the luminance channel. You don't need as much detail in the color, so you bin it to take advantage of the increased SNR.

For narrowband imaging, the Ha signal is almost always the strongest. Thus, people will do the same trickery… take the Ha data in bin1 (and potentially use it as a luminance channel as well)… then gather the S2 and O3 in bin2 to take advantage of the SNR. I haven't done it. If I'm going to use bin1 for narrowband imaging, I do it for all the filters. I adjust the data collected instead. So, I might run a sequence like 1xHa, 4xS2, 4xO3.

The biggest downside to the bin1 data is the size. This is easily mitigated with more disk, more RAM and faster/more CPU cores. Unfortunately, in my case, I have a non-upgradable processing box (a 2019 MacBook Pro). It takes a long time to chew through those 47MP subs

In my opinion it depends on how much time you can dedicate to an image and where you are imaging from. Bin1 is cool and all, but I often find myself imaging from either a highly light polluted area or with limited time (think 5 or 6 hour integrations) from a dark site. With that in mind I just choose Bin2 to soak up the light and go for maximum SNR. I suppose you could argue that shooting at Bin1 and down sampling gets you the same or close results, but I'd rather have the full dynamic range and low shot noise without risking super long exposures at a low gain with Bin1. A light drizzle can then get me to under 2 arcsec/pixel anyways.

I think the only telescope I would be willing to go all out on with Bin1 would be any telescope that is F/3 or below just to compensate for the loss of SNR per sub.

I suppose you could argue that shooting at Bin1 and down sampling gets you the same or close results, but I'd rather have the full dynamic range and low shot noise without risking super long exposures at a low gain with Bin1

Shot noise, whether from the signal itself, or from sky background, would depend exclusively on the number of photons captured. Whether you capture in Bin 1 and then do a 2x2 bin, or capture directly in 2x2 bin, the impact of shot noise will be identical. The SNR in Bin 1 will be lower simply because the image scale is different. All you are doing in binning is determining when to change the image scale.

So far I'm seeing many reasons to start shooting in Bin 1, kinda makes me regret all the time I put into my projects shooting in Bin 2. Next clear sky we'll be running bin 1, I'll come back to this thread

One thing to note when using bin1 with a ZWO 294MM Pro. It is always in HCG mode, so you don't get a big drop in noise at unity gain. Also, unity gain in bin1 is 108 (not 120 like in bin2). If you don't have it, I suggest getting SharpCap and running a sensor analysis. Then you can use that data to calculate an "optimal" exposure length (either in SharpCap itself or NINA). That should give you a good starting point for sub lengths for different gains. I use gain 0 and gain 108 for broadband and narrowband, respectively. Of course, your. mileage may vary given different optics and sky conditions

Lars, I have a similar setup for my wide field work (Askar FRA 400, ie 72mm and F=400mm), and usually use bin 1. It is worth noting that in bin 1 the 294mm always uses the HCG mode (the one that is used in bin 2 for gains above 120), so you're note really losing something when taking the subs in bin 1 and software bin later, should SNR be the issue, but if you image in bin 2 you cannot revert to the smaller pixel scale later on. I see three areas where bin2 would be the option of choice: i) for long focal length work (like with an SCT) and typical (ie non Paranal or Manua Kea like) seeing conditions, you're severely oversampling, ii) if the file size is an issue and iii) if you are in a particular dynamic range situation, which brings you to gain well below 120 in bin 2

Matthias

Jonny Bravo:
One thing to note when using bin1 with a ZWO 294MM Pro. It is always in HCG mode, so you don't get a big drop in noise at unity gain. Also, unity gain in bin1 is 108 (not 120 like in bin2). If you don't have it, I suggest getting SharpCap and running a sensor analysis. Then you can use that data to calculate an "optimal" exposure length (either in SharpCap itself or NINA). That should give you a good starting point for sub lengths for different gains. I use gain 0 and gain 108 for broadband and narrowband, respectively. Of course, your. mileage may vary given different optics and sky conditions .

I've been shooting Bin 1 exclusively at gain 108 since this is "unity", however, looking at the read noise/gain chart recently, it dawned on me that there's very little read noise penalty to shooting at 0 gain in Bin 1 - unlike the steep read noise penalty for shooting under 120 in Bin 2.

One of my biggest challenges with this camera in either mode is preventing blown out star cores, so I'm definitely gonna experiment with Gain 0 for LRGB.

Lars, I have a similar setup for my wide field work (Askar FRA 400, ie 72mm and F=400mm), and usually use bin 1. It is worth noting that in bin 1 the 294mm always uses the HCG mode (the one that is used in bin 2 for gains above 120), so you're note really losing something when taking the subs in bin 1 and software bin later, should SNR be the issue, but if you image in bin 2 you cannot revert to the smaller pixel scale later on. I see three areas where bin2 would be the option of choice: i) for long focal length work (like with an SCT) and typical (ie non Paranal or Manua Kea like) seeing conditions, you're severely oversampling, ii) if the file size is an issue and iii) if you are in a particular dynamic range situation, which brings you to gain well below 120 in bin 2

Matthias

Thank you so much for the reply! I've tried bin1 for the first time yesterday, unfortunately clouds rolled in quickly so I won't be able to share anything as of right now.

data looked promising though!

Jonny Bravo:
Using combinations of bin1 and bin2 takes advantage of the way our eyes work. We see changes in detail from brightness far better than we see changes in detail from chrominance. It's what makes LRGB imaging work so well. Basically, we capture the detail with luminance, and then paint it with the RGB. Thus, the idea is to get as much detail as you can in the luminance channel. You don't need as much detail in the color, so you bin it to take advantage of the increased SNR.

For narrowband imaging, the Ha signal is almost always the strongest. Thus, people will do the same trickery... take the Ha data in bin1 (and potentially use it as a luminance channel as well)... then gather the S2 and O3 in bin2 to take advantage of the SNR. I haven't done it. If I'm going to use bin1 for narrowband imaging, I do it for all the filters. I adjust the data collected instead. So, I might run a sequence like 1xHa, 4xS2, 4xO3.

The biggest downside to the bin1 data is the size. This is easily mitigated with more disk, more RAM and faster/more CPU cores. Unfortunately, in my case, I have a non-upgradable processing box (a 2019 MacBook Pro). It takes a long time to chew through those 47MP subs .

Thanks for your suggestion but I still have one question. If you shot Lum with bin1 and RGB with bin2，than how to do LRGB combination since they have different resolution？

@wspba i complete my liner edits and denoise on the 2x2 color data and then star align it with the 1x1 luminance data as the reference. Pixinsight upscales it to the resolution of the bin 1 data.

Brandon Tackett:
@wspba i complete my liner edits and denoise on the 2x2 color data and then star align it with the 1x1 luminance data as the reference. Pixinsight upscales it to the resolution of the bin 1 data.

I got it，thank you very much.

Brandon Tackett:
@wspba i complete my liner edits and denoise on the 2x2 color data and then star align it with the 1x1 luminance data as the reference. Pixinsight upscales it to the resolution of the bin 1 data.

As I was reading this I was wondering how to combine the bin1/bin2 data in PI. Thanks for answering!

I'm wondering if you can get away with using bias instead of darkflats in Bin1, or perhaps just a single, non-time matched Bin1 master darkflat in place of bias.

I am working on reshooting my darks/darkflats right now, and can't see a significant difference between 1.25s and 3.00s darkflats in Bin1 at gain 108. If I subtract the 1.25s exposure from the 3s, there are nonzero pixels, demonstrating increased noise in the longer darkflats, but no obvious amp glow pattern.

I know this cannot be done in Bin2. where subtracting the 1.25s master darkflat from the 3s reveals an ampglow pattern already beginning to form.

I'll also add a caveat to my statement above for using Bin1 Gain 0 for LRGB: this cannot be done for very short exposures without running the risk of failing to swamp the read noise. I did this experiment when trying to shoot M42 a few years ago. Using an f/5.5 system, 15s exposures at Gain 0 had patterned banding in the shadows that could not be calibrated out. This pattern was reduced, but not eliminated altogether with 60s exposures, by which point the core of M42 was already saturated.

So I should instead have said that I plan to revisit using Gain 0 for LRGB on fainter targets with longer exposures.

My experiences with QHY version should be comparable with those of ZWO. The default 2x2 bin requires less tracking accuracy of the mount and less computers resources. I paired my QHY294m with Redcat51 ,CEM25p to a 10 years old laptop. It was more difficult to have good results from the unlock 47MP mode comparing with the 11MP mode.

CS
Kijja