What is binning ?

[Daz1969]

Hi all , am getting further down the road towards astro imaging and keep coming across the term 'Binning' What does that mean ? Have i got it ? or do i need it?

Cheers in advance as always:hello2:

Daz

[steppenwolf]

Daz,

Binning is the process of combinning (sic) pixels together to make 'super pixels'. If you have a sensor that is 1000 pixels wide and 600 pixels deep and you 'bin' the image 2 x 2 then you end up with the pixels grouping into groups of 4 pixels, each group forming one 'super pixel'. The sensor thus becomes one with 500 pixels wide x 300 pixels deep but the fewer 'super pixels' will be 4 times more sensitive than the original pixels.

[Daz1969]

Ah cheers for that Steve! That sounds really interesting , Is it something you do with your capture software settings or is it something you do when you are processing the image?

Daz:icon_salut:

[Whippy]

It's something you do when you're capturing the data. It comes in really handy if you have a mono camera and you want to do colour. What most people do is take the mono with no binning (let's say 60 minutes' worth), then do the colour binned. As the binned pixels are 4 times more sensitive, then you're able to cut the time needed to do the colour data to a quarter. So for an hours' worth of mono, you can take just 15 minutes' of each colour filter rather than an hour for each. The good part is that when you blend the mono and RGB images together, you get the colours without losing the detail of the mono image.

It also helps to get a rough focus on older USB1 cameras as it takes less time to download each image.

Tony..

[Daz1969]

Cheers Tony, Thats facinating stuff ! This learning curve is a great one. Would i not use binning with a colour ccd camera as iam yet to buy one and dont know if i should go mono or colour

Have just bought a Heq5 pro mount for my 8" Sct and got myself a cheap 60mm refractor for guiding with.

Daz

[steppenwolf]

I have used binning with my SXVF-H9C which is a one shot CCD camera - I use this mode when taking Hydrogen Alpha images but I do this in the full knowledge that because of the Bayer Matrix (RGB filters) overlaid on the sensor, I am not getting much advantage from doing so - I do it to destroy the Bayer Matrix and to gain a small amount of extra sensitivity. Binning really comes into its own with a mono camera. However, I also use it with my colour CCD for focus purposes as I get greater sensitivity and a smaller file to download so I can take a faster series of images while I focus.

[Daz1969]

Mmmm,A dark science indeed, Very interesting though,Many thanks for all your help.

Daz

[piprees]

I realise this thread is pretty old but I'm just getting into greater details after getting a ZWO 1600mm cool.

So my question is, if I do 16x300 luminance exposures should I then apply 2x binning with 8x300 or 16x150 for the RGB filters?

Kind regards,

P.

[ollypenrice]

8 hours ago, piprees said:

I realise this thread is pretty old but I'm just getting into greater details after getting a ZWO 1600mm cool.

So my question is, if I do 16x300 luminance exposures should I then apply 2x binning with 8x300 or 16x150 for the RGB filters?

Kind regards,

P.

It very much depends on your pixel scale. If your binned pixel scale is less than about 4 arcseconds per pixel then it might be worth binning the RGB. (Why 4"PP? No good reason, it's a guess plucked from the air to define the scale beyond which I personally would not bin colour. I don't expect everyone to agree.)

Another factor concerns your processing method. Early on, you'll be happy to apply luminance globally over your RGB data. As you slide into the obsessive processing trap you may want to avoid applying luminance to your RGB stars, in which case you'll want nice unbinned RGB stars! Personally I never bin colour. I want to be able to bring those nice little RGB stars, with unsaturted cores, into my final image.

Olly

[MattJenko]

Binning also gets complicated when dealing with CMOS or CCD chips. CCDs can be binned at a hardware level, and given the noise characteristics, this can improve signal to noise ratios. With CMOS chips, it is different, in that most (maybe all, not sure) are software binned only. However, with the low noise of modern CMOS chips, the concern about this meaning binning with CMOS is pointless is lower, but it will not result in a better SNR as I understand things.

[Ande]

Like many others on here, my little brain is trying to piece together the purpose, and practicalities of all of these weird and wonderful concepts. Is there any merit in using binning with a DSLR?  Also, being a beginner, is it just something I should sidestep for the time being, until I’m more experienced? I’ve recently bought an ASIair Pro, which has yet to see first light, and binning seems to be one of the many functions on offer.

[Phillyo]

1 hour ago, Ande said:

Like many others on here, my little brain is trying to piece together the purpose, and practicalities of all of these weird and wonderful concepts. Is there any merit in using binning with a DSLR?  Also, being a beginner, is it just something I should sidestep for the time being, until I’m more experienced? I’ve recently bought an ASIair Pro, which has yet to see first light, and binning seems to be one of the many functions on offer.

You can't bin a DSLR due to it being a one shot colour camera. That being it'll have a bayer matrix over the sensor that is something like RGGB. Binning is used with mono cameras (someone correct me if I'm wrong).

[Ande]

20 minutes ago, Phillyo said:

You can't bin a DSLR due to it being a one shot colour camera. That being it'll have a bayer matrix over the sensor that is something like RGGB. Binning is used with mono cameras (someone correct me if I'm wrong).

Thanks for clarifying. That is one less idea for me to try to get my head around, which is no bad thing. Consider the notion binned

Edited January 14, 2021 by Ande

[jacko61]

On 14/01/2021 at 15:23, Phillyo said:

You can't bin a DSLR due to it being a one shot colour camera. 

Does that hold true for one shot colour CMOS/CCD too ? e.g. ZWO ASI183MC pro? (or any other zwo 'MC' for that matter).

 

Graeme 

[Merlin66]

All CMOS chips cannot be hardware binned (like the CCD chips)

The best you can do is to bin in software after acquisition.

[JamesF]

5 hours ago, jacko61 said:

Does that hold true for one shot colour CMOS/CCD too ? e.g. ZWO ASI183MC pro? (or any other zwo 'MC' for that matter).

Some one-shot colour CMOS cameras do support "binning" and I'm pretty sure that ZWO's OSC cameras do, though it may be a process that happens in the driver software rather than on the camera itself.  I know there are interpretations of what binning means when you have an OSC image, but I'm far from certain that everyone agrees it means exactly the same thing.

For example, if you requested a binned image from two different manufacturers cameras with the same mono sensor, I think you could be reasonably sure that the same process would take place in each instance.  If you requested a binned image from, say, an Altair camera and a ZWO camera with the same OSC sensor, could you be sure that the same process had taken place for both?  I'm really not sure.

It may be the case that there's a single "most sensible" method for handling binning of OSC frames, but unless they show their working I'd not want to assume that all CMOS OSC camera manufacturers do that.

James

[JamesF]

Just checked with the ASI224MC that I happen to have on my desk at the moment.  It supports binning.

To add to the confusion, the ASI SDK also allows for cameras to use "hardware binning" (which is also supported by my ASI224MC), possibly suggesting that OSC binning can be done in firmware on the camera or in the SDK if it is not supported or turned off in the camera.  I don't know if there's any guarantee that exactly the same process will be followed in each case so that the same data coming off the sensor would result in the same image being returned to the user regardless of the binning method used.

James

[vlaiv]

It is best to bin CMOS data in software as this gives one complete control of how exactly binning is performed.

On 14/01/2021 at 15:19, Ande said:

Like many others on here, my little brain is trying to piece together the purpose, and practicalities of all of these weird and wonderful concepts. Is there any merit in using binning with a DSLR?  Also, being a beginner, is it just something I should sidestep for the time being, until I’m more experienced? I’ve recently bought an ASIair Pro, which has yet to see first light, and binning seems to be one of the many functions on offer.

Of course you can bin DSLR data - once it has been recorded - in software.

Binning is just summing group of adjacent pixels to form a single value.

- It can be done in hardware. With CCD sensors (aka true binning) - actual electrons are shifted into same place - hence the name "hardware" binning, before they are converted into number. Upside to this approach is that binned value is subject to read noise - so resulting value has only one "dose" of read noise.

- It can be done in firmware - with CMOS sensors. This is the same thing as doing it in software - values are added after they have been converted to numbers. Downside is that each value gets "dose" of read noise prior to binning. Sometimes people think that this form of binning is not true binning and that it does not provide same benefits as hardware binning (one in CCDs) - that is not true - it is the same thing except for the read noise.

- You can bin bayer matrix to get mono value

- You can bin pixels after debayering to increase SNR. However - do pay attention that if you use debayering method that relies on interpolation - then binning 2x2 of such data does not make much sense - it does not improve SNR but instead increases pixel to pixel correlation (sort of blur).

- If you want to bin OSC data / data from DSLR - first understand that actual resolution and pixel count of your camera is not what you think it is. Learn to split debayer / or use super pixel mode. Once you recover color information like that - then it makes perfect sense to further bin x2 or x3 to increase SNR. Do bare in mind that this procedure significantly reduces size of your image from what you think you have in the first place (but in reality that is not so).

Say you have sensor that is 6000x4000 pixels (24MP sensor). You really don't have 24MP sensor there. You have 1/4 of red pixels, so only 6MP for red color, you have 1/4 of blue pixels, so another 6MP for blue color, and although remaining 1/2 pixels are green - and that would imply 12MP - in reality you have 2x6MP images. In another words - your OSC sensor is not 24MP or 6000x4000 sensor - it is 6MP or 3000x2000 sensor that can produce RGB at the same time (and in fact green will have twice as many frames the other two colors - which is good since luminance from such data consists of about 80% of green color - so the least noise in lum).

In any case, once you have 3000x2000 real pixels - you can bin those x2 to get 1500x1000 or x3 to get 1000x667px images. That may seem rather small by today's standards - but it really is not. 1500x1000 is larger size than most people use to view image posted online (most have full HD displays with 1920x1080 but never look at images in full screen - image displayed here on SGL is about 1500x1000 in size).

In the end - binning improves SNR of your data at expense of sampling resolution - and in most cases that is sacrifice you can afford - since with shrinking pixel size these days - most people are oversampling anyway.