How do you BIN images?

[carastro]

I had heard a lot about this and understand the basic principle of it, but cannot find out how you go about BINNING images.

How is it done please?

I use a DSLR.

Thanks

Carolem

[Merlin66]

Binning only really applies to CCD cameras.

There's no auto binning I know of for the DSLR. Some imaging packages (AstroArt) can do 2 x2 pseudo binning after the fact.

In the "normal" CCD chips four pixels can be "joined" together to make one "larger" pixel ie 2 x 2 binning.

HTH

[dph1nm]

Binning can be done in hardware or in software. Harware binning is better, as it reduces the read noise per binned pixel. In *both* cases the binned pixel will have higher s/n than the original unbinned pixels (but at the expense of resolution - actually rather like decreasing the F-ratio of your scope).

For B/W cameras there is not problem, you just add adjacent pixels. However, there is obviously an issue with colour cameras as the adjacent pixels in the RAW image are different colours. There are some colour CCDs which can be truly (i.e hardware) binned, i.e. red pixels only get binned with other red pixels etc - not sure if any DSLRs use them though, and I am sure the Canon 450D doesn't.

In principle you could software bin RAW DSLR output in the same way, but I don't know if there is any software out there which will do it.

Failing that you can just bin up the final colour image - you will get some s/n improvement doing this, but not, I suspect, as much as doing it 'properly'.

NigelM

[studdedsole]

Nebulosity will load raw CR2 files and bin 2x2 & 3x3, never used it so haven't got a clue of the impact on the final stack. Might try it and see what happens.

[ollypenrice]

Yes, with a DSLR or OSC you'd destroy the colour information so far as I can see.

Anyway, Carole, don't do it! The claim often put about that it multiplies sensitivity by four just doesn't work out. You maybe gain about 1.5 and you lose badly on the quality of stellar images.

I never know why but sometimes this defect is worse than others. I recently shot a whole long night of binned RGB and the stars were so bad that I had to de-star the binned colour and add it for noise reduction to a night's unbinned.

At long FLs binning with CCD is routine but that is at seriously long FLs.

Olly

[carastro]

Thanks every-one.

I think I'll give it a miss then. Though I might take a look at Luminosity some time out of curiosity.

Carole

[theastrodragon]

If you have a processing package like Astroart you can experiment by splitting the colour image into its R, G, B components.

You can then bin each of them separately and recombine.

That way you can play around and see if its worth it.

[dph1nm]

The claim often put about that it multiplies sensitivity by four just doesn't work out. You maybe gain about 1.5 and you lose badly on the quality of stellar images.

2x2 binning gains a factor of 2 in s/n (per pixel) not 4 (it is the square root of the number of pixels, so 3x3 gives 3x etc). If read-noise is an issue then this factor of 2 may be larger (potentially up to 4, I guess, if all your noise was read-noise). If your were dominated by some sort of fixed pattern noise, or have pixel-pixel correlations (as colour DSLR images do), I guess you could find it <2, but this would be pretty unusual for a mono CCD.

As for the image quality, this depends what native resolution your pixels start with and the seeing. If you are oversampling then you really should be binning.

NigelM