NGC7635

[GiorgioF]

Undecided if post it or not.

Let's say I'm satisfied about this image but not as much I would. I feel it can be improved. Maybe also on colours but honestly, with my eye colour blindness, not easy.

TS80ED flattened and asi1600mm.

Eq5 (I want a better one...) And postprocessed in Pixinsight.

80 light of 300" in Ha and O3.

How to boost the already captured O3?

How to improve the process?

C&CC really appreciate.

Clear skies!

[vlaiv]

There is no much OIII in this target and you have captured it.

If you have 80 x 5 minutes in both OIII and Ha that would make it 6 hours and 40 minutes in each channel - that is plenty to make good image.

Since you attached 1600 x 1200 image - that means that you don't mind smaller image. You can boost your SNR by binning your data. Bin x2 will give you x2 boost in SNR and about 2200 x 1700 image, while bin x3 will give you x3 boost in SNR and something like 1500 x 1100 image. You can choose which one is better depending on how large you want your image to be.

Next - you can use Ha layer to be luminance layer and use Ha+OIII just for color. Background is a bit too light - it can be darker and there are noise reduction artifact - maybe ease a bit on noise reduction.

[GiorgioF]

On 14/01/2020 at 18:40, vlaiv said:

There is no much OIII in this target and you have captured it.

If you have 80 x 5 minutes in both OIII and Ha that would make it 6 hours and 40 minutes in each channel - that is plenty to make good image.

Since you attached 1600 x 1200 image - that means that you don't mind smaller image. You can boost your SNR by binning your data. Bin x2 will give you x2 boost in SNR and about 2200 x 1700 image, while bin x3 will give you x3 boost in SNR and something like 1500 x 1100 image. You can choose which one is better depending on how large you want your image to be.

Next - you can use Ha layer to be luminance layer and use Ha+OIII just for color. Background is a bit too light - it can be darker and there are noise reduction artifact - maybe ease a bit on noise reduction.

Thanks Vlav for the detailed feedback. I really appreciate.

I agree about binning but with my cmos there is no advantage unlike ccd.  I will try to darken a little the background.

Just a note, I explain maybe in a bad way. 6 hours and 40 is the total H and O together.

Again thanks for your suggestions.

[vlaiv]

10 minutes ago, GiorgioF said:

I agree about binning but with my cmos there is no advantage unlike ccd

That is not true. Any sort of binning produces improvement in SNR - regardless if it is software or hardware type of binning. It is same thing as stacking - you average values and average will be closer to true value than single values.

Difference between hardware and software binning is in amount of read noise you end up with. All other noise sources are treated the same. If we observe 2x2 binning - with hardware binning you will get slightly above x2 SNR improvement (how much over x2 depends on how big contribution read noise has), while with software binning you will get exactly x2 SNR improvement - per sub. When hardware binning 2x2 - signal is increased by x4 but read noise stays the same. With software binning signal is increased by x4 but read noise is increased by factor of x2. That is only difference (and the fact that you can do software binning in numerous ways - which can actually produce better SNR results than hardware binning).

[GiorgioF]

51 minutes ago, vlaiv said:

That is not true. Any sort of binning produces improvement in SNR - regardless if it is software or hardware type of binning. 

Uhm... You make me thinking.

I always got that cmos give no improvement but your explanation has some base to be considered.

As soon as weather will allow me, I will try to gather O3 (that is always weak) with bin2x.

Thank you!

[vlaiv]

Just now, GiorgioF said:

Uhm... You make me thinking.

I always got that cmos give no improvement but your explanation has some base to be considered.

As soon as weather will allow me, I will try to gather O3 (that is always weak) with bin2x.

Thank you!

When binning cmos images - don't do it while imaging - record as you normally record - at full resolution.

Then in processing stage, after you calibrate your subs and before you start stacking - bin them and stack them binned. Since you are using PI - binning is available as integer resample tool.

Here is documentation page explaining it:

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

Just set method to average when down sampling and down sample by factor of x2 or x3. You don't have to wait for additional data - you can do it on data that you already have.

[GiorgioF]

51 minutes ago, vlaiv said:

When binning cmos images - don't do it while imaging - record as you normally record - at full resolution.

Then in processing stage, after you calibrate your subs and before you start stacking - bin them and stack them binned. Since you are using PI - binning is available as integer resample tool.

Here is documentation page explaining it:

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

Just set method to average when down sampling and down sample by factor of x2 or x3. You don't have to wait for additional data - you can do it on data that you already have.

I'll sure give a try on this! Amazing hint! This weekend, work in progress!