Ngc 1566

[rogerg (Roger)]

G'day all,

2 clear nights in one week! This is unheard of!

NGC 1566 is very favourably located for me right now, I was able to leave the scope imaging it and it got a good long shot, 9pm until 2:45am before it hit a tree! Very happy. That meant lots of exposures.

L=10x10min
R=8x10min
G=8x10min
B=4x10min

Relatively little colour in this guy as it turns out, but it was requested for my club's newsletter so did him anyway and I am happy with the result. My second LRGB object so far.

As per the last guy (NGC 1232), same res but lower compression factor at my website..
http://www.rogergroom.com/rogergroom...m.jsp?Item=527

Roger.

[Dennis]

Wow - and in COLOUR too!

Gee, Roger, you didn't hang around before putting that CFW8 to use - great image of a neat looking, lovely galaxy.

Cheers

Dennis

[Ric]

Great image Roger, nice details as well.

cheers

[Garyh]

Nicely done Roger, great detail, must have a very low surface brightness....
Cheers Gary

[strongmanmike (Michael)]

That's very good guiding for 10min subs at 2100mmFL and an LX200 even with a self guide camera - any secret?

Great looking galaxy that one really and the colour looks quite ok.

Bet ya wished ya had a bigger chip?

Mike

[rogerg (Roger)]

Thanks all

Yeap - wish I had a bigger chip, but not at the sacrifice of sensitivity

I'm keen to re-process this guy tonight. I posted it from home then when I got to work I looked at it and thought "hmm, there's less exposure detail in that than my 2 minute shots from last year, something is wrong". I put it down to my newbie status with regard to RGB imaging. It could also be because the atmosphere was very turbulent last night so the clarity/resolution isn't so good and so the image would be fainter as a result too.

Tracking - thanks, I am quite happy with it. The secret for my setup is good PEC and fast guiding and good guide star contrast. My guide exposures were 1 second with a brightness of about 2000 ADU. Compare that to the NGC 1232 shot where I had 10 second guide exposures with 400 ADU - big difference in tracking accuracy simply because it would have been making incorrect and unneccessary movements much of the time. The PEC on my LX200 (classic) has got my PE down to something increadably low - I haven't measured it but without autoguiding at worst my stars end up oval shaped, not lines - so thinking about the image scale that must be about 2 arc seconds of PE. I'm often amazed at the class of this now relatively old instrument, that it can track that accurately.

Roger.

[atalas]

Nice shot Roger with nice core detail .

[Lee]

I like it! Very nice Roger..... a nice looking galaxy

[Striker (Tony)]

Nice one Roger.

Some nice exposure time too boot.

Your a bit of a master with this faint stuff...well done.

[rogerg (Roger)]

OK, stop the press. Ignore that trash I posted before, I don't know who's camera that came from!

This is the real deal!

Oh what a difference a bit of different RGB processing does on the same raw data!

This I am much more happy with.

Roger.

[Lee]

I'd be ecstatic with either Roger, but do prefer the second..... excellent work.....

[Striker (Tony)]

So what did you do for enhancement Roger..please tell.

[rogerg (Roger)]

Previously I simply:

1) reduced in CCDSoft
2) aligned in CCDSoft
3) stacked each channel in CCDSoft
4) combined in CCDSoft using the "colour"/"combine colour" (forgotten exact wording) menu option.
5) saved as BMP and edited the levels and curves again in photoshop. Also removed a slight gradient by making a fake flat from the image.
6) sharpenned a little in RegiStax v4

To get a nice image (without the core being burnt out) I needed to keep the levels quite 'spread out' - seeing as CCDSoft doesn't allow curves on FITS files (only 8bit bmp's for some bizzar reason). Combining these then performing levels & curves on the resulting image in photoshop resulted in the first image. I pushed the combined image's levels and curve as far as I was happy to in photoshop.

I then bought a license to FITSPlug by Eddie. Big leap forward.

In my second attempt I:
1) reduced in CCDSoft
2) aligned in CCDSoft
3) stacked each channel in CCDSoft
4) opened each channel in photoshop, adjusting it's curves and levels such that the fiant detail was visible but the core wasn't burnt out, then saving as a .TIF file.
5) combined the new .TIF files in CCDSoft using the "colour"/"combine colour" (forgotten exact wording) menu option.
6) saved as BMP and edited the levels and curves again in photoshop. Also removed a slight gradient by making a fake flat from the image.
7) created a second layer, adjusted it for the faint detial using levels only, used a mask to hide the burnt out centre, merged the 2 layers. (couldn't do this the previous time - simply not enough faint detail in the combined image to make it worthwhile)
8) sharpenned a little in RegiStax v4

Changing the levels and curves of the original FITS before combining them to colour made a huge critical difference - I had much more room to play after combining the chanels. I simply couldn't do this without the FITSPlug.

I think I'm going to re-process every CCD photo I have with the FITSPlug now Even non-colour it means I'll be able to do curves on it, making a huge difference.

Hope that help someone out there.

Roger.

[iceman (Mike)]

Excellent image Roger, beautiful.

[Striker (Tony)]

Intersting Roger,

I was about to post this question.

Do you enhance each color channel before combining or combine then enhance...????

I thought you would combine then enhance but you have done the opposite.

Anyone know the benifits to either method...?????

[rogerg (Roger)]

Tony, the way I see it is this..

The reason I'm having to alter the levels & curves of the images prior to combining them is really a side effect of CCDSoft's limitations. If CCDSoft could combine the images and save the resulting image as a 32 bit FITS file, there would be no point doing it before comining the colours.

However, the resulting image from the combine in CCDSoft is not a FITS file, it's a BMP/TIF which doesn't hold the same bit depth as the FITS files.

Because the resulting combined image is only 16 bit at the most, the ability to get the faint detail in that combined image is limited - it's been chopped off in the bit conversion from 32 to 16.

So to counter-act this, I need to adjust the levels and cruves such that the information I want is within the 16 bits that will be left. Then combine the images resulting in the 16 bit combined image, and not losing any of that faint detail that would have been chopped off.

Perhaps there are other applications (MaximDL? ImagePlus? AstroArt?) which can combine FITS files resulting in a 32bit file. I would expect in this case there to be no difference in doing the levels etc before or only after combining the LRGB.

Perhaps I am right (could be wrong), and perhaps that makes sense

Roger.

[Striker (Tony)]

It sure does Roger..thanks mate.

very much appreciated.

[iceman (Mike)]

Is the camera a 32-bit camera? Is there that much information to begin with?

[rogerg (Roger)]

I'm not sure if the camera is 16 or 32 bit ... there's every chance it's 16bit.

My theory might be right but my numbers wrong, perhaps it should be:
Camera = 16bit
CCDSoft FITS = 16bit
CCDSoft TIFs = 8bit
CCDSoft BMPs = 8bit

.... so data is still list.

I'm quite confident there is a data loss in the combined image in CCDSoft vs the FITS as the difference is quite obvious visually when working with the images, but whether it's 32 vs 16 or 16 vs 8 I'm not so sure..

Roger.

[iceman (Mike)]

TIF's and BMP's are 16-bit, i'm quite sure..

Check the camera specs to see whether it's 16 or 32, I guess..