Ngc 7496

[rogerg]

Galaxy - NGC 7496.

I haven't quite got the hang of LRGB processing so I'm not sure on the correctness of the colour in this image.

Reasonable image quality in general though, so not too bad.

Roger.

[Astroman]

Nice one Roger, looking a little too yellow, but its a good effort. The only way to get it is to keep doing it, so you are well on your way with the processing.

[Dennis]

Hi Roger

That is a very nice image. Nice round stars, good detail and congrats on the LRGB effort. I'm planning to dip my toes into LRGB shortly, so I'll be hot on your heels and ecstatic if I can produce an image such as this.

Cheers

Dennis

[Garyh]

nice one Roger!! a nice image for your start into LRGB
Nice natural colors but like Andrew mentions maybe a very little on the yellow side? well guided but!..:thumbsup:

[jase]

This is a fine effort Roger, certainly at an .86 arcsec/pixel combination. Nice round stars. If anything, I wouldn't skimp on the colour exposures. You've got 9x5min shots for luminance which is great, but your RGB are only a couple of exposures. Try to take at least three exposures per channel to allow you to perform a median or sigma reject/clip combine. If you've only got two, then you can only average or sum combine, both of which will not cancel out noise in an effective manner. The richer your RGB data, the more you can stretch it to bring out the colours. Looking forward to seeing future developments.

[rogerg]

Thanks very much all for the positive comments, much appreciated.

I keep wanting more RGB data but not getting there before the cloud rolls in or I need to go to sleep. I think perhaps I will drop my ratio from 3:1:1:1 to 2:1:1:1 and get more in that way.

I would like more contrast in the colour range, and I'm guessing that would come if I had more colour data?

I discovered when I packed up last night (after submitting this post) that cloud had rolled in and covered most of the sky. My guide star was fluctuating a bit but I didn't think to go outside and look. I wonder how much of the night I was looking through cloud! I think the guide star started fluctuating about half way through the night.

Jase, I'm keen to hear more about median and sigma clipping vs just averaging. I only use average and add. Sometimes I add ever second exposure and then average the result of that. Should I be doing something different (median?) for colour frames vs luminance frames for best results? (assuming I have enough data)

Roger.

[RB]

Very nice capture Roger, it's always hard when the clouds interfere and you just can't get the exposures you want.

[jase]

Try to work on an L to RGB exposure ratio of 3:1 but depends on the object and goals (i.e what you are trying to achieve). Your luminance data is critical as this is where all the resolution comes from. If you’re producing an LRGB, the RGB data doesn’t need to hold exquisite resolution. In such cases, 2x2 bin your RGB. This will allow you to acquire much better S/N in a shorter time.

The more RGB data you have the easier it is to work with. If its good, it can be stretched without introducing considerable noise. Adding noisy RGB data to a smooth luminance image defeats the purpose.

Ok, will keep things brief as this is really an entirely new topic;

Sum (or add as you put it) adds up the pixel values in all images you’re combining.

Average adds up pixel values in all images (just like Sum), but then divides them by the number of images you’re combining

Median takes the medium/middle values of the pixels in all images you’re combining. Pixels that fall outside this range are rejected.

Sigma Clip/Reject/Combine is a combination of Median and Average with other techniques to detect outlier pixels. The pixel values that deviate the most from the mean will be discarded providing the deviation is greater than the number of standard deviations configured by you. The standard deviation value is sometimes known as the Sigma Factor. Based on the Sigma Factor a new mean is calculated without the discarded value and is assigned to the pixel in the combined image

These are the main four algorithms. There are variations on these introduced by tools such as RC Console in MaximDL and CCDStack etc. For example Image Plus has a Sigma Average and Sigma Median. Ultimately they all follow similar principals.

Sum and Average are not effective a reducing noise. Some would disagree. Based on experiments have conducted with noisy data, average has an advantage over sum, but marginally. Median combine works very well to suppress noise; however there is also a loss in signal when using the algorithm. To circumvent this, you need plenty of data. Median works well for combining calibration frames. Finally Sigma X (various permutations) offers the greats flexibility at suppressing noise and enhancing signal. It does take some experimenting to get the Sigma Factor right. If incorrect you’ll lose too much valuable data.

There are also other things to consider when imaging RGB in particular image normalisation. In order to accurately present each colour channel in a balanced manner, the background pixel values of each combined R, G, B channel should be similar to each other. Some tools such as MaximDL perform the normalisation automatically as part of the colour combine process. If one channel for example blue is brighter than the others due to light pollution etc, it will dominate and through the balance out. You can use pixel math to rectify this. There are many other tricks to handling these tasks.

Cheers

[rogerg]

Thanks for the various bits of info Jase, some useful stuff for next time.

[Bassnut]

Jase, given CCD stack seperately allows noise reduction before combining, I have found I use sum mostly to keep signal level high. Do you think sigma or average still offers an advantage even after 1st doing noise reduction ?.

[rogerg]

I have found both sum and median provide more noisy results than average with my ST7 images. If I use sum it is only to then have those results averaged. Hence I typically stay clear of both those. That's just my experience.

Roger.

[rogerg]

Btw, for the fun of it, this is the same raw data but having run a Sum on the Luminance channel. RGB are still averages (couldn't be bothered re-doing them als0 right now).

Much more noisy than the original, but obviously the data that is there in both is more obvious here.

There's a happy medium somewhere.

[Bassnut]

Roger, what software you useing?.

[jase]

Fred, CCDStack is an exception. Stan Moore has produced a remarkable piece of code when it comes to noise suppression. The Sum algorithm does absolutely nothing to control noise. As you combine noisy images, not only does your signal increase, but so does noise at the same rate. This is where CCDStack smarts come in to play by performing significant analysis before the combine routine. I feel certain that using median or sigma with CCDStack would deliver an improved result. I'm not sure if you've tried Sigma by Ray Gralak - http://www.gralak.com/Sigma/. Ray is also the developer of PEMPro. The code works quite well, but I personally like Russ Croman's RC Console for Sigma as it give you a nice percentage of how many pixels have been rejected (similar to CCDStack). While I do have a copy of CCDStack, I have not used it enough to make a quantitative comment on its different combine methodologies.

The common thread along all algorithms is large quantities of data (as you are well aware). These algorithms will not magically turn your average astro image into an APOD candidate. Only when you present the algorithms with such volumes of data will they work well for you.

[rogerg]

Reduction and stacking in CCDSoft then LRGB combine & the rest in Photoshop.

[Bassnut]

well, its brighter, umm, I wont bang on about that thumbnail res again, that would be irritating ;-). Nice image, but the stars seem a bit bloaty. Colour balance is nice though, not bad.

[jase]

Roger, There is certainly more spiral structure present in the new image. Perhaps not as clean. You could have potentially stretched the first image a little more which could have delivered a similar result to the latest image, but with less noise. How far you can stretch the data signifies its quality. Short or lack of exposures will result it more noise when stretched.

You're not doing the entire LRGB process in CCDSoft are you? You'll find you will have much great control combining the lum and RGB data in photoshop. Simply register the two before opening them in photoshop otherwise you'll need to nudge (and/or skew/rotate) the layers.

[Bassnut]

OK Jase, I understand, different algorithims depend on the number of subs processed, some suit minimum data, some rely on more. And I know sum is the worst on noisy data. I only ment that with carefull noise reduction beforehand, I couldnt see any reason for not using sum, given from my understanding, it best for noisless data. I havent tried all the 3rd sigma combines you mention, but I have found with the ones I have tried, the differential results were marginal, and the end result differences were quickly buried in subsequent processing. Ofcourse I also understand the need for processing depends on the quality of the raw data and combining method used. This process comes with experience, the better the earlier processes, the less time and work required in PS ;-)

[Bassnut]

Im agree with you totally there Jase, L and RGB should be merged in PS for a million reasons (jeez, and dont bother aligning in PS ;-)., this is a given.

[rogerg]

I like the original the best, it has nicer colour and is not pushed. Low noise level. Usual story, I know more data is desirable and required to reach the brightness of the latter while retaining quality, but I have very limited time for astronomy these days :(

I do the combine of LRGB in photoshop, yes. I tried twice in CCDSoft and gave up on that useless tool they have for it in there. I find it very easy in photoshop.

It's unfortnate that CCDSoft doesn't have sigma stacking. CCDSoft v5 is falling behind competitors in many regards, I hope SB soon bring out a new version, it has been a while.

Thanks for the interesting discussion on the image processing. :)

[h0ughy]

wow what an informative thread. Nice image Roger, and I like the way there has been constructive advice throuhgout. Not that I have a CCD, if I did I would have gleened a lot from this.