Is it possible to make the Horsehead more apparent with these data?

[Epicycle]

Hi,

The tif image below is a stack of 27 subs of 20 seconds each with 25 dark frames substracted. It was processed

with Nebulosity 3 and nothing else (applying DDP and some noise reduction).

The question is: have I made the most of these data? The Flame nebula appears OK, but the Horsehead is barely visible.

Levels or Curves don't seem to help much at this stage, I thought that some of you which are more experienced might

be able to tell me if it is possible to make the Horsehead stand out with a bit more detail.

Thanks!

E.

http://www.dropbox.c...fu/Flame&HH.tif

[steppenwolf]

The big problem with the HH is that you can only detect it against the bright curtain of nebulosity behind it and that in itself is pretty dim so you need long exposures to capture it. An impediment to extracting more detail from your image is that you haven't taken any flats so there is a goodly amount of vignetting too.

You really need to aim for a minimum of 5 minute exposures of this object I'm afraid.

[lukebl]

There's potential there, but a first glance tells me that you didn't take any flats, which will make processing difficult due to the strong vignetting. Also, the exposures weren't nearly long enough, resulting in a very noisy image. Take flats and longer subs. Even an unmodded camera will pick up the horsehead with longer subs (say, 5 mins).

EDIT: Sorry, crossed posts with Steppenwolf saying exactly the same thing!

[steppenwolf]

EDIT: Sorry, crossed posts with Steppenwolf saying exactly the same thing!

One of us must be right then! :grin:

[Chris]

Apart from taking flats as mentioned above, if you are unguided like myself and you want more out of imaging Nebula with short exposures you might wan't to look into modding your camera by removing the IR cut filter, its a bit fiddly and takes a few hours but it significanlty improves the signal for the red stuff. Theres a good guide on modding various Canon cameras by Gary Honis. If you also want to keep the camera for daytime use you can swap the filter you remove for a Baader replacement filter which will allow your autofocus to still work properly, you just have to set the cameras white balance to Tungsten to get rid of the pink cast for daytime pici's.

Heres 30x90seconds on the horse and flame with my Modded Canon 350D, no replacement filter.

I didn't manage to open your pic so it will be embarrasing if I'm giving you advice and yours is better than mine, this one needs flats re-doing as well D

[Epicycle]

Many thanks for the quick responses, that's very useful! It confirms that is not just me being inoperant with Gimp, but simply that I need to take longer exposures and add flats.

Starfox: thanks for the suggestion about modding, it's one obvious thing to try but I feel a lttle reluctant to do it myself, I might send the camera to Andy Ellis though,

cheers

E.

[Chris]

I know what you mean about being reluctant, I darn't mod my 1100D just incase I broke it! so I bought a very cheap 300D from Ebay so I could give it a go at low risk, turned out well thankfully:) like you say if you just up the exposure time and add flats that will do wonders on its own, I'm not sure what kit your using but if you can get up to at least 60 second subs then you should be happier. if you get time can you post a jpeg or PNG on here , it would be interesting to see what very short subs show exactly? I don't know why I can't see your pic on drop box?

[Epicycle]

Hi Starfox: once you click on the link above it will take you to a drop box page, to see the image you need to download it by clicking on the upper right corner. To see the HH you may need

to may need to stretch it further and make a bit of an effort to spot it.

My equippment is a Megrez 72 and a Canon 1100, but my mount is very inadecuate for long exposures (I should remedy this soon I hope),

E.

[steppenwolf]

I don't know why I can't see your pic on drop box?

Did you click the 'download' button?

[Chris]

Oh yes thats done it thanks:) I ignored the download button thinking it meant do you wan't to download drop box:D

Yes the Horse is definately hiding away in the Vignetting.

[Epicycle]

Yes, it's hidding in the vignetting. One last basic question. Suppose I keep fix the exposure time to 20sec, since SNR is proportional to the square root of the number N of exposures one expects

to be able to resolve more detail just by increasing N. Any idea about the minimal N is so that I will have a passable detail of the HH? Maybe some of you have experimented with this.

Thanks,

E.

[Chris]

would a signal to noise ratio of 31 be enough? because you would need 1000 exposures!

I guess you would have to estimate what percentage of noise pixels in an image would begin to ruin an image, if it was say 5% then you would need 400 exposures, I think this would be about right if one in twenty pixels on my TV were noise I would still be able to make out whats going on with quite a lot of detail, I think?

[Epicycle]

400 sounds about right. I may be wrong with this, but I was thinking that SNR is also proportional to exposure time so everything else being equal to get something like your picture above

one would need N=(4.5)^2 x30 which is a bit over 600. But your picture is certainly more than passable! So 2 nights of hard work might do the trick...

[Chris]

I think if you don't mind doing the graft it would be a very interesting little experiment, I've certianly not seen anyone stack any where near 400-600 short 20-30sec subs before! it would be good to see if the estimated math matches the practical results, other uncertanties aside like seeing etc

Good luck if you decide to go for it:) I think the most 20 second subs I've taken on one object was 60 x20 seconds on the Leo Triplet. I did 160 x 120seconds on M51 and that looked pretty smooth:)

Chris

[Epicycle]

Well, I've done a partial experiment. Below is a stack of 128x20s of HorseHead and Flame. The HH is now much more apparent and if my

Canon 1100 were modded it would be a lot more colorful! I think I'll pass on stacking 400 subs! But this one shows that if I were

to do so I would get something roughly twice as a good which is consistent with the maths above.

Anyway, I'm not that good at processing, so if anyone would like to have a go at the stacked file to see if more detail can brought out it would

be great, Here is the link:

https://www.dropbox....1:Flame_SL.jpg]

[Epicycle]

This seems a better attachment.

[Jannis]

I think if you don't mind doing the graft it would be a very interesting little experiment, I've certianly not seen anyone stack any where near 400-600 short 20-30sec subs before! it would be good to see if the estimated math matches the practical results, other uncertanties aside like seeing etc

Good luck if you decide to go for it:) I think the most 20 second subs I've taken on one object was 60 x20 seconds on the Leo Triplet. I did 160 x 120seconds on M51 and that looked pretty smooth:)

Chris

Well, this might be a bit off topic; but i'm currently stacking about ~1000 6sec frames now (550d, 6 sec exposure @ ISO3200 50mm F/1.4@2.2), not sure if my data could help on prooving/disprooving the math about many short subs vs few long subs.. o.o

unsure how it will turn out though, as it's taking a good few hours to process. Stacking with Median Kappa-Sigma clipping (still haven't really understood when to use what stacking method in DSS, lol)

can upload a single raw and the stacked TIF to dropboks later if anyone would like to take a look at the data to calculate the difference in S/N ratio.

i'm only stacking the centre of the image though as i only used a normal tripod without any tracking, so it was moved manually over the ~2 hours it took to take the nerly 1100 frames.