Jump to content

When to call it a day on a target


swag72

Recommended Posts

I have captured the following data on my target

Luminance - 5.75 hours

RGB - Over 10.5 hours combined

So should I call it a day and move onto somethng else? I've not stacked or started processing it yet, so don't know what it looks like so far and whether it would benefit from any more data.

How do you decide when enough is enough? Would I gain much more benefit from additional data?

Link to comment
Share on other sites

Sara, it is a case of diminishing returns. More data means less noise, but only up to a point. So there does indeed come a time when you have enough data, but you wont know until you process it.

As a minimum I like to have 20 Light frames, but I prefer 30 to 50 where possible. It all gets a bit mathsy and square rooty for my liking if you get too deep into into the numbers :D

Longer exposure narrowband images can work well with lower numbers of subs. For very faint targets, like the really faint outer halo of M57 for instance, you can never have enough data!

Tim

Link to comment
Share on other sites

I think that Tim has hit the mail on the head for me. It seems that the number of subs is equally as important as the total exposure.

I've recently managed 60 minute subs, and now as a matter of course set a run of 20 or 30 minute subs going... But even with 5 hours of imaging time I only end up with 10 - 15 subs.

So when I come to process I'm expecting lots of data and little noise and I'm not getting it. Not enough subs to avoid noise...

So I agree that you'll never know till you process your data if you need more or not.

For me (and this goes against the grain and something I've only really thought about in the last few days) I think in future I'll be getting lots of different length. So the randomness of noise can be minimized.

Sorry for going off topic.

Ant

Link to comment
Share on other sites

An interesting thought Ant, so basically different length subs will all contain different amounts and noise patterns, stacking them together will allow the stacking software to eliminate it more effectively?

I think that Tim has hit the mail on the head for me. It seems that the number of subs is equally as important as the total exposure.

I've recently managed 60 minute subs, and now as a matter of course set a run of 20 or 30 minute subs going... But even with 5 hours of imaging time I only end up with 10 - 15 subs.

So when I come to process I'm expecting lots of data and little noise and I'm not getting it. Not enough subs to avoid noise...

So I agree that you'll never know till you process your data if you need more or not.

For me (and this goes against the grain and something I've only really thought about in the last few days) I think in future I'll be getting lots of different length. So the randomness of noise can be minimized.

Sorry for going off topic.

Ant

Link to comment
Share on other sites

I saw a graph somewhere showing an example of the effects of #subs on read noise; can't find the darn thing anywhere now... Essentially, the graph flattens out quite quickly, so the more subs you have, the more you will need to give any significant further effect.

Link to comment
Share on other sites

An interesting thought Ant, so basically different length subs will all contain different amounts and noise patterns, stacking them together will allow the stacking software to eliminate it more effectively?

I've not really had a chance to try it - but what got me thinking was a 4 x 30 minute sub final image had more noise in it than 25 subs of 5 minutes - I was able to process the 5m subs a lot smoother than the longer subs.

These were taken different night - and of course they may have been taken in different conditions / altitude etc...

But we constantly talk about more subs needed to avoid the random noise...

Would

30x5

21x10

3x60 (total 560 minutes, 9 hours)

be better than

9 x 60 minutes

I'm starting to think that maybe they would be - this may already have been discussed to death on here, I don't read all the thread noadays...

Ant

Link to comment
Share on other sites

I basically eneded up doubling the time spent on a target every year endign up with a min of 8 hours worth off OSC DSLR data and dependign onthe target a couple of hours of DSLR Ha for good measure...

In the end I was getting so little done with the weather we have been having in Wettest West Wales that I lost interest...

Going back to basics and Imaging with SLR Lenses (Fastish Primes) may just relight the fire ... especially as I will travel to darker skies to get the targets I can't reach from home...

Peter...

Link to comment
Share on other sites

If I told you John, you may think it will turn out OK with all the data I've got. Lets put it this way - The first lot of data and process was sent to the recycle bin, the processing was truely awful. If this doesn't come out much better then it will never even see the light of day!! So I won't raise the level of expectation!!!

Link to comment
Share on other sites

More subs of lower length will always give better results than less subs of long length due to signal to noise reduction. You might loose a bit of detail if you take this too literally.

I stopped going for very long subs after testing this theory out a few months ago. I compared 10m Ha to 30m Ha and there was not much difference in detail picked up, even at F2.8 that I am running at.

I now aim for 30+ subs per channel. I can get away with this as @F2.8 I can use low sub lengths such as 2 or 3 mins compared to 8 - 10 on the older MN190 @F5.7

HTH

Link to comment
Share on other sites

If the target is close to the brightness of the sky background / LP do you need to stretch it further to raise it above that background? Does that lead to more noise and so ideally more data is needed in the first place?

Link to comment
Share on other sites

Longer exposure narrowband images can work well with lower numbers of subs.

I think that on NB more is definitely better. I like get 5 to 6 hours per filter as a rough guide.

There does come a point when the major stuff stops improving and any improvements seem to be on firming up any edges.

Link to comment
Share on other sites

If the target is close to the brightness of the sky background / LP do you need to stretch it further to raise it above that background? Does that lead to more noise and so ideally more data is needed in the first place?

This is what I find to be true.

Rob

Link to comment
Share on other sites

I think that on NB more is definitely better. I like get 5 to 6 hours per filter as a rough guide.

There does come a point when the major stuff stops improving and any improvements seem to be on firming up any edges.

If I did 5 or 6 hours it would only be 5 or six subs :(

What I have found with very long exposure Ha images, is that the detail in the dark nebula really starts coming to the fore. The longest exposure I have taken so far is two hours I think, or maybe three. There does come a point when you get all you are going to get.

@Ant, not sure your thinking is correct there buddy, when you come to stacking your images you'll be throwing away signal and detail gained in the longer exposures by combining them with shorter subs surely?

@Sara - You will likely get hundreds of images and half done projects that never come to fruition. I'd say less than a third of the images I start ever see the day when they are fully processed. I have really good data from star parties 2 years ago I haven't got round to yet :D

Link to comment
Share on other sites

The trouble with really long subs is that stars get blown and bloated (not to mention planes and satellites). I always try and keep the brightest stars under control by not letting them burn out. Also I find that there comes a point when after stacking the image really does start to smooth out. Before this point the noise is noticeable, after not so. It will of course get smoother the more over the figure but that is a tipping point imo. I find that 12 images is the minimum and after that the improvement is very noticeable. I have done hour long subs and still found that to be true.

I would much rather have 12 x 20 minute subs than 4 x 60 minutes anytime.

Link to comment
Share on other sites

This thread is a fascinating read.

I think the amount of data and length of subs really is target dependant. The most data I've ever collected on one target is the Cone nebula at 5 hours of 5 minute subs. I was doing test processing as I gathered the data and found that up to about 4 hours, there was noticeable improvement in signal to noise ratio. The last hour of data seemed to make very little further improvement. (I was working with an unmodded DSLR at the time). I tend to stick to the limit of 4 hours now. I do find that the number of subs is very important as it is that which seems to dramatically reduce noise. I wouldn't consider less than about 25 for most targets.

Having said this, as I said before, it is target dependant. I've been imaging a few clusters recently and found that to retain star colour and to avoid star bloating, the optimum exposure seems to be about 1.5 minutes. As the stack doesn't need much stretching, I've managed to get results I'm happy with with as few as ten 1.5 minute subs. Longer subs would bring out mo0re fainter stars, but I don;t want them to distract from the main feature anyway.

Link to comment
Share on other sites

I use an unmodded DSLR and I found about 4-5 hours to be the limit too. My Rosette image was 71x 4min @ ISO800. I had an additional 30 or so 6min subs at ISO400 but it didn't help too much and introduced a moonlight gradient that I couldn't easily get rid of so I didn't use the extra subs in the end.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. By using this site, you agree to our Terms of Use.