Return to the Cloudy Nights Telescope Reviews home page


Astrophotography and Sketching >> CCD Imaging & Processing

Pages: 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | (show all)
Inverted
sage
*****

Reged: 01/19/13

Loc: LP Land
Re: 5 or 30 minute exposures for narrowband? new [Re: pfile]
      #5657003 - 02/01/13 07:21 PM

Quote:

Quote:


Edit: by the way, just to clarify, I'm asuming there is a higher ratio of pattern noise to random noise. I think that was obvious, it re-reading I see how it could be interpreted that I am saying we are calibrating out random noise. Of course you can't. That's why I keep saying if we have low read noise AND good calibration. As I said I sti don't have a sense o the magnitudes. That is what I'm interested in though. Hopefully once I have a chance to read Craig's article, that will be more clear.




here is how you quantify the read noise.

make a really good master bias frame.
subtract the master bias frame from a bias sub.
the result is the read noise in that bias sub.




Thanks that makes sense now and looking back, I see Frank mentioned that the FPN is technically considered separate from the read noise. I think in practice though, it seems if people are taking insufficient bias frames, then a significant amount of FPN will remain in the bias subs. I think this may look like random noise, until it converges into a fixed pattern though. Either way, it could be worse than the poison noise, if your not modeling the FPN well as fixed noise tends to be more noticeable than random, even if it "looks" roughly random. It would also seem that it would make the read noise look higher than it is, if measured how your describing though. Again unless I'm missing something...


Post Extras: Print Post   Remind Me!   Notify Moderator  
vpcirc
Post Laureate


Reged: 12/09/09

Loc: Merced CA
Re: 5 or 30 minute exposures for narrowband? new [Re: Inverted]
      #5657130 - 02/01/13 09:04 PM

I just had a long discussion with a very advanced imager who shared a few of his thoughts and experiance.
1. Science says one thing reality says another.
2. Exposures times necessary can change from night to night and can be dramatically different depending on your equipment so there is no easy answer to "how long and how many"
3. There is no such thing as too long with narrowband
4. Longer images and more of them as a general rule will always produce higher contrast and more resolution than shorter images with the same total time.
5. With modern rejection methods, even bad data is usuable to a point.
6. It never hurts to add in extra data and the limit on how much is not proven.
7. It's now very common for luminance data for example to contain 20 hours of data.

I eat a little crow stating the 16 frame rule, but I do feel correct in the philosophy that total time does not equal total time. The longest images you can take with your given equipment to reach that total time will achieve far great results than stacking a bunch of short exposures.


Post Extras: Print Post   Remind Me!   Notify Moderator  
korborh
scholastic sledgehammer
*****

Reged: 01/29/11

Loc: Arizona
Re: 5 or 30 minute exposures for narrowband? new [Re: vpcirc]
      #5657200 - 02/01/13 09:50 PM

I disagree with #1: Science is based on observations from reality (nature).

Regarding #3, there are lots of other reasons to avoid very long subs (tracking issues, focus shift, star saturation, airplanes, cosmic rays etc.). Question is for a particular equipment, shorter exposures may indeed be better than longer (for the final outcome) so going too long with narrow-band will not always apply.

#4 cannot be always true (true if read noise is too large). Not sure about resolution - like mentioned above, long exposures have more chances of being messed up (tracking, focus etc.) that can make resolution worse.

Edited by korborh (02/01/13 10:08 PM)


Post Extras: Print Post   Remind Me!   Notify Moderator  
David PavlichAdministrator
Transmographied
*****

Reged: 05/18/05

Loc: Mandeville, LA USA
Re: 5 or 30 minute exposures for narrowband? new [Re: korborh]
      #5657327 - 02/01/13 10:59 PM

Like I said, you expose for as long as your equipment and skyglow (NB being the exception) will allow or to the point of diminished returns.

David


Post Extras: Print Post   Remind Me!   Notify Moderator  
Alph
Carpal Tunnel


Reged: 11/23/06

Loc: Melmac
Re: 5 or 30 minute exposures for narrowband? new [Re: vpcirc]
      #5658696 - 02/02/13 06:30 PM

Quote:

I eat a little crow stating the 16 frame rule,




It wasnít really your statement. You simply quoted someone you thought was an authority on the subject. This goes to show what you read in hobby magazines has to be taken with a grain of salt.


Quote:

but I do feel correct in the philosophy that total time does not equal total time. The longest images you can take with your given equipment to reach that total time will achieve far great results than stacking a bunch of short exposures.




It depends. If the Poisson noise dominates the readout noise then stacking will be statistically equivalent to one longer exposure. It is simple arithmetic. If your anonymous SME knows something what scientists donít know then you should encourage him to publish a paper in a professional journal.


Post Extras: Print Post   Remind Me!   Notify Moderator  
vpcirc
Post Laureate


Reged: 12/09/09

Loc: Merced CA
Re: 5 or 30 minute exposures for narrowband? new [Re: Alph]
      #5658748 - 02/02/13 07:16 PM

I should clarify that. When he said total time doesn't equal total time, 10 hours of 10 min shots will not show the same resolution and contrast as 10 hours of 15 min shots. I don't think he meant from a noise perspective.

Post Extras: Print Post   Remind Me!   Notify Moderator  
cn register 5
scholastic sledgehammer


Reged: 12/26/12

Re: 5 or 30 minute exposures for narrowband? new [Re: vpcirc]
      #5658813 - 02/02/13 07:56 PM

Why? What evidence do you, or this unknown prophet, have?

Post Extras: Print Post   Remind Me!   Notify Moderator  
orion69
scholastic sledgehammer


Reged: 05/09/10

Loc: Croatia
Re: 5 or 30 minute exposures for narrowband? new [Re: cn register 5]
      #5658839 - 02/02/13 08:09 PM

Quote:

Why? What evidence do you, or this unknown prophet, have?




Why don't you try if you don't believe. Of course that depends of the sky quality. If sky is poor and your subs are too long noise would prevail.

Imagine this: you have Astrodon 3nm NB filter and shoot one object for 6 hours. First you take 180x1 min and then 6x30min subs. If sky is fair do you really think that those series of subs would produce image of same quality?


Post Extras: Print Post   Remind Me!   Notify Moderator  
vpcirc
Post Laureate


Reged: 12/09/09

Loc: Merced CA
Re: 5 or 30 minute exposures for narrowband? new [Re: orion69]
      #5658895 - 02/02/13 08:43 PM

If you have any questions in regards to narrowband I suggest contacting Don Goldman at Astrodon. He is an optics expert and would be able to answer your questions much better than I from a scientific standpoint. I'm just a struggling imager trying to figure this all out myself.

Post Extras: Print Post   Remind Me!   Notify Moderator  
Inverted
sage
*****

Reged: 01/19/13

Loc: LP Land
Re: 5 or 30 minute exposures for narrowband? new [Re: orion69]
      #5658919 - 02/02/13 08:59 PM

Quote:

Quote:

Why? What evidence do you, or this unknown prophet, have?




Why don't you try if you don't believe. Of course that depends of the sky quality. If sky is poor and your subs are too long noise would prevail.

Imagine this: you have Astrodon 3nm NB filter and shoot one object for 6 hours. First you take 180x1 min and then 6x30min subs. If sky is fair do you really think that those series of subs would produce image of same quality?




I wish the sky around here was fair. Last night I had to throw out 20% of my subs due to passing clouds, that would have really sucked if a cloud passed through a 180 minute exposure. Also, cosmic rays, satalites etc.. are a problem everywhere, as are moments of bad seeing IME. Further not everyone's equipment will work consistently well for that long of an exposure. The good thing though is camera noise is becoming lower and lower and less of an issue if stacking images (and many technuqies such as sigma stacking and dithering may not work well without at least a reasonable amount of subs, maybe somewhere around 10 or so subs...) From the website,Mike posted as a reference, therer are examples using real world data, which showed cutting the exposure time in half, may only increase the required total integration time by a measly 10%. the irony is they even made recommendations similar to the ones people have been making in this thread.http://www.hiddenloft.com/notes/SubExposures.pdf

No one doubts that under perfect circumstances with perfect tracking and perfect skies going long with nb will improve the image some. In the real world things aren't perfect, I think that is all anyone has said in this thread.

(Also BTW, I think a big part of the issue hasn't been that Mike is saying it's better to go long, everyone is entitled to their opinion, and in many cases I'm sure that's true. I believe much of the discussion stared via his insistence that you have to go long and anyone that says otherwise is apparently wrong. I know I'm not right all the time, that's why discussions like this are useful, I really learned a lot once I finally put two and two together, but if you approach it with your mind made up already, then it all doesn't matter...)


Post Extras: Print Post   Remind Me!   Notify Moderator  
vpcirc
Post Laureate


Reged: 12/09/09

Loc: Merced CA
Re: 5 or 30 minute exposures for narrowband? new [Re: Inverted]
      #5658956 - 02/02/13 09:29 PM

That would be a misunderstanding. I said in my case John Smith and others all told me I needed to go longer with my camera and my dark skies after looking at my data. Longer images in narrowband, from what I'm being told, will produce far better contrast and resolution. The same applies to broadband but has more limits based on equipment skies etc. the disagreement has been over the benefits of long vs short. I know for example Don Goldman shoots his narrowband subs at 30 min. Why, because even though he could easily shoot far longer but he doesn't want to lose images to satellite trails etc. The majority of us firmly believe shooting longer duration images will produce better images. Too often I see new imagers wanting to believe its just as good to take short images because when they try to shoot a 10 min sub they have star trails and other issues because they haven't learned or its too much of a pain to properly polar align their mount. I can speak from experiance, it wasn't that long ago that was me.

Post Extras: Print Post   Remind Me!   Notify Moderator  
Inverted
sage
*****

Reged: 01/19/13

Loc: LP Land
Re: 5 or 30 minute exposures for narrowband? new [Re: vpcirc]
      #5659055 - 02/02/13 10:45 PM

No worries, I've learned a lot from the resultant discussion, which is all I really care about. I'm not really sure what the continued debate is about then though but that is another matter.

Post Extras: Print Post   Remind Me!   Notify Moderator  
freestar8n
Post Laureate
*****

Reged: 10/12/07

Re: 5 or 30 minute exposures for narrowband? new [Re: Inverted]
      #5659325 - 02/03/13 03:26 AM

I have put some plots together to help convey this stuff in pictures, which maybe I should have done a long time ago. This is all based on the core, simple noise model used by many web sites to calculate the impact of read noise - so if I made a mistake I welcome corrections - but if people have a problem with the noise model itself - I am simply using it, not saying it is perfect.

For people who want to do a similar calculation based on their own parameters, please take note of the following:

1) Measure your own camera parameters (gain/read noise) because they may differ from spec. It is easy to do and you don't need master frames - just a few biases and lights.

2) If you bin, measure the gain and read noise in binned mode also because the results may not be what you expect.

3) To measure sky background, make sure you measure with a calibrated frame so you don't include the bias offset in the sky background. Checking the pixel value of a raw light vs. a raw bias will help confirm that the delta you measure is due to sky background. You want to measure with a "black" area of the light.

The main points of the noise model are:

1) The dominant noise is due to sky background and read noise

2) The sky signal increases linearly with exposure time, and since it is Poisson the sky noise goes as the sqrt of exposure time

3) Read noise only contributes once when each exposure is read, and is independent of exposure time

4) The sky and read noise add in quadrature rather than a simple sum. So if they are both 10, the combined noise is 14. If one is 10 and the other is 20, the combined noise is 22.3.

The following plots are only of the background noise (sky+read) - and they don't show signal at all. SNR is not described here - only background noise. What is important here, and in calculating subexposure, is the *relative impact of read noise* on the total noise. I show this as the ReadImpact value, which is the added impact of read noise divided by the sky noise. There is no hard value for what is acceptable here - but if the sky noise is big, there is little to be gained by reducing the read noise impact to a very small value.

Frank


Post Extras: Print Post   Remind Me!   Notify Moderator  
freestar8n
Post Laureate
*****

Reged: 10/12/07

Re: 5 or 30 minute exposures for narrowband? new [Re: freestar8n]
      #5659328 - 02/03/13 03:30 AM Attachment (17 downloads)

Here is a plot showing a high noise camera at a bright site and the benefit of increasing the sub exposure time from 100s to 300s. The blue curve is the sky background noise and the black arrows are the read noise being added at the end of each subexposure. The red curve shows the combined noise. The gap from the blue to the red curve is the penalty you get from read noise. If this is big compared to sky noise, then you are being hurt by read noise and should use longer exposures. But if the sky noise is big already - longer exposures won't help much.

In this case, longer exposures are a help, and lower the total noise.

Frank


Post Extras: Print Post   Remind Me!   Notify Moderator  
freestar8n
Post Laureate
*****

Reged: 10/12/07

Re: 5 or 30 minute exposures for narrowband? new [Re: freestar8n]
      #5659331 - 02/03/13 03:34 AM Attachment (11 downloads)

Here is the same situation, but at a dark site. Note that longer exposures have a much greater impact on reducing the noise, so you have a bigger motivation to go longer. But in each case, the total noise is less than the noise at a bright site. This helps convey that a dark site you want to go longer - but even if you don't, you will benefit from the dark site.

Frank


Post Extras: Print Post   Remind Me!   Notify Moderator  
freestar8n
Post Laureate
*****

Reged: 10/12/07

Re: 5 or 30 minute exposures for narrowband? new [Re: freestar8n]
      #5659333 - 02/03/13 03:37 AM Attachment (11 downloads)

Here is one more, showing a high read noise camera at a dark site, vs. a low read noise camera at a bright site - with the same exposure. Note that the total noise is the same - but you would have been much better off with the low noise camera at the dark site.

Frank


Post Extras: Print Post   Remind Me!   Notify Moderator  
cn register 5
scholastic sledgehammer


Reged: 12/26/12

Re: 5 or 30 minute exposures for narrowband? new [Re: freestar8n]
      #5659463 - 02/03/13 07:37 AM

Thanks for this Frank, your bright sky, high noise example seems to be fairly close to what I have. Going from 100 to 500 secs has an obvious good effect but it looks as if going much longer than this is at the point of diminishing returns because there's not much room between the red and blue lines and that's all that can possibly be gained by exposing for longer.

The dark sky results show that increasing exposure beyond the 5 minutes of your example would be useful, and this explains why the experts at dark sky sites insist that longer exposures are beneficial, they are much more so for them.

I wonder, how long professional astronomers expose for, they have perfect equipment at very dark sites and are trying to get the most out of very dim objects. Does anyone know how long the individual exposures in the Hubble deep field were?

As for the claim that increasing exposure length, even from from 10 to 15 minutes, makes a difference to resolution and contrast we have been given no evidence that this is so, other than claims that experts say so. I can't see any rational reason for this so am not prepared to take the word of experts without any reason.

Chris


Post Extras: Print Post   Remind Me!   Notify Moderator  
freestar8n
Post Laureate
*****

Reged: 10/12/07

Re: 5 or 30 minute exposures for narrowband? new [Re: cn register 5]
      #5659482 - 02/03/13 07:56 AM Attachment (5 downloads)

Thanks - I think that plot does probably match your situation - especially since you binned - which effectively increases your sky flux, or effectively halves your f/number. But it may also increase your read noise so I recommend measuring it.

Here is another plot with a longer time scale of 2 hours, showing that same 300s subexposure plot (5 minutes) compared to 20 minute subexposures. The noise is heavily dominated by sky and the increase from 5m to 20m only drops the noise by 10, from 98 to 88. It's not clear to me this improvement would even be visible after stretching - but even if it was visible - you would just stretch a bit less and it still wouldn't have much tangible effect.

Frank


Post Extras: Print Post   Remind Me!   Notify Moderator  
Inverted
sage
*****

Reged: 01/19/13

Loc: LP Land
Re: 5 or 30 minute exposures for narrowband? new [Re: freestar8n]
      #5659518 - 02/03/13 08:44 AM

Thanks Frank great work! It really helps to visualize the models impact, rather than just talk about parameters. I'm curious how much changing the read noise would effect the dark site and bright sites. What if we had a 5e- or 3e- camera? Surveying cameras with listed read noise, 10e seems to be the high end for any camera on the market today. The average looks closer to 7 and even relatively inexpensive Atiks median is about 5, and go down to 3. Heck, if you really want, I think Apogee has some high end cameras <2e-

Post Extras: Print Post   Remind Me!   Notify Moderator  
freestar8n
Post Laureate
*****

Reged: 10/12/07

Re: 5 or 30 minute exposures for narrowband? new [Re: Inverted]
      #5659531 - 02/03/13 08:52 AM

Thanks - yes, I sure hope this clarifies the discussion and key points.

As for lower read noise - as I've been saying - it all depends on your camera, optics, and site. You can turn a "bright" site into a dark one just by increasing the f/ratio or using a narrower filter. It's all about the actual sky flux you are receiving and the actual read noise you have.

The equations involved are very simple - but there are some confusing sub-topics. The most confusing one is "optimal sub-exposure length." A long one may indicate a problematic situation - or a very good one.

I mainly encourage people to take some careful measurements if they are interested in this stuff - and be sure to do all calculations in electrons. It is a big no-no to take the square root of adu to find the noise. No, no, no. Must be converted to electrons first.

Frank


Post Extras: Print Post   Remind Me!   Notify Moderator  
Pages: 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | (show all)


Extra information
14 registered and 20 anonymous users are browsing this forum.

Moderator:  Oldfield, droid, Scott in NC, bilgebay, WOBentley 

Print Thread

Forum Permissions
      You cannot start new topics
      You cannot reply to topics
      HTML is disabled
      UBBCode is enabled


Thread views: 7584

Jump to

CN Forums Home




Cloudy Nights Sponsor: Astronomics