Question about flats and LPR

[MishMich]

Given the setup for flats should be the same as for lights, if I image using a light pollution filter, then the flats should use this also?

M.

[jgs001]

That's correct. The imaging train must be exactly the same, including the focus point for taking flats.

[MishMich]

So, cannot be done in hindsight then.

I have a photographic 'white balancer', essentially a white circle of nylon material in a frame, about 20" across. If I wanted to use this, is it best done in daylight, and should it be held against the sky or receiving light from the sky, or at night with light shone from behind or lit from the front? If lit from behind, should it be held up to the telescope, or at a small distance? If I take flats during the day, how do I achieve the same focus the I will use on the stars at night, as there will be nothing equivalent to focus on?

Thanks.

[jgs001]

Really you should not change the configuration of the imaging system before taking flats, as even a slight misalignment can upset the process. As to the best way to produce them, there are all sorts of ideas and suggestions. I use notepad, maximised, on my lappy screen, held over the objective of my scope, set the camera to Aperture priority and fire away.

Another option is one of EL Panels

both of these work in the dark and therefore can be done whenever you have finished the run, prior to dis-assembly.

[themos]

Yes, after the proper imaging run when your battery that feeds the mount is dying, You don't need tracking for these! I tried the white t-shirt method a foot or so from the aperture illuminated by a camera flash. Vary the position of the flash with each frame so any biases are washed out.

[MishMich]

OK, thanks.

Next question. I have tried to frame my lights with darks. So, if I take 5 x 2 min followed by 5 x 3 min, it goes:

dark: 1 x 2 min

light: 5 x 2 min

dark: 1 x 2 min

dark: 1 x 3 min

light: 5 x 3 min

dark: 1 x 3 min

etc., but from what I have read, there should be an odd number of correspoding darks, and around 9 for median stacking. Looking at the way DSS works, darks are grouped alongside corresponding lights, so if I only take 5 frames, I'd end up with more darks. Of course, I could increase the number of lights - if time permitted, but often it doesn't. What is the solution? I take shorter and longer frames because it is not immediately obvious whether there are serious errors in the tracking, and find that I may get 4/5 good frames at 3 mins, or 1/5 - so not always easy to know until processing how many have actually worked out. Guiding will no doubt help eliminate that, but I am still inclined to work with two sets of exposures.

So, would 1 dark before and 2 after lights be sufficient, or will 2+3? Otherwise, most of the session will end up taking darks.

Then what about flats? I understand bias frames are very short exposures, but flats are supposed to be the same exposure & ISO as the lights, aren't they? Where does a flash come in to this? Do I need to work on flats in the same way as darks - uneven number of flats each set matching the exposure time(s) of the lights? Is three enough for each exposure? Is there a particular light that is better - I have a bright torch, I also have a portable neon light for using in dark places like the loft; it gives quite a diffuse light. Would that sort of light held behind the white balancer be OK? At the end sounds good - then I can leave things to snap away while I pack everything up.

Vignetting was not a problem, until I started pushing the processing as far as I could, then it became noticeable. I am wondering whether a dew-shield-extension can affect this? As it seems to have affected the one's I did with a dew-shield worse, and more one side than the other. For short small aperture scopes with a large FOV, are dew shields better constructed as a cone rather than straight out from the scope?

M.

[jgs001]

I'm not sure on the darks, I tend to stick to one exposure length per session. Even if it's short, lots of short subs will still get a good result, and it makes darks much easier. Just take 11 or so at the end of the session. As for the flats, they need to be so the histogram peaks at about 70% on the SLR. The easiest way to get this, is to set the Camera to Aperture Priority mode, this will give you the shutter speed you need (I got that info from the DSS help).

[MishMich]

Thanks,

just want to make sure I have everything right & firmly in my head for next time I get a chance to see the sky - I have done darks, but not flats or bias frames - so I can get the most out of it.

So you only need one set of flats, exposure time of lights is irrelevant for them, it is that they are at the same focus. Makes more sense now. Makes more sense to do at end of session as well.

The only exposure-matching is the darks. I'll try and stick to one length of exposure, it will make things easier.

OK, what about flat-darks? Are these necessary? Presumably they have to match the flats' exposure?

Last session, I got a dim flame and hint of horsehead. Any ideas what sort of lengths exposure I'd need to brighten that up a bit? I think I ended up with about 20 minutes' worth of usable frames in all.

I really need to move on from M42 and M45 until I am better at this, and M31 and M33 just don't seem to be in a good position for me. I need to start looking for other things to image, but I'm pretty limited to south, SW, SE, and straight up. Any suggestions for other things I could be looking at at the moment?

M.

[jgs001]

M, That sounds about right... For the flats though, don't move anything... the components need to remain in the same optical alignement also. You are correct for flat darks. I used to do those, but I don't anymore. The exposures are so short for the flats for my setup, that there doesn't seem to be any point.

Have a look at some clusters... There's some lovely ones in Auriga M36, M37, M38 (see if you can get the smiley face in shot with M38). The Double cluster between Cassie and Perseus is another beaut. M35 in Gemini also.

I tried an HH and flame with my unmodded camera last year. I didn't get many subs, and was stretching to 3 minutes and it was barely there. I'm planning on trying again at some point, possibly not this year though and aiming for 10 to 15 minutes. Although I suspect that Alnitak will be a really big problem at those lengths...

[simon hicks]

M,

The Darks are the most difficult IMHO! They should be at the same exposure length, ISO setting (or gain) and temperature as the Lights. And the more the better....10 or so being a good starting point. Its usually the temperature bit that makes them difficult. So just get them 'close' and go for it.

Cheers

Simon

[Catanonia]

M,

The Darks are the most difficult IMHO! They should be at the same exposure length, ISO setting (or gain) and temperature as the Lights. And the more the better....10 or so being a good starting point. Its usually the temperature bit that makes them difficult. So just get them 'close' and go for it.

Cheers

Simon

eaiser if you have a timing mechanism such as the EOS Utility and it's batch processing.

The other night I took a load of images of 30mins each but come processing time had no darks of 30mins.

So next night was raining, went to observatory, put lense cap on and left it imaging via EOS utility 10 x 30min darks

You will find that you will stick to certain length subs and ISO

I always use ISO 800 and subs of 5, 10, 15 or 30mins so very easy for me.

Easiest night imaging i ever did

[themos]

flats are supposed to be the same exposure & ISO as the lights,

NO! Flats can be short exposure. They are the only ones that need doing at each session. Darks and bias can be shot any time (but you need to watch the temperature for darks and match them with the temperature+ISO+exposure of the light frames).

[Psychobilly]

I dont bother with darks anymore... but thats just me....and i'm not going to discus the reasons...

If you do take your darks in daylight watch that the the cover on the scope is properly light tight and also the cameras viewfinder etc..

I had a lovely set of gold tinted images of the inside of my Megrez 72 cover where the light had leaked through the felt thats around the inside edge...

Billy...

[MishMich]

Yes, thanks, I got to there further on in the thread. This was the bit that was confusing me.

Flats - short exposure (depending on light, appx midway in the gauge, or 70% on the histogram), several of them. Need taking with same setup, focus, filters, etc., while attached to telescope, but at lowest ISO.

Darks - same exposure as lights, same ISO, ideally same temperature, taken before and after a sequence of lights (or at intervals if lots) - but not necessarily in same session. No need to be attached to scope.

Bias - shortest exposure time possible at same ISO as lights (not sure how many) - can be taken any time, any conditions. No need to be attached to scope.

Flats best taken before dusk, but easiest taken using artificial light source at end of imaging session.

Flat darks should match ISO & exposure of flats, but not necessarily worth bothering with.

The only thing I'm still a bit vague about is the numbers.

11-19 darks

How many flats?

How many bias?

M.

[jgs001]

I don't use bias frames. I always got a weird and hard to process out green tint.

[themos]

I've no idea what flat-darks are. Flats require light, darks require darkness, how can flat-darks exist?

I'd stick with the same ISO for flats and make sure they don't' have ANY saturated pixels.

Shoot 100 bias, they're easy.

With flats, as many as you can but at least 10, I'd say, and up to 50.

[themos]

I don't use bias frames. I always got a weird and hard to process out green tint.

Even after PixInsight LE's green killer?

[simon hicks]

The only thing I'm still a bit vague about is the numbers.

11-19 darks

How many flats?

How many bias?

The numbers really depend on what is practical and how much impact any more of them will have. You take lots of them so that when they are combined to create a master flat or master dark or master bias the effect of combining them reduces the random noise present in each individual frame giving a lower noise master. The randon noise goes as 1/n^2, where n is the number of frames. So to halve the randon noise you need to square the number of frames. So if you combine 4 dark frames you get roughly half the random noise as two dark frames. To halve the noise again you need to increase the number of frames to 16....so its gets more and more difficult to have a big effect.

So essentially you want as many as possible. Its easy to run off say 50 bias frames....they only take maybe 1/4000th of a second each plus maybe a few seconds to download each one. And flats are usually only a second or two....so again its easy to run off a bunch of 50.

However, if your Darks take 10 minutes each, then running off 50 would take many hours.....so you usually put up with less....maybe 10 or 20. Just a case of practicality.

And there's no 'magic numbers' of them....you don't need to stick to even or odd numbers of frames.

Cheers

Simon

[MishMich]

Flat-darks (or dark-flats?) are darks for the flats, which deal with the sensor noise and hot pixels in the flats.

Presumably I don't need to bother with them?

M.

[simon hicks]

You don;t 'usually' need to bother with them. If your flats are 1 second then this won't usually be enough time for any significant dark signal to build up....so no need for them.

But if your flats are say 30s or longer (maybe for very narrow band imaging) then the dark signal could be significant and it may be good to include them.

So it depends on the exposure time of the flats.

[MishMich]

Ah. That makes sense. So, only need flats and darks. M.

[MishMich]

And maybe even not darks?

Thanks Peter, my reply was to the post before yours, BTW.

I get the idea that the process of stacking would deal with any hot pixels and noise, unless tracking is so accurate that it reproduces sensor errors across many frames, which then get replicated in the result. But I did notice in my first subs that there is a build-up of glow around the sensor terminals. This appeared to get worse the longer the exposures and the longer I image (I was imaging with live-view enabled, now I switch it off once everything is set up). When I stacked these, it was reproduced, but not obvious, but with image enhancement, particularly DSO enhancement like in Noel's actions, it was amplified. Before I tried darks, I had to crop this out, but with darks I found it was dealt with.

M.

[Catanonia]

Well I learnt something too..

My flats are under 1 second, so no need for dark flats...... sweet.

Question, BIAS at 1/4000th second, i assume these are with lense cap OFF....

[Merlin66]

A Bias is a very short duration Dark. ie no light

I just compensates for the CCD read noise as the image is "moved" from the CCD.

Ken

[Catanonia]

arhhhhh, gotcha... I was doing darkflats or thought i was... Aperture mode with lense cap on... A mix really.

Ok so 1/4000th second with cap on = Bias Sweet makes it MUCH easier