Some examples of flats please.

[ollypenrice]

I think you should point away from the sun. But don't ask me, I can't do 'em!!

A DSLR should give you a histogram on its screen. The peak should be a third of the way to the right.

Olly

[Uranium235]

Makes perfect sense thanks !

 do you think pointing the scope directly up might help ?

Rob

Yes, point it straight up for sky flats. And you have to be quick becuase the brightness level is dropping constantly.

Also, Im not sure you should be doing your flats at a different ISO to your lights - it might introduce a noise issue.

[Robp]

I also think I need to improve my flat taking.

Can you use any LED light panel to take flats for example these:

http://www.amazon.co.uk/Amzdeal-Recessed-Ultra-slim-Corridor-Conference/dp/B00LUI1MWQ/ref=sr_1_4?ie=UTF8&qid=1425472635&sr=8-4&keywords=LED+Panel

http://www.amazon.co.uk/LED-light-panel-30x-30/dp/B0094G2NUG/ref=sr_1_3?ie=UTF8&qid=1425472635&sr=8-3&keywords=LED+Panel

or does it have to have specific properties?

I usually try and use my laptop screen to make the flats but that is so awkward. Somethinng dedicated would be lots better.

[Xplode]

I also think I need to improve my flat taking.

Can you use any LED light panel to take flats for example these:

http://www.amazon.co.uk/Amzdeal-Recessed-Ultra-slim-Corridor-Conference/dp/B00LUI1MWQ/ref=sr_1_4?ie=UTF8&qid=1425472635&sr=8-4&keywords=LED+Panel

http://www.amazon.co.uk/LED-light-panel-30x-30/dp/B0094G2NUG/ref=sr_1_3?ie=UTF8&qid=1425472635&sr=8-3&keywords=LED+Panel

or does it have to have specific properties?

I usually try and use my laptop screen to make the flats but that is so awkward. Somethinng dedicated would be lots better.

Those panels will probably work fine, but you might have a problem with them being too bright so you could use some paper sheets between them and the scope.

For taking flats i use a panel like this (Huion L4S)

If your flat doesn't have darker corners and a brighter middle it won't be right. I get this using the flattest flatfield astrograph on the planet, the Tak FSQ106, and even so I get this kind of thing:

The peak of your histogram for flats should be between one third and two thirds of the way from left to right. I prefer to be nearer a third. It depends on the rig, but a third is always good for me.

I can't see any reason whatever for needing to use the same ISO but I don't image with DSLRs so maybe there is a reason?

Flats should be calibrated with their own darks but the good news is that a master bias will do for a 'dark for flats.' There is no statistical difference between a master bias and a master short dark frame.

Olly

I think your vignetting comes from the filter wheel and adapters.

Here's a flat taken with my 100Q with my 6D.

I think the low vignetting comes from the scope being a little slower and that i have gone for the M68 system all the way to the Canon adapter, only there i convert to M48.

Vigneting on the bottom comes from the mirror.

[rharrison]

So here's my master flat I'm working with, as you can see it's adding some annoying gradient at the bottom.

40 exposures at 400iso 1/40 sec and 20 flat darks. I will later do 20 at the lowest iso.

The image is M81 a quick test with 8 3 min guided subs no calibration apart from the master flat. I plan to make this my main target over the next few weeks.

[rharrison]

Yes, point it straight up for sky flats. And you have to be quick becuase the brightness level is dropping constantly.

Also, Im not sure you should be doing your flats at a different ISO to your lights - it might introduce a noise issue.

Arghh see that's what I've heard ! I certainly don't want to add noise.

Same iso or not is the question at the moment. I think what Olly said makes sense I can't see how it could effect it especially as there would be such low noise at 100iso with flat darks too.

[Davey-T]

I can't see any reason whatever for needing to use the same ISO but I don't image with DSLRs so maybe there is a reason?

Olly

Some software (possibly Maxim) complains if they don't match.

Dave

[rharrison]

I think you should point away from the sun. But don't ask me, I can't do 'em!!

A DSLR should give you a histogram on its screen. The peak should be a third of the way to the right.

Olly

Ok so the peak just left of the middle, slightly dark... ?

[rharrison]

Some software (possibly Maxim) complains if they don't match.

Dave

Well DSS does too but surely if there's 20 at only 100iso and darks there would be such low noise anyway ?

[sharkmelley]

You guys have taught me something - I didn't know that DSS complains if your flats are not the same ISO as your lights - but then I don't use DSS.  I read elsewhere that if you take flat darks at the same ISO as your flats then DSS is happy again.  I can see why DSS does this and it does make sense.  It also really makes sense to take your flats at ISO 100 if you can.

Regarding the gradients at the bottom of the image, these are caused by the edge of the mirror that folds up in the way of the light path during an exposure.   It is really, really difficult to create a flat that totally removes that bottom gradient, so I removed my mirror:  http://www.markshelley.co.uk/Astronomy/Projects/Mirrorless/canon550mirrorless.html   It solves the gradient problem but you can never use the viewfinder or auto focus again

Mark

[Uranium235]

Yep there is a reason why DSS moans when you give it different ISO calibration frames, its something to do with the subtraction of read noise (which changes at different ISO settings). Typically, the read noise increases with a lower ISO setting. 

[rharrison]

You guys have taught me something - I didn't know that DSS complains if your flats are not the same ISO as your lights - but then I don't use DSS.  I read elsewhere that if you take flat darks at the same ISO as your flats then DSS is happy again.  I can see why DSS does this and it does make sense.  It also really makes sense to take your flats at ISO 100 if you can.

Regarding the gradients at the bottom of the image, these are caused by the edge of the mirror that folds up in the way of the light path during an exposure.   It is really, really difficult to create a flat that totally removes that bottom gradient, so I removed my mirror:  http://www.markshelley.co.uk/Astronomy/Projects/Mirrorless/canon550mirrorless.html   It solves the gradient problem but you can never use the viewfinder or auto focus again

Mark

Nice ! I don't think I'll be removing the mirror as it's my girlfriends camera haha

Is the gradient caused by the mirror moving or the fact that part of it obscures the light path when it's up if so bad design or what ?!

[sharkmelley]

Is the gradient caused by the mirror moving or the fact that part of it obscures the light path when it's up if so bad design or what ?!

It obscures the light path when it's tucked away.  Lenses are designed so the light paths don't hit the mirror so it is only a problem with telescopes and it's mainly the fast F-ratio scopes that are particularly troublesome.

[glowingturnip]

just as an aside, I've noticed that the batch-preprocessing script in Pixinsight doesn't seem to handle the flats properly, at least in my experience.

As far as I can tell, if you load up the script with all your flats, bias and darks in one go, then it will subtract the darks from the flats before producing a master flat, which is clearly wrong since darks and flats are wildly different exposure times.  This leads to a noisier flat which leads to a very noisy calibrated light frame since flats are applied multiplicatively.

Work-around seems to be to run the script twice.  First time, just load up all the bias frames and flat frames and one light, no darks - run that and generate the master bias and master flat (and you can bin the calibrated light you get out of it).  Then run it a second time with your new master bias, master flat, and all your darks and lights - seems to calibrate correctly then

[KennySharp]

I've been using an open Notepad on my pc to create flats. That was the advice I picked up from the Budget Astro website tutorial on calibration frames.

[KennySharp]

I've been using an open Notepad on my pc to create flats. That was the advice I picked up from the Budget Astro website tutorial on calibration frames.

[frugal]

just as an aside, I've noticed that the batch-preprocessing script in Pixinsight doesn't seem to handle the flats properly, at least in my experience.

As far as I can tell, if you load up the script with all your flats, bias and darks in one go, then it will subtract the darks from the flats before producing a master flat, which is clearly wrong since darks and flats are wildly different exposure times.  This leads to a noisier flat which leads to a very noisy calibrated light frame since flats are applied multiplicatively.

PI should scale the darks to the exposure time of the flats. So if you have 600s darks and 1/10s flats, then it will divide the dark by 6000 before applying it (I think it is slightly more complicated than that, but it appears to be the general approach).

[glowingturnip]

agreed, I've not had much luck with the scaling though (dark-optimise button clicked or unclicked)  think i'd rather use the bias as a dark for the flats

[AstroAdam]

I've found that flats vary considerably depending on what scope/camera combination you're using. 

My f/7 102ED produces reasonably 'flat' flats with my 1100D.  With the C8 with f6.3 focal reducer, I used to get very vignetted flats. 

My 70ED flats are almost completely 'flat' so to speak.  

This has been the same with whatever method I've used (t-shirt twilight, lightbox, laptop screen, iPad screen etc).  

[ollypenrice]

I think your vignetting comes from the filter wheel and adapters.

Here's a flat taken with my 100Q with my 6D.

flat_BINNING_1_integration_RGB_VNG.jpg

I think the low vignetting comes from the scope being a little slower and that i have gone for the M68 system all the way to the Canon adapter, only there i convert to M48.

Vigneting on the bottom comes from the mirror.

You're right Ole. I should have gone for unmounted filters rather than 2 inch mounted but I didn't realize that at the time. However, the parallel scope in the tandem uses unmounted and, after flat correction, there is nothing to choose between them. Of course, this shows the value of flats! The situation is also much better in the TEC140 with the same camera because the light cone is longer and thinner at F7 than at F5 in the Taks.

Olly

[sharkmelley]

I've been using an open Notepad on my pc to create flats. That was the advice I picked up from the Budget Astro website tutorial on calibration frames.

The poor man's light panel - it works very well indeed

Mark