Problem with flats

[blinky]

I took some flats tonight and added them to my stack of M51 data from last night. The problem is that when I stretch the image I get wierd streaks. This is only evident if I add Flats for some reason? I have included a stretched version of the mast flat file that DSS created and it looks pretty good to me!

[roundycat]

unless the flats are doing something weird in the software you are using what is the background adu in your sub frames? How well exposed is the galaxy? It looks a lot like co-related noise. Longer exposures called for.

Dennis

[blinky]

What and how do I measure ADU? I dont think I could go with longer subs, the light pollution last night was really bad and 8mins was pretty much as far as I could go!

[RobH]

What software are you using for capture?

Yes, 8 minutes should be plenty....how many subs?

Remember that you need to bias subtract your flats and your lights too, or they won't calibrate correctly.

Cheers

Rob

[Starflyer]

Are you doing bias frames as well Craig? Don't know if it will sort your problem but I know it's recommended that if you use flats you should also use bias.

Simple to do; same ISO, shortest exposure possible, lens cap on, take a good few (I do 50-60) and add them to DSS.

Edit: Rob beat me to it

Hope you get it sorted,

Ian

[blinky]

That could be the problem then! No I aint taken any BIAS frames....

[RobH]

It doesn't look like a bias problem.

That will normally manifest by the flat either under or overcompensating.

Did you subtract darks.

With 8 minute subs it is likely you'll need to dark subtract if you're using a dslr.

I don't need to with my Atik16HR but that is very low noise.

When I used to use a DSI, I had similar streaks if the darks weren't matched to the lights temperature wise, even by as little as 2 degrees.

Cheers

Rob

[blinky]

I did use darks and they should have been around the same temperature. I took the darks at the end of the session so maybe it was a bit colder then.

TBH I have taken some bias frames and if they dont work I think I will dump the idea of flat frames alltogether!

[RobH]

Don't give up on flats!!!

It took me 7 months of trying to figure them out until I realised that you needed bias frames too!

The flat you have there looks fine.

What capture software are you using?

Rob

[blinky]

I tootk some bias frames before packing up last night so we will see how they go. I only ever use DSS for stacking. is there something better out there?

[blinky]

Well I just tried it with the BIAS Frames. These were 1/4000th of a second @ ISO 1600 same speed as the lights/darks.

It appears to have made it no better! What am I doing wrong?

[RobH]

I'll be home early next week if you want to send me some fits files....a couple of lights, flats, darks and bias.

I'll give them a look over in maxim if you want.

Cheers

Rob

PS....you're using DSS....Just noticed....sorry!

[Tim]

I used to get that with a dslr Craig, never really got to the bottom of it.

I would suspect that the streaks are actually in the lights, but that only by adding flats do you see how 'bad' the streaks are.

What stacking method did you use?

Are the streaks worst in the red channel by any chance?

as for the flats, how did you gauge the exposure with a dslr?

[blinky]

The streaks are worst in the red channel as fort he exposure I just looked at the histogram on the 350d's screen and made sure it was not overexposed.

[Tim]

if its just with flats, forget them and use pixinsight for getting rid of vignetting? You'll need to get the chip cleaner though to eliminate bunnies.

[blinky]

The chip is now nice and clean! I do use Pixinsight but was wanting real flats as I assume they are better.

[ngc2403]

Flats do more than just cure vignetting they get rid of fixed pattern noise and high signal levels. It looks to me like the bias frames are not doing there job right?

try stacking lots of flats, then lots of bias and then save them as master bias and master flat

next do the same for the master dark and remember you need lots to get rid of that noise.

stick them into DSS and you should get a better result by;

removing the bias from all lights, darks, flats,

subtracting darks from the lights,

dividing the lights by the flats

done.

if you are lost, just look at it this way. Noise is just four things;

Read noise - fixed by bias - a constant in every image

dark noise - no prizes here fixed by darks - increases over time

fixed pattern noise - fixed by flats - Vignetting / uneven lighting, pixel problems

shot noise - fixed by nothing - this is a fundamental limit to the camera

hope this helps you

[narrowbandpaul]

hi craig.

ADU, or DN, is the brightness of each pixel in the image. It is the number the camera spits out once it has finished measuring the electrons recorded in each pixel.

How bright are each of your flat frames...they need to be about half way to overexposed, or slightly brighter. Then how many did you average together.

You need to go for about 20 or more flats, or all you will do is add more noise. Which might be what we are seeing here.

So, if you havent already, take flats at 1/2-3/4 to overexposed and take like 40 of them. If the problem persists or is not improved over the previous case, then the flat wasnt to blame.

I would try this first

paul

[blinky]

I am going to give that a try over the next couple of days. I think they may not have been exposed enough so will try leaving the DSLR set to auto exposure and see how they turn out.