What am I doing wrong with my flats?

[simmo39]

I have just started to take flats after getting a ASI 1600mm. I had never bothered before but it seems they are needed but im having a bit of a love hate relationship with them. Here is mt intergrated image with flats. 

AS you can see I have still got a lot of vignetting at the corners.

Here is a single sub before processing.

Although there is some vignetting its better than the above image.

Here is the flat im using.

From my limited experiance it dosent  look to bad to me. I generated it using APT flats aid. below is a single flat sub

Has anybody got any idea where im going wrong? 

Many thanks in advance.

 

[johngm]

With my limited knowledge to, those look rather dark for flats. Make sure your histogram is in the centre when taking them, and do not alter anything in the imaging train.

John

[simmo39]

27 minutes ago, johngm said:

With my limited knowledge to, those look rather dark for flats. Make sure your histogram is in the centre when taking them, and do not alter anything in the imaging train.

John

I was wondering about that but that is what APT flats aid throws out for use. They are 20000 adu thingy which is about what flats should be isnt it?

[johngm]

I'm not sure about ADU measurements or APT. I use the histogram and BYEOS. I am sure APT shows the histogram. Might be worth trying. I will try and Dig out one of my flats.

John

[alacant]

Hi. I stacked your jpgs in Siril and stretched. The result was fine to my inexperienced eye. How are you stacking the flat and light frames?

On Siril I use median stacking with multiply normalisation. 

[johngm]

A Single flat of mine. The blue hue is due to my CLS Clip filter. This was taken with a Modified Canon 600D.

 

I hope this is of any help.

 

John

[spillage]

 

I have just posted a similar thread about the asi1600 and flats/calibration frames.  I have seen over on CN that some use down to 13k but I could not see that the gain setting had been mentioned. I to was sure that 20k should be about right, I think my issue is the panel being too bright as my exposure is .00247 with the L filter.

 

Can you post an photo of your image train and the kit used in it.

[geordie85]

It appears that your flats are over correcting whihh means your ADU is too high. Try lowering the ADU and stacking again. 

Its all trial and error till you find what works best for you. 

With my Atik 383L I'd use flats with an ADU of 23000, but with my qhy183c I need flats of around 18000 to 20000. Anything more than that and they over correct 

[vlaiv]

Did you take darks at all?

Glow on right side looks like dark current glow - meaning no regular darks.

Over correcting flats usually mean - no flat darks were used.

How do you calibrate your frames?

 

[simmo39]

1 hour ago, vlaiv said:

Did you take darks at all?

Glow on right side looks like dark current glow - meaning no regular darks.

Over correcting flats usually mean - no flat darks were used.

How do you calibrate your frames?

 

Hi, i have been using Flat darks aswell as flats and darks. I have been following the tutorial on the PI sight.

[simmo39]

2 hours ago, geordie85 said:

It appears that your flats are over correcting whihh means your ADU is too high. Try lowering the ADU and stacking again. 

Its all trial and error till you find what works best for you. 

With my Atik 383L I'd use flats with an ADU of 23000, but with my qhy183c I need flats of around 18000 to 20000. Anything more than that and they over correct 

Hi Geordie, I will try 18k tonight to see if that helps. thank you.

[simmo39]

2 hours ago, johngm said:

A Single flat of mine. The blue hue is due to my CLS Clip filter. This was taken with a Modified Canon 600D.

 

I hope this is of any help.

 

John

Thanks john, not much difference to mine.

[simmo39]

2 hours ago, alacant said:

Hi. I stacked your jpgs in Siril and stretched. The result was fine to my inexperienced eye. How are you stacking the flat and light frames?

On Siril I use median stacking with multiply normalisation. 

Hi, Im using very similar settings to you. as for the Flats and darks im following the PI tutorial.

[ollypenrice]

It's important to understand that stretched flats contain very little real information. Depending on the stretch they can show huge vignetting and terrible dust bunnies or hardly any vignetting and the mildest of bunnies, but they are the same flats. What we need to know is the range in ADU from the dark parts to the bright.

I agree with those who think the flats have over corrected, possibly due to an issue with dark flats. Unfortunately my own experience of this problem proved to be very intractable and I never did get to the bottom of it. Does your capture software have alternative capture modes? It is vital that the capture mode for the lights be the same as that for the flats.

I also wonder about linearity. Will the short flats subs have the same level of linearity as the long light subs? They do, more or less, in CCD but what about CMOS?

Olly

[vlaiv]

 

14 minutes ago, ollypenrice said:

I also wonder about linearity. Will the short flats subs have the same level of linearity as the long light subs? They do, more or less, in CCD but what about CMOS?

 

According to multiple sources ASI1600 is pretty much linear:

https://www.cloudynights.com/topic/554803-more-asi1600mm-cool-statistics-linearity-and-more-zwo-settings/

http://www.astrosurf.com/buil/CMOSvsCCD/index.html

[simmo39]

10 hours ago, ollypenrice said:

It's important to understand that stretched flats contain very little real information. Depending on the stretch they can show huge vignetting and terrible dust bunnies or hardly any vignetting and the mildest of bunnies, but they are the same flats. What we need to know is the range in ADU from the dark parts to the bright.

I agree with those who think the flats have over corrected, possibly due to an issue with dark flats. Unfortunately my own experience of this problem proved to be very intractable and I never did get to the bottom of it. Does your capture software have alternative capture modes? It is vital that the capture mode for the lights be the same as that for the flats.

I also wonder about linearity. Will the short flats subs have the same level of linearity as the long light subs? They do, more or less, in CCD but what about CMOS?

Olly

Hi Olly, Thanks for the reply. I have to say I dont know but tonite Im going to try a lower ADU to see what happens.

[rockinrome]

23 hours ago, johngm said:

A Single flat of mine. The blue hue is due to my CLS Clip filter. This was taken with a Modified Canon 600D.

 

I hope this is of any help.

 

John

This is the kind of thing I get with 1200D unmodded with CLS filter too.

Regards

 

[carastro]

I was teaching some-one to image with a DSLR using APT and the Flats plan did not produce usable flats.  Him being new at imaging did not really know what was right in the way of a flat, but I could see they were not doing their job.  

In the end we just manually worked out how to do the flats by taking short shots until the histogram was about 1/3 in from the left.  

Carole 

[simmo39]

4 hours ago, carastro said:

I was teaching some-one to image with a DSLR using APT and the Flats plan did not produce usable flats.  Him being new at imaging did not really know what was right in the way of a flat, but I could see they were not doing their job.  

In the end we just manually worked out how to do the flats by taking short shots until the histogram was about 1/3 in from the left.  

Carole 

Hi Carole, After trying again last night at 18k adu. there was a slight improvement but still not right. I think i will have to learn how to do it properly.

[carastro]

Quote

After trying again last night at 18k adu. there was a slight improvement but still not right. I think i will have to learn how to do it properly.

I am trying to recall how we did it as it was a while ago, but I think it was the option just to take an image without using a "plan", and we called up the histogram so we could see what each sub produced.  I recall the exposure getting shorter and shorter (fractions of a sec I think) until we got the histogram in the right place.

Carole