No correlation between the master dark and the target frame

[star gazer]

I'm new to PI, but have already had some success.

Today though I am getting an message that I don't remember on my previous runs.

The Dark scaling factors are set to 0 and I get a warning 'No correlation between the master dark and the target frame'. Does this mean that i should set the scaling so that the frames correlate? Not sure how?

The reason for asking is that I haven't managed to get a good result in terms of dark frame subtraction so far.

I think I need a little help.

I'm using a DSLR with the darks and light with the same duration and iso, and roughly the same temperature.

Any advice?

[sreilly]

I have no experience with the DSLRs but you would need to have a master bias so scaling can be done. You are making your master darks and bias frames in PI?

Steve

[star gazer]

I'm not using any bias frames currently.

I am making the Master Dark frames in PI

If I do image calibration using a Master Dark and the inspect the result after applying the Sreen Transfer Function (auto), then this is the result:

http://dl.dropbox.com/u/16060721/Astro%20Help/calibration%20noise.png

Would you say my calibration result is poor? I think it is but what can be done to improve this? I can only think my master dark is poor.

What about the horrible lines? I think this is typical canon sensor banding. May be I can deal with this later?

[Nocturnal]

What does your stack look like if you do not subtract darks? Do you have enough dark frames to make it low noise? How long is the total exposure length of your lights?

Looks like it's simply a short exposure which will generally be noisy. Dark subtraction does not remove noise. It adds noise. It removes dark signal though, but that's not the same. In some cases dark subtraction does not improve image quality.

[star gazer]

Ok this is what I am working with

10 x 5 sec
10 x 15 sec
10 x 30 sec
10 x 60 sec
10 x 120 sec
10 x 150 sec

The link in my previous post is to the calibrated  10 x 5 group of subs.

I plan to merge all these using HDR function in PI.

[Nocturnal]

I think that 6 series of exposures if complete overkill for an HDR shot but hey, whatever floats your boat

How does your calibrated 10x60 look? Are you calibrating each exposure length with appropriate darks?

[star gazer]

It's a stab in the dark to have 6 sets go exposures. I know that M42 is very bright in the core and very faint at the edge. I hope that I have all the differences in dynamic range covered. Why is it overkill? What would you have done?

[Nocturnal]

I have taken good images of M42 with 2 sets of lights. With a DSLR you may need 3 because of reduced dynamic range. If this is your first M42 image I would start with 1 set of exposures and then do 2 exposures for your next. Of course there's a lot of 'it depends' and 'your mileage may vary'. Since you're only signing as 'star gazer' I don't know who you are and how experienced you are at this, sorry if my suggestions are inappropriate. Without context about someones experience and equipment it's hard sometimes to give the right advise.

So tell us, how long have you been doing this and do you have some past images to share?

[star gazer]

Hi Sandar,

You are right I should have introduced myself.

I am very new to imaging, so have next to no experience. I have a good knowledge of day light photography, with some but not in depth knowledge of PS. My knowledge of PI is only what I have gleaned from Harry's tutorial videos.

I am imaging with a WO Megrez 72 TV field flattener and my un-modded Canon 7D. Also I am running without any guiding. I'm imaging in rural Wiltshire (UK) where there is some LP, but I think not so bad. I can see a reasonable view of the milky way on a clear night. The nearest town is 5 miles away I am also using a Hutech IDAS LP filter.

I haven't tried stacking anything other than the 5 second group, but I will try your suggestion with the 6o seconds tomorrow.

What is troubling me at the moment is that there is a lot of; for want of a better way of putting it, gridded noise. Please take a look at the link in my second post in this thread how can I remove this? I think this will be in the 60 subs so I will need to banish it some how.

Thanks for your advice.

Tim

[pfile]

the image you posted looks pretty normal for a short exposure of the sky with any canon DSLR. you'll notice that the banding noise is much reduced in the brighter areas of M42.

how long is that exposure?

if you hope to pick up any of the dust around M42 you will have to expose long enough so that the dust is out of the read noise of the camera.

generally speaking, "no correlation between master dark and target frame" usually occurs when you are dealing with short subs and there is just not much dark signal there, either in the light or the dark. you'll probably find that you're getting this message on your <60s subs.

for something like M42, i don't even bother to calibrate the frames that are used to recover the trapezium... i just stack 5 or 6 of them and merge them in with HDRComposition.

[astropixel]

It's a good idea to dither. Dithering will remove a large percentage of banding and ambient noise associated with some DSLR's. Bias frames are essential with short exposures as there is not much dark/thermal noise.

[star gazer]

Thanks for your help pfile.

how long is that exposure?

It's just 5 secs. I realise now that is probably too short. But as you can see from my earlier post I do have longer duration subs

10 x 15 sec
10 x 30 sec
10 x 60 sec
10 x 120 sec
10 x 150 sec

I will process these over the weekend

[star gazer]

Thanks astopixel

It's a good idea to dither. Dithering will remove a large percentage of banding and ambient noise associated with some DSLR's. Bias frames are essential with short exposures as there is not much dark/thermal noise.

I'm not sure how you dither and at what point in the workflow. For instance is dithering applied to the light or dark frame, or maybe during calibration? Would you mind giving me a bit more explanation?

Thanks again

Tim

[Cleon_Wells]

Tim, on the Q of dither, check this link out.
http://www.adass.org/adass/proceedings/adass99/O6-02/
 That said, I would start sort of where Sanders said, closer to the basics, don't Calibrate, just convert your 14bit CR2 files to .fit files with the BatchFormat Conversion script . This procedure seems simple but it's full of pit falls.
First go to the Format Explorer tab, check the attached image for the settings.
Second click BatchFormatConversion under Scripts/batch processing changing the default settings see attached image.
Image Integrate some of your longer image sets, this is a starting point.
Next you can make a MasterBias form around 50 iso100_1/4000 subs with the Camera covered with aluminum foil to keep out the light, convert with BatchFormatConversion
read this,
http://pixinsight.com/forum/index.php?topic=2570.msg19019#msg19019
work with the MasterBias first, try Calibrating some of your lights say 6 of the longer sets, see how the dark back ground level changes between unCal and Calibrated lights.
Then try to make a MasterDark work, and then a MasterFlat...
Cleon

[star gazer]

Hi Cleon,

Thanks for the detailed explanation. I will be taking yours and others advice for sure.

One interesting thing I have noticed is your defaults for the FIT conversion do not work well for me. If I view the fit using your defaults, as is, or after debayer. the result is very green. My defaults ( for my 7D) are to uncheck everything and then go for bilinear interpolation.

Thanks again