M42 and friends.

[JohnH]

I was really trying to be patient with this image.

26th February 2021 -
Luminance: 7.5 seconds x 714 (of 800)
27th February 2021 -
Red: 20 seconds x 101
Green: 20 seconds x 142
Blue: 20 seconds x 129
This gave me just over 49Gb of data!

Gain - 143

ZWO ASI1600 MM Pro attached to Sharpstar HNT15028 on iOptron CEM25P. Processed in Pixinsight, sharpened in Topaz Sharpen AI and scaled in Topaz Gigapixel AI.

Location Bromley (Bortle seven to eight). Moon 99 - 100% (One day I will be able to image on a night with no moon) ~ 80 degrees away.

The gain was set too high! I still managed to over expose the Trapezium area in 7.5 secs L and 20 secs RGB and ended up with a ridiculous amount of data to crunch through. I may have knocked the collimation slightly as I set up, the diffraction spikes are a bit odd in places. Thanks to Roys image I used Arcsinh Stretch.

When processing I TRIED to be more structured and had a little experience from my few previous images.

However, I am still "confused of Bromley" -

I have now seen comments to the effect that, with a CMOS camera/DSLR I should not be using Bias frames and also I should not/cannot "optimise" Dark Frames.

In this image I have used Bias but I have not optimised my dark frames. I would welcome people's views on what is correct.

All that being said, I am quite pleased with this one.

Regards,

John

[Mac]

Nice image.

I have now seen comments to the effect that, with a CMOS camera/DSLR I should not be using Bias frames

Yes and No.

A bias frame is the shortest image that your device can capture, this is actually the readout noise from your device, ALWAYS use a bias image if possible, the reason why is as follows,

Dark frames first.
Your ccd is heat sensitive,the hotter it gets, the more noise is generated in each of the wells.
The colder it is the less noise, this is one of the reasons why CCD's are often cooled to reduce the thermal noise

Now take your camera, when you take a photo in daylight, there is a huge amount of light, this will fill the wells up and you get a nicley exposed image ALWAYS A SINGLE IMAGE, NEVER STACKED, this data is WAY above the noise threshold for that CCD device and you would not notice this noise, unless you pixel peek in the really dark areas of your image, as for readout (bias noise) not a chance of seeing it.

Now step in to your night photo, there is so little light that you have to do a long exposure, the longer the exposure, the more heat the CCD is exposed to, the more noise is going to be generated, this is why Darks are taken at the same length as the exposure, this is a noise image generated by the heat of your CCD and device.

NOW.
On that image is also the readout noise from your CCD, this is independent of the Heat noise.
With the shortest time interval on your ccd, you are unable to generate heat into the CCD thus only taking the image of the read noise,
This is sometimes where you might see banding issues with the read out,

Now, we are only talking about very very small values for this data, but when you stack 10-20- 1000 images, this read noise is always in the same place.
This noise data will stack nicely as it is a process of reading the chip which produces this noise and its always in the same place.
Unlike heat (random noise),

This is why we take the bias from away from the dark as it removes the read noise, the same for the light frames.

So ideally, set up all your equipment (in the day  ),

Once you are ready to take images you should do the following..
Camera on and cooled (if possible)
Bias frames, (this is your read noise) the mistake here is people take 100 images in very quick succession well were talking about 1/8000 of a second, so its not going to take long to rattle loads of these frames off. (allow 10 seconds between image for the heat to dissipate on the CCD). Just reading the data will cause heat.

Next take your light images.
The first images should be binned, good or bad, the reason, your camera is now working and will heat up, it might take 5-10 mins for the heat distribution to stabilize.  Once the camera is working, then your heat noise will be the same across each image, you might find the top left is noisier due to it being hotter, ect,

Then Flats, same again as lights,
Then Darks,

So in a nut shell your images contain the following data.

Bias  - Read noise
Dark   Read noise + Thermal Noise
Flats   Read noise + Thermal Noise + Obstruction noise.
Light   Read noise + Thermal noise + obstruction noise + THE DATA

The only one that matters is the light,
If you take the others then you end up with a cleaner final image.
One other thing, always take the darks, flats , bias frames on the same night, even small changes in temperature can affect the images.
If your using a filter wheel, then automate alternate light and dark frames throughout the night, that way if there is a large temperature change then at least you will have
a dark to match the light (not that it really matters at our level of photography)

Mac.

[Carole]

I know nothing about CMOS cameras, so cannot help on that front.

The sky background looks balanced and the Nebulosity looks not too bad considering the short length of subs.  The Trapezium is always a difficult one in this target because of the high dynamic range, but there are processing procedures that can be utilised where you combine a very short image of the trapezium area with a longer exposure for the dimmer areas.  Or simply (as you are using short exposures) a gradual layering process where you do not stretch the Core too much.  I do not use Pixinsight so you will have to ask those that do what the process is with that.

Overall there are some nice aspects of this image, but also some things wrong. 
There is something very odd about the diffraction spikes strangely enough not with the brightest star.  Also the stars towards the corners (coma perhaps), some stars are burnt out, could be the result of less experienced processing.  Also all the halos are pink, did you use too much colour control in red perhaps?

You are definitely making some progress though.  Practice makes perfect. 

Carole

[JohnH]

Hi Mac and Carol,

Thanks.

I think the theory behind not using Bias frames is that CMOS reading is too random. As I said, I did use Bias here and will probably continue to do so (it ain't too broken so I won't try to fix it). Thanks for the explanation about heat accumulation and spread it was new to me, I will take that to heart and apply it.

I may have misunderstood the performance graphs for the ASI1600 but, as I read it, a gain of around 140 (0.956052 electrons per data number) gives low ?the lowest? read noise. before that I was using gain of 0 which I think I will go back to.

I suffer from Coma which I am trying to cure by finding the optimum distance between the reducer and the flattener. I have a dew heater on the secondary mirror and the cable is taped along the top of one of the spider vanes - I strongly suspect that to be causing problems. When lockdown is over I will go hunting for a cleaner way of securing the cable. Also in the corners the diffraction spikes do not run through the centre of the stars (like a Bhatinov mask?) so there may be:
a) curvature issues
b) collimation issues
c) completely different issues.

I probably did go too red, I was trying to get rid of a greenish cast but I would have been better off strengthening the blue.

There is more detail in the original .xisf file which was lost on conversion to .jpg - I am not sure what to do about that.

I certainly need practice but for the moment "like a dog's walking on his hind legs. It is not done well; but you are surprised to find it done at all" .

John

[The Thing]

Hi John,

All good stuff. I'm terrible with colour balancing.

FYI I stopped using BIAS/Offset frames with my ASI294MC Pro, I use darks, flats and dark flats. Bias frames produced weird and wonderful effects worthy of a psychedelic happening for me. Your CMOS may be better behaved but my reading shows that most people using modern CMOS astro cameras have stopped using bias frames. Also don't use any dark optimisation, only matched darks, as CMOS dark current appears not to be linear as it was with CCDs so you can't extrapolated values based on time.

[Roberto]

This is a very nice result John! 

You have been given very good advice here.  For my ASI2600MC, I only use Bias, Flats and Darks.  The Flats are so short that PixInsight uses the Bias to calibrate them.  It all seems to work in WBPP 2.0.  Darks are almost unnecessary with these modern cameras (and the short exposures) but they don't hurt.  I only image 150s or 300s with the ASI.   That range seems to cover all the RGB objects I can do from London.  I use 0 Gain and 50 Offset for RGB. Other people use higher gain for narrowband but I do that with a traditional CCD.

Roberto

[NoelC]

John
Very good well done!
Can't comment on all the above; but your using nice short subs to good effect.
Super.

Noel

[MarkS]

That's a great result John!

Regarding bias, the ASI1600MM has a Panasonic sensor whose bias levels go a bit strange for exposures shorter than 0.1sec, hence the advice to use flat darks (a.k.a. dark flats) instead of bias frames.  The advice is unnecessary for most CMOS cameras but you can never go wrong using flat darks instead of bias.

Since you are using a cooled camera there is no need to "optimise" Dark Frames.  Just take darks with the same exposure length, temperature, gain and offset as your light frames.

Mark

[JohnH]

Pixinsight is a really steep learning curve! I decided to give this data another go.

Here is my newest attempt, I have used the original data to fill in a bit of the overexposed core of the nebula. All of the old problems are still there but I am amazed how much additional information was lurking.



Regards,

John

(Never knowingly under processed!)

[Roberto]

WHAT AN IMPROVEMENT! All that signal was hiding in there.  Once you learn to use PxI, you'll never go back.   I think you can move the black point further to the left as there is some clipping and you'll see even more detail.

Roberto

[Carole]

Absolutely, great improvement.  I agree with Roberto's comments about the clipping. 

You are moving in the right direction. 

Carole 

[Mac]

Very nice image.
There is a lot of data there,
Mac.

[JohnH]

Thanks everyone.

At the moment the clipping is hiding other problems caused by mistakes when imaging! I will try to tinker a bit to see if I can dodge them another way.

John

[RobertM]

Lovely image John !

Yes I agree about the black point too.  Have you checked your camera sensor for tilt, if not it's worth covering that off just in case.

Robert

[JohnH]

Hi Robert,

I suspect that there might be tilt (differential focus across the field of view) but I do not know how to check for it.

John