19 replies to this topic

[RoscoeD]

(If this is not the right forum, please direct me to the correct one)

 

Currently buying equipment and researching my brains out in preparation for spring when it will be warm enough to spend hours outside making mistakes 😎

 

I have a Nikon D750 (not modified nor do I plan to do so as I primarily use it for terrestrial work [I assume Nikon put an IR filter in there on purpose so I'd rather not remove it...]) and I have a ZWO ASI294MC Pro (color) camera being shipped as I type.  Note that I'll probably start with the DSLR as it will be a bit easier to learn on (T mount also on order).

 

As I research, I'm reading about flats, lights, darks, bias's etc... my head is spinning.  I find posts that talk about pieces of the process but I can't get my head wrapped around the entire process, begging to end as it were.  Can someone point me to a post or any good website that explains the complete process?  I'd kinda like to have the whole picture (so to speak) before I start diving in (the engineer in me).

 

Thanks!

[rj144]

Calibration is explained pretty well here:

 

http://deepskystacke...tdarkflatoffset

[acrh2]

(If this is not the right forum, please direct me to the correct one)

 

Currently buying equipment and researching my brains out in preparation for spring when it will be warm enough to spend hours outside making mistakes

 

I have a Nikon D750 (not modified nor do I plan to do so as I primarily use it for terrestrial work [I assume Nikon put an IR filter in there on purpose so I'd rather not remove it...]) and I have a ZWO ASI294MC Pro (color) camera being shipped as I type.  Note that I'll probably start with the DSLR as it will be a bit easier to learn on (T mount also on order).

 

As I research, I'm reading about flats, lights, darks, bias's etc... my head is spinning.  I find posts that talk about pieces of the process but I can't get my head wrapped around the entire process, begging to end as it were.  Can someone point me to a post or any good website that explains the complete process?  I'd kinda like to have the whole picture (so to speak) before I start diving in (the engineer in me).

 

Thanks!

The gist of calibration is expressed in this simple equation from ASTAP:

 

DSO signal = 

I'll rewrite it like this:

 

DSO signal = (image-dark)/(flat - flatdark).

 

So take a light image, then correct it for the dark signal by subtracting the dark. Correct that for optics imperfections by dividing by the dark-corrected flat.

So, here's what you need to do, once you figure out exactly how to collect those darks, flats, and flatdarks, as described in the previous point.

 

1) At home, prior to imaging the sky, create a set of darks.

2) Collect lights.

3) After collecting lights, get a flat panel, put it on top of the telescope and collect flats.

4) Then cover the telescope and collect flatdarks.

5) Load all 4 of the data sets into DSS and let it do its magic.

 

P.S. I've never used a DSLR but I cannot even begin to imagine how using a DSLR could possibly be easier than using a dedicated astro camera.

Your 294 should come with all of the adapters and cables necessary. Then you just connect it to the telescope, plug in a usb cable into a laptop and start recording with Sharpcap or NINA or some other software.

Edited by acrh2, 15 January 2022 - 02:11 AM.

[ks__observer]

Glad you posted formula but one tweak:

 

"Calibrated DSO signal" = (image-dark)/(flat - flatdark).

 

Bias can substitute for flatdark.

 

Darks subtract fixed pattern noise.

[acrh2]

Glad you posted formula but one tweak:

 

"Calibrated DSO signal" = (image-dark)/(flat - flatdark).

 

Bias can substitute for flatdark.

 

Darks subtract fixed pattern noise.

Technically true.

 

Flatdark = Bias + darkcurrent.

 

So if Bias >> darkcurrent, then you can substitute Flatdarks with Biases.

[ks__observer]

Flatdarks = bias + FPN Signal

"Dark current" is an additional source of noise.

"Dark current" is separate and distinct from FPN image inherent even in a dark frame.

[acrh2]

Flatdarks = bias + FPN Signal

"Dark current" is an additional source of noise.

"Dark current" is separate and distinct from FPN image inherent even in a dark frame.

Why are you talking about noise?

You are complicating things unnecessarily.

 

Btw, the equation I posted, Flatdark = Bias + darkcurrent is the definition of bias frames.

[ks__observer]

People think dark frames subtract "dark current."
Darks subtract FPN.
"Dark current" is a temperature dependent noise source.
Stacking darks adds dark current noise as the cost for removing FPN.

Edited by ks__observer, 15 January 2022 - 07:43 AM.

[acrh2]

People think dark frames subtract "dark current."
Darks subtract FPN.
"Dark current" is a temperature dependent noise source.
Stacking darks adds dark current noise aa cost for removing FPN.

I am sorry. That is incorrect.

 

I don't know what "people think," but dark frames subtract both bias and dark current.

By definition, dark frames contain bias signal and dark current signal, i.e. Flatdark (or Dark) = Bias + darkcurrent, where Bias is a signal recorded with the shortest exposure, and darkcurrent is a signal recorded after that.

 

Dark frames don't subtract FPN. FPN can be calibrated out of lights by flat frames:

https://en.wikipedia...pression_of_FPN

 

*Using* dark frame calibration adds noise, true, but *stacking" darks reduces that noise.

 

When you say FPN, are you referring to Amp Glow by any chance?

 

In any case, using bias frames in place is flatdarks for a modern CMOS sensor is lazy, imo, and is asking for trouble. But it depends on the particular sensor.

 

NINA has a Flat Wizard module, which allows you to take flatdarks immediately after taking flats with the proper exposure, and it literally takes less than 5 minutes for the whole thing.

[danny1976]

You can do the pre-processing manually or you can automate it with scripts. 

 

Scripts in Siril are very easy. You only make 4 folders lights/darks/biases/flats and Siril does the whole process of calibrating, registering ... and you end up with the final stacked image ready to process.

Edited by danny1976, 15 January 2022 - 08:40 AM.

[matt_astro_tx]

Yep.  That's what I do.

 

Also, astrobackyard has a good primer on calibration frames on his web site. 

[ks__observer]

Darks add dark current shot noise to calibrated image.

Wiki is usually correct but:

Darks subtracts both FPN and subtracts amp glow.

Edited by ks__observer, 15 January 2022 - 08:16 AM.

[rishigarrod]

You might want to skip the DSLR step and go straight to your 294mc Pro. They both have a learning curve and in many ways I find a dedicated camera simpler. Most people start with a DSLR because they have one. With the DSLR you have no temperature control so you have to take your darks after every session (when you are tired and really just want to go to bed). With the 294mc you can build your entire dark library inside. Do it in a cool spot at home. If it is 20 degrees C and you want to cool to -10 C then the camera can struggle.

 

A couple of things on calibration frames for the 294mc. Darks are critical. They need to be same exposure, gain, offset and temperature (as for all cameras). The amp glow is significant with the 294MC but good darks remove it completely.

Make sure your flats have an exposure time of at least 2 seconds for the 294mc (someone far smarter than me probably knows why).

[danny1976]

I agree about skipping the dslr. With a cooled camera you can take darks whenever you want and reuse them. If you have 300s lights then your darks also have to be 300s. For 10 darks that means you can end your session 50min earlier. It is better to take more darks.

 

A dslr has raw files that you have to convert to fits files in Siril. Another extra step. Your cooled camera directly outputs fits files.

 

And another tip, don’t use biases but instead take darkflats. The process is the same. In Siril just replace the biases by darkflats (but don’t rename the biases folder otherwise the script wont work). Darkflats are with the same settings as flats but with lenscap on.

Edited by danny1976, 15 January 2022 - 09:03 AM.

[jonnybravo0311]

Roscoe,

 

For astrophotography, you take four types of images. These are referred to as lights, darks, flats and flat darks (with your 294, you do _not_ use biases). I'll explain.

 

Lights: the stars of the show (pun intended). These are the pictures of the stars/nebulae/whatever

Darks: used to calibrate out the fixed pattern noise (hot pixels, amp glow) and bias signal of your Lights. See discussion above between ks_observer and acrh2

Flats: used to calibrate out imperfections in your optics like vignetting and dust motes

Flat Darks: used to calibrate out the fixed pattern noise (hot pixels, amp glow) and bias signal of your Flats.

 

The 294 series cameras are notoriously finicky. With a lot of other cameras, you can get away with things like using bias frames in lieu of flat darks and just not taking any darks at all (camera like the ASI2600 is an example).

 

So, why should you bother taking these calibration frames? Well, a picture is worth 1000 words, so below I present the same data. The top set is stacked with absolutely NO calibration frames. The bottom set is stacked with all the calibration frames. I'll let you judge which you think is the better data set .

 

Not Calibrated

 

 

 

Calibrated

 

[RoscoeD]

Lot's of good discussion here, greatly appreciated!!

 

This site was also good (at least to this newb) of filling in some of the gaps (such has how to take a dark flat).

[ks__observer]

This site was also good (at least to this newb) of filling in some of the gaps (such has how to take a dark flat).

Perfect example of so-called experts getting it wrong -- or at least vastly oversimplifying that dark frames remove "thermal noise."
Dark frames do not calibrate out dark current temperature shot noise.
Darks calibrate out DSNU -- dark sensor non-uniformity -- one type of sensor FPN.
https://www.cloudyni...ight/?p=9891860

Edited by ks__observer, 15 January 2022 - 02:41 PM.

[unimatrix0]

 

I have a Nikon D750 (not modified nor do I plan to do so as I primarily use it for terrestrial work [I assume Nikon put an IR filter in there on purpose so I'd rather not remove it...]) and I have a ZWO ASI294MC Pro (color) camera being shipped as I type. 

 

The IR filter is in there, because regular daytime photography of landscapes and cats and dogs do not need to capture any faint light anything near infrared, including hydrogen alpha emission. 

DSLRs, mirrorless are designed to work 99.99999% on broad daylight and on planet Earth and people want to see the pictures of colors and lights that can be seen by human eyes. 

If you plan to capture anything other than broadband emitting lights (galaxies, star clusters, reflection nebulae) you still can with unmodified, but your camera is limiting that light. 

It's sort of trying to navigate inside a dark house at night in sunglasses.  You can maybe still see a bit if there is some light available, but surely see a lot better without the sunglasses. 

 

At night, things changes, since we are chasing faint light emissions that are - most of it- invisible to see to human eyes.  Cameras are better than human eyes and they are able to record nearly the entire spectrum of light. 

At the night sky, all emission nebule emits a specific wavelength of light  the most, it's ionized hydrogen and the filter that Nikon put in there is blocking it, or let's say it blocks most of it.  That's why many people remove that filter, trying to save some minutes or hours of imaging.  But then again, there are those who are fine with it, and want to keep their camera stock.  For this purpose- I have an unmodified camera I use for daily stuff and broadband (Pentax) and I also bought a used DSLR with the purpose of modifying it and specifically used for astrophotography only. 

Edited by unimatrix0, 16 January 2022 - 12:14 AM.

[RoscoeD]

The IR filter is in there, because regular daytime photography of landscapes and cats and dogs do not need to capture any faint light anything near infrared, including hydrogen alpha emission. 

trust I'm aware of all of that.  In my recently-retired job I was intimately familiar with IR systems.  In the case of DSLR we know the sensor is capable of detecting light at wavelengths that the human eye cannot.  We also know that those wavelengths for us at different locations than visible light...letting that light through can cause a slight out-of-focus fuzziness int eh final image.  Hence DSLR makers filter that light out.  I'd prefer leaving it that way. 

[TelescopeGreg]

The issue isn't so much what we can see, it's that much of the light coming in from deep space is that of hydrogen alpha at 656nm.  The filters in an unmodified DSLR kill most of this light, making it very difficult to image much of the stuff out there.