vlaiv

1 hour ago, dannybgoode said:

Thanks again @vlaiv  

Would I be right in thinking then that binning can be useful to prevent oversampling (or undersampling - I can’t quite get my head around which!) with long focal length scopes?

 So for example using my SX674 mono with my TMB 105/650 gives me a result of 1.43” pp so about right. 
 

However if I wanted to use my Meade 10” SCT with an FL 2500 then binning 2x gives me 0.74 or using my f6.3 flattener/reducer 1.18 so just about spot on?

Yes indeed - one way to look at binning is as increased pixel size, so you are right, it is a way to combat oversampling (low "/px number means high resolution and oversampling - it is inverse in pretty much same way wavelength and frequency are inverse of one another)

2 minutes ago, dannybgoode said:

Thanks @vlaiv - that’s the chestnut I was looking for. Am I right in thinking that it is pixel size in um (apologies on the phone so can’t get the proper character!). 
 

So if my pixel size is 4.5 um then I just use 4.5?

Yes - if you have 4.5um pixel size and for example 500mm FL - resulting sampling rate is in arc seconds per pixel - 1.86"/px

Btw - check out top of this page - menu on SGL - there is Resources menu, and first item there is astronomy tools - it will give you some basic tools for calculations / fov, etc ... (although I don't agree with some of the things there - like CCD suitability)

206.3 * pixel size / focal length

(pixel size in um and focal length in mm)

If you want to be exact - that is for mono cameras, for color cameras it is twice that value

6 minutes ago, Lockie said:

Hi vlaiv. Thanks for your well considered reply, very sensible. Well, being brutally honest I always end up wanting to take quick pics of things if the past is anything to go by.

E.g. If I'm looking at the Moon and planets, it's just a few minutes observing before I want to grab some AVI footage with a webcam. If the Moon and planets aren't about I like to try and find a few DSO's to observe, but again I do like to grab some images either by piggy backing a camera and lens for longer exposures at shorter focal length or go a bit deeper with some unguided 1 to 2 minutes max exposures with DLSR or mirrorless camera attached to a scope. 

I just need an OTA and mount with some kind of RA tracking. I have a modded webcam, and a couple of mirrorless cameras (all in my sig). I have some bit's and bobs such as Plossl EP's, collimation tools, focus mask, a decent diagonal, and an RDF. I would probably need a new T ring for my camera, and would probably opt for the Fuji with it's inherently weak IR filter as if should work reasonably well on nebs as a stock camera body.

Something like the EQ5 pro and 130pds would work well. I see the goto as a bonus really. Only really expecting some kind of tracking for my budget

I guess in summery imaging is more important to me, but I'm not fussy about long exposure guiding, goto, PC's being involved and all that until I get an obsy back further down the line   

 

I still feel that I don't fully understand what ticks the boxes for you. You say it is imaging, and I can understand that, but what level of imaging?

Maybe it would help you more if I told you "my story"?

This is something that I would not recommend you yet, but it is something that is pretty much in line with your original question. I'm awaiting (impatiently - there was postal workers strike that ended just days ago) for a shipment (just cleared customs) of setup that is supposed to do it all - and guess what? It costs about the same as your budget is.

It's Mak102 + AZGti (and a few more bits and pieces - that is why I hit your budget mark). I intend it for following purposes:

- grab&go / lunar peek scope to be used on terrace

- testing out "do it all" setup capable of: planetary imaging (with suitable CMOS camera), planetary observing, DSO imaging both in traditional role, but also with small sensors and reducers and in EEVA role - with eyepiece projection setup.

I'm mentioning this because difference between occasional snap and quite usable EEVA (and fairly decent planetary imaging) and a bit more serious imaging is - mount. AzGTI has been mentioned before in this thread and yes - it is very interesting solution but I doubt it can be used for serious AP. On the other hand I think it will do very well in EEVA role - even in AZ mode without the need for wedge / counterweights to do EQ configuration.

Maybe your need for DSO imaging can be met by EEVA approach instead of traditional AP?

4 minutes ago, Rob Sellent said:

As such, all options should be explored when observing, or at least those which maximise the observer's experience

This is much in line with what I've gained from this thread. Up until now I was under impression that x2 aperture in mm is based on theory (and it is but with a bit flawed assumptions) and that was one of my primary drivers in eyepiece selection - you need one eyepiece that will show what scope can deliver under optimum viewing conditions.

It turns out that chasing short FL comfortable eyepieces is not going to provide that because x2 aperture is not that magical figure. It sort of lifted that burden of aiming for particular magnification based on physics of things and set me free to experiment with both lower and higher magnifications depending on target (primarily without fear that I'll miss out on some finest detail if I use lower power eyepiece that is sharper due to less mag and also due to atmosphere).

Who does not want it all for £600-700?

It can be done, I think, but you need to list your priorities. Can you make a list of priorities from high to low on these (if I got that right):

- observing planetary

- observing DSO

- imaging planetary

- imaging DSO

- solar (this is probably at the bottom of the list).

Does your budget include camera? (I really hope not) and what sort of camera did you have in mind (or one that you have, or have special budget for ...).

I'll give my "initial" recommendation of setup that would cover above with close to equal priorities:

Sky-Watcher EQ5 PRO Go-To - £569

Sky-Watcher Explorer 130P-DS - £179

Grand total of ~£750

Does not include accessories that you might need but also might have - EPs, barlows, as mentioned camera, finders, etc ...

You can save quite a bit of money by going with EQ3 Goto version and purchase all accessories (except camera) that you might need, but then again - how important is imaging for you (this is option if DSO imaging is lower scoring item on the list).

Ok, in this particular case - flat file is wrong and such flat file can certainly lead to above artifacts.

Here are my recommendations:

- shoot flat darks as well - maybe you already doing that, but you did not include one flat dark in this set so I figured I should mention it anyway - these are shot in the same way you shoot regular darks - except matched to flats (exposure, gain, offset of flats).

- since you are using OSC camera - pay attention to histogram when shooting flats. Your current flat file has been over exposed and there is clipping to the right on histogram. OSC flat histogram is going to have three peaks instead of one. Common wisdom when shooting flats with mono camera is to have peak at 75-85% of histogram. When doing OSC, you want your right most peak to be in this position.

Here is histogram of your flat:

and this is how you want your flat histogram to look like:

(actually this one is a bit under exposed - you want your right most peak to be not at the middle but about 3/4 to the right).

5 minutes ago, John said:

I'm not sure about it either come to think about it. It's probably a bit beyond the scope of the question asked by the original poster anyway

Yes - it occurred to me that I should have checked if initial post was in beginner section before I started whole theoretical discussion.

6 minutes ago, John said:

I think often we see the Encke Minima rather than the Encke Division.

 

Not sure about that. Minima is very low contrast feature and is quite extended and soft:

Image itself is not telling much.

Can you post single unaltered fits from camera of following:

one light sub

one dark sub

one flat sub

one flat dark sub

of an image that displays this issue?

3 hours ago, jetstream said:

@vlaiv is it possible that resolving stars splits differ from resolving objects such as planets with regard to the MTF for the latter? I find Peachs obervations to be true from my own observing.

from Peach

 

"Understanding Resolution and Contrast

Two points it is important to understand is the resolution a telescope can provide, and how the contrast of the objects we are imaging affects is related to what can be recorded. Its often seen quoted in the Dawes or Rayleigh criterion for a given aperture. Dawes criterion refers to the separation of double stars of equal brightness in unobstructed apertures. The value can given given by the following simple formula:

115/Aperture (mm.) For example, a 254mm aperture telescope has a dawes limit of 0.45" arc seconds. The dawes limit is really of little use the Planetary observer, as it applies to stellar images. Planetary detail behaves quite differently, and the resolution that can be achieved is directly related to the contrast of the objects we are looking at. A great example that can be used from modern images is Saturn's very fine Encke division in ring A. The narrow gap has an actual width of just 325km - which converts to an apparent angular width at the ring ansae of just 0.05" arc seconds - well below the Dawes criterion of even at 50cm telescope. In `fact, the division can be recorded in a 20cm telescope under excellent seeing, exceeding the Dawes limit by a factor of 11 times!. How is this possible?.

As mentioned above, contrast of the features we are looking at is critical to how fine the detail is that we can record. The Planets are extended objects, and the Dawes or Rayleigh criterion does not apply here as these limits refers to point sources of equal brightness on a black background. In fact it is possible for the limit to be exceeded anywhere up to around ten times on the Moon and Planets depending on the contrast of the detail being observed/imaged."

http://www.damianpeach.com/simulation.htm

I think we are talking about different things here - resolving and detecting.

Here is counter argument to what is written above that might explain what I mean by that - what is angular diameter of a star? Much less than 50mas angular width of Encke division. Looking at the list of resolved stars - about 1mas. But there is much more unresolved stars that have smaller angular diameters still - and we see them all.

In that same sense we see Encke division.

That is detecting a feature and yes it does depend on contrast of the feature (as does resolving).

Now let's discuss resolving a feature - imagine Encke division is not a single clean gap, but something more like this:

(actual image of Encke gap)

How separated two gaps need to be before we can detect that they are separate features? What is their distance at which we can resolve them and say - look it's not single gap, there are two gaps with ridge between them (or there are two stars rather than one in that binary system).

That is the meaning of resolve vs detect.

Yep, if focuser does not move - loosen screw underneath it - it could be focuser lock screw.

I don't remember having one on my 8" Dob, but it looks like focuser lock screw - and if it is tightened - focuser won't move when you rotate knobs.

24 minutes ago, Richard136 said:

The TS scopes are of course an option. Is it worth the nearly 200 euro extra for the 72mm? Glass is stated as better and the mounting looks more substantial too. Though at that price I can have a WO ZS73

TS Photoline 72mm is the same class instrument as WO ZS73 - both in aperture but also in optical quality (probably, I can't be 100% certain. I do own 80mm TS F/6 APO triplet and it is indeed very good optics). Same class focuser - 2.5" R&P.

There are few things that may sway you to TS side - like lower mass (2.2Kg vs 2.8Kg - probably due to carry handle, tube length and longer dovetail), shorter tube in TS model (315mm vs 390mm) and longer back focus.

I'm just guessing longer back focus - since tube is shorter in TS model and both are F/6 scopes with same focal length of ~430mm. TS website states 135mm focus position distance from end of 2" receptacle and 73mm focuser travel - that is going to be usable on both 2" and 1.25" diagonals "out of the box" - most 2" diagonals have around 110-115mm optical path while 1.25" have about 75-85 mm optical path (only prisms have significantly shorter path).

WO has couple of things going for it - local dealer which is much better in terms of transportation and any issues that you might have with scope, and stronger brand name which I suspect will be a plus in resale value.

Btw - both scopes I listed being TS models are available from Altair Astro:

Cheaper model - 70mm and 2" focuser:

https://www.altairastro.com/starwave-70ed-f6-travel-refractor-telescope-with-2-crayford-focuser-finder-diagonal-eyepiece.html

Better model - 72mm model with 2.5" focuser:

https://www.altairastro.com/Altair-72-EDF-Refractor-Deluxe-CNC.html

(you have "CNC - deluxe" model as well as a choice - £50 more but you get better focuser with rotator and optics test report).

If astro photography is not important for this scope and you are going to use it for grab&go, mega finder and travel scope - I would personally consider cheaper 70mm scope. Same aperture, could have some color on high magnifications - but I would not expect any serious observing on 70mm so that would not be too big of an issue. It will have edge over 61mm as travel scope - not much but I would not mind having 10mm more aperture in roughly the same "package" in terms of size and weight.

If you are still keen on WO61 have a look at these two scopes:

https://www.teleskop-express.de/shop/product_info.php/info/p10095_TS-Optics-PhotoLine-60mm-f-6-FPL53-Apo---2--R-P-Focuser---RED-Line.html

and

https://www.altairastro.com/60EDF-ED-R-Refractor-Telescope.html

and keep in mind that most of these scopes are made in PRC or Taiwan and just branded differently

15 minutes ago, lenscap said:

Vlaiv, I have skimmed through the "Magnification" section of this  source

and it seems  that the 2D figure is reached by assuming 4 arc min as a resolution limit for the eye, not the 1 arc min in your first analysis, which accounts for the difference between the 2D & D/2 results.

So I suppose the question is, what is a practical, realistic resolution limit for a typical, dark-adapted eye observing a double star through an appropriate telescope exit pupil.

 

Well worth a read for all interested.

2 minutes ago, Merlin66 said:

The resolution of the human eye is usually stated as 2 arc min. Fantastic eyesight (few and far between) can get down to around 1.5 arc min.

The Dawes Limit for visual observing (generally double stars) depends on the ability to discern the low contrast peaks of the star PSF maxima.

A good test is to determine the minimum magnification to resolve double stars....

I managed to resolve the Double double in Lyra at x49 (Genesis, 10.5mm TV plossl). The separation is 2.3/ 2.6 arc sec , giving 1.88 arc min for the eye resolution.

 

 

If you read above quoted link you will see that often quoted angular resolution of human eye of 1 arc minute (or 2 arc minutes for line pair) is not representative of human eye resolution at telescope eye piece.

In fact 1 arc minute resolution is limit made by pupil size at daytime - using same 1.22 * lambda / D formula - for example 200mm telescope has Rayleigh criteria of ~0.64" and 2mm pupil in daytime will have x100 larger airy disk - so resolution is then 64" or 1' 4" - or about 1 arc minute.

With eye pupil being larger (dimmer light) - eye aberrations start messing around and blurring the image.

When at eyepiece - telescope corrects for eye aberration in two ways - first eye pupil is no longer aperture stop and it does not cause blur. Diopter is corrected by focusing and any aberrations due to shape of eye ball are reduced by smaller exit pupil. Resolution is dictated then by physical "sampling" - spacing of cells on retina - which are about 2um and for human eye and with 22mm of focal length we have about 18"/px (or third of arc minute "per pixel" so real resolution is a bit less than 1' if you need 2-3 "pixels" to sample point source).

One thing is clear - Maximum useful magnification is misnomer and should read something "Minimum resolving magnification". Higher magnifications will still show you same level of detail and at some point due to dimming start messing with contrast, which can again lead to detail loss since human eye has "static" contrast ratio of 100:1 and can only see certain number of shades - so decreasing brightness will lead to some shades that we can distinguish become indistinguishable. I guess that would be hard to quantify exactly.

I just found another interesting text on human eye resolution - so I'll just paste a link to it with citing few important bits and then I'll stop spamming this topic further

https://www.quora.com/At-what-screen-resolution-is-the-human-eye-incapable-of-detecting-an-increase-in-resolution

Quote

Most people are capable of appreciating resolution far beyond that however. A large majority of people have visual acuity sharper than 20:20 after corrective lenses.

and this:

Quote

In the image above, even if you can only barely resolve the lines from each other, you will still be able to see that the lines on the bottom are shifted slightly to the right relative to the lines on top. Humans can detect the position of line segments down to above 0.1 arc minutes, more than 10X finer than the resolution of the eye.

I tried to find source of x2 per mm of aperture rule, and was not able to find it - I was able to find similar derivation and it also gives lower value:

http://www.rocketmime.com/astronomy/Telescope/MaximumMagnification.html

It starts by assuming that you need 2 arc minutes of separation between the stars - giving this logic:

while I used Rayleigh criteria and visual acuity of 1 arc minute. Reality is probably somewhere in between (neither everyone has 20/20 vision, but mind you, you are using telescope and far/nearsightedness don't count for sharpness - if you need glasses because of that you can say that you have 20/20 or better with telescope - unless you have astigmatism or such) - one needs between 1' or 2' for Rayleigh criteria.

In any case, above website concludes:

or in another words - maximum useful magnification is x1 aperture in mm (twice larger than my above value because they used 2 arc minutes instead of 1 arc minute for derivation).

Btw, I have no idea so far where x2 rule originated from, and wikipedia says:

with emphasis on "citation needed" part, or:

This means that even wiki reckons it's "hearsay"

Budget is also a huge factor for me, and so far, EPs that I like the most are ES - 82 and 68 lines. Deciding factor is somewhere between sharpness and ease of use in sense of eye placement/relief.

While not the most comfortable EPs out there - these are certainly sharpest EPs (ES82 11mm) that I've tried.

I've yet to try out 5.5mm 62 degrees from ES and I'm hoping it will be equally sharp. Ergonomics of it are fine according to simple - "hold it against white wall and check how difficult it is to see field stop" test.

I'm not sure that I was understood properly - above I'm implying that maximum useful magnification is not x2 aperture in millimeters but rather something like x0.45 aperture in millimeters.

Take for example 8" scope - you should be able to see all there is to see at something like x90 power and going above that will not show you more detail. Now, I did not make this up - I just followed the logic that lead to x2 aperture in millimeters - and just got different number instead - that is why I said that my math should be checked.

In any case this deserves to be checked at eyepiece as we all know that seeing allows us to go to x150 or there about on most occasions. Question is - can you see a feature on x150 mag that you can't see on x90 mag (a small crater on the moon, or detail in planetary atmosphere or separation between double stars at the edge of resolving power).

I'm not implying that it will not be "easier" to see the same on higher mag - I'm just saying that in theory you should be able to see it on lower magnification as well.

https://www.firstlightoptics.com/pro-series/sky-watcher-evostar-72ed-ds-pro-ota.html

or if you want something mechanically better, then look at this one:

https://www.teleskop-express.de/shop/product_info.php/info/p8866_TS-Optics-Photoline-72mm-f-6-FPL53-und-Lanthanum-Apo---2-5--R-P-focuser.html

(retractable dew shield, better rings and very good 2.5" R&P focuser)

Again, in SW 72ED class, there is this scope:

https://www.teleskop-express.de/shop/product_info.php/info/p1151_TS-Optics-70-mm-F6-ED-Travel-Refractor-with-modern-2--RAP-Focuser.html

that should be mechanically better.

12 minutes ago, R26 oldtimer said:

Well, next week over the holidays, I'll have the chance to shoot some subs under the stars, and get some darks and bias with the camera properly shielded from outside light, so we can come to a conclusion as where this anomaly comes from, external light or internal ir light or some strange internal calibration.

Vlaiv I want to thank you again for all the help you're giving me.

You are welcome. Once you shoot all of those subs, should you wish for me to take a look at them again - just mention me here or send me a PM - I'll be happy to help.

12 minutes ago, Rob Sellent said:

I'm being dumb, but I don't get this, @vlaiv. I'm not disputing the math - I couldn't nor wouldn't. But I don't get how maximum useful magnification for a given eye is x0.45 the diameter of a scope's lens or mirror in mm. I apologise for calling you up. I've no doubt I've missed something  but it's interesting what you've written and I'd like to understand it

Sure, no problem, I'll explain in detail reasoning behind all of that and math is written above.

When starting off in astronomy, max useful magnifications was one of those things I read about and adopted. At some point when I started being interested in resolution, both in terms of imaging (planetary and deep exposure) and visual I gathered - either read about or concluded, but I do think I read about it somewhere that maximum useful magnification of the telescope is defined as magnification that matches resolving power of telescope and that of human eye.

Maybe best explained on pair of double stars that have just right separation - first part is telescope and depending on aperture size of telescope (under perfect or non existent atmosphere) following can happen:

You can clearly distinguish two stars given their angular separation, you can just make out that there are two discs and you can't be certain if that is some weird elliptical object or two stars.

This separation is related to size of airy disk and is equal to its radius - from max value to first minimum. This is also called Rayleigh criterion for telescope resolution (visual).

Next part in the equation is to match that image which is really tiny if there is no magnification - for usual amateur apertures this criteria gives about 1" resolution (or distance between stars that can be resolved) - that is too small angular separation to be seen by naked eye without magnification.

In fact, according to "eye science" - visual acuity of average human is about 1' - one arc minute, you can read up on this topic here:

https://en.wikipedia.org/wiki/Visual_acuity

But for simple explanation - on eye exam if you have 20/20 vision (considered average / regular for healthy human) this letter:

is 5 arc minutes in size, so in order to know it is E letter - you need to be able to distinguish features of 1/5 in size or 1 arc minute (otherwise you could think it is some other letter as you would not be able to see gaps or little bar or wiggles - each of them is 1/5 in size compared to whole letter - which is sort of matrix of 5x5 elements)

Now if you can magnify two stars enough so that separation between them visually is 1 arc minute, then you should be able to see them as two stars.

In order to magnify something that is close to 1 arc seconds to size of 1 arc minute - you need magnification of x60 (1 arc minute is 60 times larger than one arc second).

For actual math, you can see my earlier post, but bottom line is - I was under impression that x2 is calculated by matching resolving power of the telescope and resolving power of human eye with some magnification, but once I actually tried to do math in response to this thread - it turned out that factor is not x2 but rather x1/2 - so I figured that maybe it was originally miscalculated - it can be easily mistaken as 1/value if you are not careful what multiplies and what divides in math above.

One more important thing to note is that you can use higher magnifications but you will not see additional detail - all the detail that telescope can resolve - so can you at this magnification (threshold magnification), and at higher magnification things will be larger but no additional detail will be seen. This is why it is called maximum practical magnification - like with sampling - you can oversample but you will get larger blurrier image without additional detail.

Now this is interesting, I just did the math while trying to confirm old rule - x2 aperture size, and I think that someone messed up calculation long time ago

It is not x2 aperture diameter in mm, but rather aperture in diameter / 2.

Let's go thru all of that together to see if I made a mistake, shall we?

Angular resolution of telescope is 1.22 * lambda / diameter, where lambda is wavelength of light usually taken to be 550nm - green light where human eye is most sensitive, and diameter is taken for our case in mm (we will need this to convert units later) - this gives angular resolution in radians (we will need to convert to arc seconds or similar later)

Angular resolution of human eye is said to be 1 minute of arc on average (I assumed that this is for 20/20 vision but I now found why is this - and it can be more in some cases - but let's go with 1 minute of arc).

Maximum useful magnification of the telescope is magnification that is needed so that resolving power of telescope is matched to resolving power of human eye - any more magnification and you will not see additional detail - all detail that is there by telescope resolving power will be resolved by human eye - you will be able to see it.

magnification * 1.22 * lambda / diameter = 1 minute of arc

Let's see what is magnification based on diameter after we convert to same units

magnification = 60 * diameter * 1000 * pi / (1.22 * 0.55 * 180 * 60 * 60)

- first 60 converts from 1 minute of arc to seconds of arc

- 1000 converts from micrometers to millimeters - matching diameter and wavelength of green 550nm = 0.55um = 0.55 / 1000 mm

- pi converts radians to degrees  ( rad * 180 / pi = deg)

- 180 is other part of converting radians to degrees)

- 60 * 60 - is converting degrees to arc seconds ( 60 arc minutes in degree, 60 arc seconds in arc minute)

magnification = diameter * 3141.5 / (1.22 * 0.55 * 180 * 60) = diameter * 3141.5 / 7246.8 = diameter / 2.3

You can test that I'm right using online airy disk diameter calculators like this one:

http://www.wilmslowastro.com/software/formulae.htm#Airy

For 200mm scope it gives airy disk size of - 1.28" and we need half of that for Rayleigh criteria (airy radius - or distance from center to first minima) = 0.64"

How much times we need to magnify that to make it 60" or 1" that our eye can resolve - 60/0.64" = x93.75

Expressed in diameter of scope (we said it was 200mm) this would be 200 x something = 93.75 -> something is 2.133333

(difference being use of 510 vs 550nm in calculations, but if you put 550 instead you get airy disk size of 1.38" or resolving power of 0.69" or magnification of ~87, or 200 / 87 = 2.298 = 2.3 )

It looks like maximum useful magnification is in fact only about x0.45 diameter and not x2 diameter in millimeters????

13 minutes ago, Ryan_86 said:

That's amazing Vlaiv condsodering how little data I have. Did you use just sub 8, 9 & 10 of the Ha. I will attempt using your workflow👍🏼

Much appreciated pal. 

Ryan

No, I stacked all 10 of them. I did use outlier rejection (that is sort of hot pixel correction) to prepare subs. I then used "split bin" - it's a sort of binning technique where you split your subs into larger number of subs without actually binning them and then stack all of those - which is very much like first binning then stacking but has some advantages - like dealing with remaining hot pixels and not introducing pixel blur. I did bin x3 in this fashion - that is why resulting image is x3 less in height and width than original data (there will be no difference when viewed at screen size - like here in forum post, but you will see difference if you zoom to 100%). I opted to bin because for this data - SNR has priority over detail captured (and let's face it - stars are not quite round to be viewed at full resolution).

In the end I used special stacking method that I developed that accounts for different SNR of individual subs - a bit like PixInsight weighted average, but mine does not assign weight to whole frame but bunch of weights to groups of pixels and noise estimator is different. I used that approach because of different temperature between Ha subs.

For other channels I just used sigma reject stacking - they are pretty uniform in SNR.

Last step was to remove background - again I used my algorithm for that - it's pretty straight forward - you start by iteratively selecting lowest value pixels in the image (like sigma reject - but instead of rejecting you actually look at those pixels as background) and fitting plane thru those "background" pixels only - that background is subtracted and process is repeated (some pixels that were too bright because they were part of gradient are now in background and we need to examine them as well) iteratively until residual gradient is rather small (less than 0.000001 or something like that).

You have resulting stacked and wiped linear data attached as fits above - so you can download those and have a go at processing that as well.

Do have a look at StarNet++. Last image was done with it. Here it is similar workflow except I did following:

- Ha luminance stretched layer was processed in StarNet++ to get starless version. Stars were extracted as regular layer + starless layer set to difference (or subtract or whatever operation is for subtraction) - that leaves only stars.

- Color layers were processed in StarNet++ and then combined. After combination I did channel mixer to get right "tone" to the image - much easier since you don't have any stars and you don't have to worry about odd magenta colors. Raise level of green in red channel to get that golden tone and remove green cast (converts green to yellow), and subtract a bit of blue channel from red to get better blue tones. Raise a bit of green in blue channel to get that "sky like" blue rather than deeper sort of marine blue.

- Apply starless Ha luminance to resulting color composition as described above

- add stars as top player with screen blend mode (I think it was screen - one that only lets light parts - stars from top layer).

I think I'm predictable at these things - numero uno for both screen size and 1:1 (100% zoom) setting.

Second one just says: "I'm trying too hard".

Mind you, I would probably be less critical of second one if I've seen it alone - meaning its also good processing, but I prefer subtlety of first one better.

2 hours ago, R26 oldtimer said:

Ok, I'm back. Here is the google drive link to a folder of fresh bias frames shot at iso800, 1/4000, .arw.

https://drive.google.com/open?id=1xU836mlPuEo7TlNXZcawPaubdt3_86Lk

Ah, lovely - these are really nice looking bias subs - very uniform and very much like good bias subs.

Look at their stats (again - FitsWork used to convert):

That is so nice - mean value is really nice at somewhere at ~0.3 with very small deviation sub to sub.

Master bias looks really uniform and flat:

and if we zoom in we can see vertical stripes that are common in master bias (either vertical or horizontal stripes are common).

What is not so good about this is that there is certain "signature" to dark subs that could be due to amp glow - which could be problematic when doing dark scaling - but you really need to try dark scaling to see if it will work.

Let's see if we can make sense from those initial darks now that we have master bias:

Quote

DSC263 was at room temperature ( around 22degrees)

DSC261 was at room temperature ( around 22degrees) but with the peltier cooler on.

DSC265 was shot outside (at temperature of around 12degrees)

DSC269 was shot outside (at temperature of around 12degrees) but with the peltier cooler on.

I just repeated what you written about conditions for initial darks so we can try to make sense out of it and to see if we can determine how good your cooling is.

DSC261:

Average signal level (dark signal after bias removal) is 0.018281393 ADU/s - I accounted for 914s exposure and that is why dark current is given in ADU per second (we can compare those values)

DSC263:

Dark signal is 0.018862519 ADU/s

DSC265:

Dark signal is 0.018966408 ADU/s

DSC269:

Dark signal is 0.019258152 ADU/s

These numbers seriously don't fit your information on how these darks were generated. Either that or there is something going on here that is "non standard" thing. Dark current doubling temperature is about 6C - so if we only observe subs that were without cooling and we look at 22C and 12C ambient - it is not unreasonable to assume that sensor was at least 6C cooler in lower ambient temperature - which means it should have about half of dark current in "colder" frames than in "hotter" - but difference between all is very small - few percent at most.

One explanation might be:

There is "amp glow" that is quite serious but does not behave like regular thermal noise - it is added in readout but does not exist in bias subs (which would be rather strange). Maybe it is light leak? If camera is modified and IR filter removed - any chance you had IR leak into camera while taking darks? This could happen if you had only plastic cover on camera - some types of plastics are transparent to IR radiation.

If you suspect this to be the case - put aluminum wrapping around the camera and redo darks to see if there is difference (do one shot with and one shot without IR protection - make sure you have same exposure length and compare mean ADU value in both - if significantly different - you have light leak).

Other explanation might be:

Camera does "internal calibration" on long exposure shots and removes "predefined constant value" from image based on exposure length - that is bad thing as there is no way you can do proper calibration then. Best that you can do is take darks of same duration as lights each time you image at the end of the session (or better yet - half at the start and half at the end - to even out any temperature difference).