theoretical question F25 v F15 when imaging

[iwols]

Hi just a theoretical question but would i notice much difference imaging at f15 compared to f25 on say jupiter using a C8 edge ,any thoughts thanks

[symmetal]

The image at f15 will be just over half the size of the f25 image.  For planetary (lucky imaging) you want a focal ratio that optimises your likelihood of getting the most detail, that is, not under or over sampling. This depends on the pixel size of your camera and also whether it's colour or mono.

For a mono camera the optimal focal ratio is equal to the camera pixel size in microns x 3.

For a colour camera it's the pixel size in microns x 6.

The theory behind these figures is well explained in this article, which is in Dutch but Google translate works perfectly. 🙂

Alan

 

[iwols]

well my options are my asi462 with 2.5x barlow giving me this

 

or do i get a asi 678 which gives me a brighter smaller image,with faster frame rate and less gain,but do i loose quality making the image the same size as i get with the 462??

or am i getting as good as i can get with my C8 edge thanks 

 

Edited September 28, 2022 by iwols

[iwols]

another one with the 462 and 2.5 barlow

 

[newbie alert]

On 23/09/2022 at 15:02, symmetal said:

The image at f15 will be just over half the size of the f25 image.  For planetary (lucky imaging) you want a focal ratio that optimises your likelihood of getting the most detail, that is, not under or over sampling. This depends on the pixel size of your camera and also whether it's colour or mono.

For a mono camera the optimal focal ratio is equal to the camera pixel size in microns x 3.

For a colour camera it's the pixel size in microns x 6.

The theory behind these figures is well explained in this article, which is in Dutch but Google translate works perfectly. 🙂

Alan

 

So the op is using a F10 SCT, so in theory he should be using a pixel size no smaller than 3.333333 to image at its native FL and any larger he should be using a focal reducer?

Doesn't seem to stack up between theory and practice to me

[symmetal]

Your ASI462 has 2.8μm pixels and being a colour camera by the calculation above your optimum focal ratio is f17.4

vlaiv suggests pixel size x 5 for colour or mono cameras which means your focal ratio should be f14.5

This suggests a 1.5x barlow to be theoretically closer to what you need. With your 2.5x barlow you're oversampling so you should resize the final image to 60% size to give the best looking image with the detail available. Your bigger image will be softer but show no more actual detail than the 60% one. The drawback with oversampling is that your recorded image will be dimmer compared to using optimal sampling, so to compensate you either increase the exposure which reduces your chance of freezing the seeing as well as reducing the framerate, or you increase the camera gain and so need a longer video to help mitigate the increased noise.

The ASI678 has 2.0μm pixels so your native f10 is exactly 5x pixel size so no barlow required. The image size will be similar to your 60% reduced one from above with similar potential detail. Your recorded image will be brighter though which avoids the oversampling pitfalls mentioned above.

Jupiter is bright so you will likely get a reasonable exposure duration, and therefore framerate with your ASI462 and 2.5 barlow, so when 60% resized, the result may not look any different compared to the ASI678 without a barlow.

Saturn and Mars are dimmer objects though so the drawbacks of oversampling will be more apparent and these will likely look better with the ASI678.

3 hours ago, newbie alert said:

So the op is using a F10 SCT, so in theory he should be using a pixel size no smaller than 3.333333 to image at its native FL and any larger he should be using a focal reducer?

Doesn't seem to stack up between theory and practice to me

iwols is using a colour camera so optimal focal ratio is x5 or x6 depending on whose reasoning you take, so 2μm pixels is pretty optimal at native FL. Any larger pixel size would then require a barlow to increase the focal ratio rather than a focal reducer. 🙂

Alan

[newbie alert]

6 minutes ago, symmetal said:

Your ASI462 has 2.8μm pixels and being a colour camera by the calculation above your optimum focal ratio is f17.4

vlaiv suggests pixel size x 5 for colour or mono cameras which means your focal ratio should be f14.5

This suggests a 1.5x barlow to be theoretically closer to what you need. With your 2.5x barlow you're oversampling so you should resize the final image to 60% size to give the best looking image with the detail available. Your bigger image will be softer but show no more actual detail than the 60% one. The drawback with oversampling is that your recorded image will be dimmer compared to using optimal sampling, so to compensate you either increase the exposure which reduces your chance of freezing the seeing as well as reducing the framerate, or you increase the camera gain and so need a longer video to help mitigate the increased noise.

The ASI678 has 2.0μm pixels so your native f10 is exactly 5x pixel size so no barlow required. The image size will be similar to your 60% reduced one from above with similar potential detail. Your recorded image will be brighter though which avoids the oversampling pitfalls mentioned above.

Jupiter is bright so you will likely get a reasonable exposure duration, and therefore framerate with your ASI462 and 2.5 barlow, so when 60% resized, the result may not look any different compared to the ASI678 without a barlow.

Saturn and Mars are dimmer objects though so the drawbacks of oversampling will be more apparent and these will likely look better with the ASI678.

iwols is using a colour camera so optimal focal ratio is x5 or x6 depending on whose reasoning you take, so 2μm pixels is pretty optimal at native FL. Any larger pixel size would then require a barlow to increase the focal ratio rather than a focal reducer. 🙂

Alan

Oh yes of course..

Correct sampling is the relationship between pixels and FL using Nyquist, which maybe more accurate than fr

[vlaiv]

4 hours ago, iwols said:

well my options are my asi462 with 2.5x barlow giving me this

First image is over sampled by factor of 2

4 hours ago, iwols said:

another one with the 462 and 2.5 barlow

Second image is over sampled by factor of 2.5

Not sure how this happened, you probably manipulated size of planet in one of both images - since they are of a different size - and that should not happen for same setup (same camera + barlow if barlow distance was constant).

In any case - try just camera without barlow, or maybe barlow element real close to give you x1.5 at maximum.

 

[iwols]

thanks guys so where does this leave me please using my 462 with a 1.5x barlow

Quote ...The ideal pixel size for OK Seeing (2-4" FWHM) seeing is: 0.67 - 2" / pixel.

This combination leads to over-sampling. Will require a good mount and careful guiding. OK for high magnification solar, lunar or planetary imaging. Might cause signal to noise issues with wide-field imaging.

 

[vlaiv]

5 minutes ago, iwols said:

hanks guys so where does this leave me please using my 462 with a 1.5x barlow

Quote ...The ideal pixel size for OK Seeing (2-4" FWHM) seeing is: 0.67 - 2" / pixel.

This combination leads to over-sampling. Will require a good mount and careful guiding. OK for high magnification solar, lunar or planetary imaging. Might cause signal to noise issues with wide-field imaging.

You can't use (flawed) calculator for deep sky imaging as a guide for planetary imaging.

Lucky type planetary imaging is completely different activity. In planetary imaging you hope to get around effects of seeing - while with long exposure deep sky imaging - there is no getting around it, and you need to account its effect on resolution of the image.

With planetary type imaging - you go by aperture alone - how much the telescope can resolve without impact of atmosphere (and hope for the best - or rather you image and hope that there will be moments of good seeing that you will then use in stacking software).

Best guideline for you in terms of F/ratio for planetary is pixel size x 4. That will give you best results. That is F/11.6.

Your scope is F/10 - so you might even skip using barlow at all (with SCTs focal length depends on primary to secondary separation, so if you move camera backward - you could even reach F/11 or above).

Alternative is to get barlow element that screws into nose piece of camera and then get short 1.25" extension in hope that you'll get barlow to work around x1.2 (closer you place it to sensor - less magnification it will give you).

 

[CraigT82]

Don’t forget that when using an ADC, the ADC introduces an aberration that gets worse with faster F ratios (hence why you should always use the barlow in front of the ADC)

I wonder at what point the ADC aberration will outweigh the benefit of critical sampling. Might be worth some investigation. 

[vlaiv]

14 minutes ago, CraigT82 said:

Don’t forget that when using an ADC, the ADC introduces an aberration that gets worse with faster F ratios (hence why you should always use the barlow in front of the ADC)

I wonder at what point the ADC aberration will outweigh the benefit of critical sampling. Might be worth some investigation. 

Do you know what type of aberrations does ADC introduce?

Regardless of that, there will be critical sampling for combination of aperture and whatever aberration ADC introduce for given F/ratio.

Critical sampling is related to the point where MTF graph hits 0:

For example, here we have comparison of effects of central obstruction. They introduce "sag" of MTF graph - but don't change critical sampling, all graphs end up in same point.

How much MTF is sagging compared to perfect aperture - represents how hard you need to sharpen to restore proper image. Some aberrations just make MTF sag more, while other can put it to zero "earlier" than it would otherwise hit it.

 

[iwols]

2 hours ago, vlaiv said:

You can't use (flawed) calculator for deep sky imaging as a guide for planetary imaging.

Lucky type planetary imaging is completely different activity. In planetary imaging you hope to get around effects of seeing - while with long exposure deep sky imaging - there is no getting around it, and you need to account its effect on resolution of the image.

With planetary type imaging - you go by aperture alone - how much the telescope can resolve without impact of atmosphere (and hope for the best - or rather you image and hope that there will be moments of good seeing that you will then use in stacking software).

Best guideline for you in terms of F/ratio for planetary is pixel size x 4. That will give you best results. That is F/11.6.

Your scope is F/10 - so you might even skip using barlow at all (with SCTs focal length depends on primary to secondary separation, so if you move camera backward - you could even reach F/11 or above).

Alternative is to get barlow element that screws into nose piece of camera and then get short 1.25" extension in hope that you'll get barlow to work around x1.2 (closer you place it to sensor - less magnification it will give you).

 

thanks so would you go for asi462 with barlow or the straight asi 678,thanks 

[symmetal]

2 hours ago, iwols said:

thanks so would you go for asi462 with barlow or the straight asi 678,thanks 

If you can get a barlow to work at around 1.5 x then the 462 would be fine to use, if not, the 678 with no barlow would be better in my opinion. 🙂

Alan

[vlaiv]

9 hours ago, iwols said:

thanks so would you go for asi462 with barlow or the straight asi 678,thanks 

If you don't want to mess with a barlow - then sure ASI678.

[CraigT82]

11 hours ago, vlaiv said:

Do you know what type of aberrations does ADC introduce?

I don’t fully understand it myself but more info here

 

https://skyinspector.co.uk/adcs-part2/

[iwols]

well im going to try the 678 will post when tested  thanks to all

[iwols]

well this looks better

 

[vlaiv]

7 hours ago, iwols said:

well this looks better

It certainly does

[inFINNity Deck]

On 23/09/2022 at 16:02, symmetal said:

The theory behind these figures is well explained in this article, which is in Dutch but Google translate works perfectly. 🙂

Alan

 

Hi Alan,

thanks for linking to my article. Yesterday, as a result of the discussion in this planetary imaging thread, I have published a second article on that Dutch forum in which I dive deeper into that matter and comparing my method, which was based on being able to solve a Rayleigh-object, with the method @vlaiv used (based on the Spatial Cut-off Frequency) and to my method but then based on the Sparrow and Dawes criteria. The differences between the four methods are small (as calculated for 540nm):

Sparrow-criterion: 3.9 x |px|
Spatial Cut-off Frequency: 3.7 x |px|
Dawes-criterion: 3.6 x |px|
Rayleigh-criterion: 3.0 x |px|

That these differences exist is due to the detail level that is resolved. The Rayleigh-criterion is not yet on the limit (the combined intensities of two objects still shows a dip in between). With the Dawes-criterion there is no visual dip, while at the Sparrow-criterion there is no mathematical dip. In practise our images are affected by aberration, lack of contrast and seeing causing the images to become oversampled (i.e. the worse the seeing, the lower we can choose the factor before we are oversampling). For this I added a section explaining how Fiji can be used to test whether our images are under- or oversampled (as kindly explained by @vlaiv to me, for which I am grateful), which shows that especially seeing has a significant effect on oversampling as a result of which the factor of 3.0 still is safe to use (in most, if not all, situations oversampling will still occur at this level).

The article can be found here: https://www.starry-night.nl/vergroting-onder-de-loep-deel-2-het-optimale-f-getal-nader-beschouwd/

Opening it in Chrome should translate it.

Nicolàs

[vlaiv]

@inFINNity Deck

This nicely describes what I tried to explain in PM.

In this particular instance there is linear dependence between features in two domains - frequency and spatial, and when you choose some object in spatial domain - you will find that it gives you some constant in which to express F/ratio.

It does not however mean that actual cut off frequency depends on that feature - only that they are correlated and that there is linear conversion constant between values.

I think it is best to always refer to true cause of band limited signal and that is aperture and fact that it has hard cut off frequency and that is what we can use as part of criteria for Nyquist sampling theorem.

[symmetal]

@inFINNity Deck Thank you for the updated article. That ties up the differences between your original article and vlaiv's calculations. 🙂

Alan