Jump to content

Narrowband

Jupiter, February 28-2018


astroavani

Recommended Posts

  • Replies 38
  • Created
  • Last Reply

Lovely Jupiter as always.

Hope that you don't mind this, but I've used this image to make small "experiment".

According to some of my research on optimum resolution for planetary imaging, with ASI290, one should go for F/11 focal ratio. Since your C14 is already F/11 instrument, in my view using x2 Powermate is just waste of SNR with no additional information recorded. Image of yours gave excellent setup to test this hypothesis. If I'm right, down sampling this image x2 should result in no detail loss what so ever. So I took your image, downsampled it x2 and then resampled it to original size, and then I made a difference between two images and here is montage of result:

resample_demo.thumb.png.2e4f9ac8a5719fc6c0a35cafbbb5e62e.png

If you look at this montage at full resolution you will notice that there is no difference in Jupiter detail at all between x2 downsampled version and original one. Only difference observable is in information and credits text and that clearly shows on difference image.

So I would recommend to try using ASI290 without powermate on C14 - it should give you shorter exposures and/or better SNR, without any detail loss (due to diffraction on scope aperture).

Link to comment
Share on other sites

1 hour ago, vlaiv said:

According to some of my research on optimum resolution for planetary imaging, with ASI290, one should go for F/11 focal ratio. Since your C14 is already F/11 instrument, in my view using x2 Powermate is just waste of SNR with no additional information recorded.

Hi vlaiv, I've always understood as 'rule of thumb' that optimum focal ratio is 6 x pixel size (microns) so that would suggest F17-F18 is best for ASI290 with 2.9micron pixels, so I'm interested in your source of F11 being optimum for that camera. If, as I understand, F18 would be optimum then F22, would be a better option than F11, but I'm very interest to hear from Avani. Regards, Geof

Link to comment
Share on other sites

17 minutes ago, geoflewis said:

Hi vlaiv, I've always understood as 'rule of thumb' that optimum focal ratio is 6 x pixel size (microns) so that would suggest F17-F18 is best for ASI290 with 2.9micron pixels, so I'm interested in your source of F11 being optimum for that camera. If, as I understand, F18 would be optimum then F22, would be a better option than F11, but I'm very interest to hear from Avani. Regards, Geof

My research was based on size of airy disk and MTF that it produces in frequency domain. So basic premise is of course Nyquist theorem.

Since MTF of telescope aperture has cut off point in frequency domain, I wondered for what frequency there is 99% attenuation, and what would be proper sampling to capture that frequency. Rationale was that there is no need to capture frequencies past cut off point - since those are effectively attenuated to 0, but capturing frequencies that are attenuated to only 1% of their original amplitude does indeed capture some information if we include frequency restoration process (such as deconvolution or wavelets). Mind you, those two are so close that you won't go wrong if you sample for cut off frequency.

Most of rules of thumb were "devised" based on optical resolution and Nyquist, so it was for some time assumed that 2 pixels per Airy disk radius is good sampling value. Then people noticed that if one uses 3 pixels per airy disk radius - one gets more detail - and this was erroneously explained that 2D Nyquist needs x3 instead of x2 highest frequency. I've seen even values of 3.3 pixels per airy disk radius mentioned.

I did not do rigorous mathematical analysis, but rather I used FFT to create Airy disk patterns from both unobstructed and obstructed apertures, and their corresponding MTFs - and then I measured at what frequency attenuation is close to 1% (or 0.01). It turns out that it is somewhere around 2.4 x Airy disk radius (if one does rigorous mathematical analysis and calculates fourier transforms of airy pattern function we will be able to have exact cut off value). I've made comprehensive thread about this research - with graphs and simulations:

 

 

 

Link to comment
Share on other sites

28 minutes ago, vlaiv said:

My research was based on size of airy disk and MTF that it produces in frequency domain. So basic premise is of course Nyquist theorem.

Since MTF of telescope aperture has cut off point in frequency domain, I wondered for what frequency there is 99% attenuation, and what would be proper sampling to capture that frequency. Rationale was that there is no need to capture frequencies past cut off point - since those are effectively attenuated to 0, but capturing frequencies that are attenuated to only 1% of their original amplitude does indeed capture some information if we include frequency restoration process (such as deconvolution or wavelets). Mind you, those two are so close that you won't go wrong if you sample for cut off frequency.

Most of rules of thumb were "devised" based on optical resolution and Nyquist, so it was for some time assumed that 2 pixels per Airy disk radius is good sampling value. Then people noticed that if one uses 3 pixels per airy disk radius - one gets more detail - and this was erroneously explained that 2D Nyquist needs x3 instead of x2 highest frequency. I've seen even values of 3.3 pixels per airy disk radius mentioned.

I did not do rigorous mathematical analysis, but rather I used FFT to create Airy disk patterns from both unobstructed and obstructed apertures, and their corresponding MTFs - and then I measured at what frequency attenuation is close to 1% (or 0.01). It turns out that it is somewhere around 2.4 x Airy disk radius (if one does rigorous mathematical analysis and calculates fourier transforms of airy pattern function we will be able to have exact cut off value). I've made comprehensive thread about this research - with graphs and simulations:

 

 

 

Thanks vlaiv, I knew that I was going to rapidly get out of my depth when I posted my comment. I tried to follow your analysis, but I drowned (glug, glug, glug....) Thanks anyway :icon_confused::happy7:

Link to comment
Share on other sites

6 hours ago, vlaiv said:

Lovely Jupiter as always.

Hope that you don't mind this, but I've used this image to make small "experiment".

According to some of my research on optimum resolution for planetary imaging, with ASI290, one should go for F/11 focal ratio. Since your C14 is already F/11 instrument, in my view using x2 Powermate is just waste of SNR with no additional information recorded. Image of yours gave excellent setup to test this hypothesis. If I'm right, down sampling this image x2 should result in no detail loss what so ever. So I took your image, downsampled it x2 and then resampled it to original size, and then I made a difference between two images and here is montage of result:

resample_demo.thumb.png.2e4f9ac8a5719fc6c0a35cafbbb5e62e.png

If you look at this montage at full resolution you will notice that there is no difference in Jupiter detail at all between x2 downsampled version and original one. Only difference observable is in information and credits text and that clearly shows on difference image.

So I would recommend to try using ASI290 without powermate on C14 - it should give you shorter exposures and/or better SNR, without any detail loss (due to diffraction on scope aperture).

 

5.jpg

Link to comment
Share on other sites

Hello Geof!
I understand your settings and we already discussed a lot about this in other forums, including in astrobin itself I have a colleague who thinks exactly like you, even thought that I would have nothing to gain by using ASI 290 with smaller pixels than ASI 224.
I've read about it, but what I can tell you is that practice does not always follow the theory. I do not mean the scientific part that is perfectly correct, but rather the part that concerns image capture and processing.
At least for me I find it very easy to capture and work the image captured with the Powermate 2X. It is difficult to explain, but it is several small details that favor the capture and processing of a larger image. Even the color seems to get better balanced.
All the great astrophotographers today work with the system between f / 20 and f / 40, usually with 4000mm to 10000mm of focal length, so I believe that some reason must exist.
I know that the ASI 290 would not bring me any more gain than the ASI 224 that has a larger pixel, but just as an example, ASI290 has allowed me to easily capture the division of Encke even when seeing is not the best, already with ASI 224 I rarely managed to capture it well, only on nights of very good seeing.
I think one of the truths is that we end up getting used to a certain setup and it is difficult to change because it changes all parameters of capture and processing.
Best regards!

 

Link to comment
Share on other sites

5 minutes ago, astroavani said:

Olá Geof!
Eu entendo suas configurações e já discutimos muito sobre isso em outros fóruns, inclusive na própria astrobina eu tenho um colega que pensa exatamente como você, até pensei que não teria nada a ganhar usando o ASI 290 com pixels menores que o ASI 224.
Eu li sobre isso, mas o que posso dizer é que a prática nem sempre segue a teoria. Não me refiro à parte científica que está perfeitamente correta, mas sim a parte que diz respeito à captura e processamento de imagens.
Pelo menos para mim, acho muito fácil capturar e trabalhar a imagem capturada com o Powermate 2X. É difícil de explicar, mas são vários pequenos detalhes que favorecem a captura e o processamento de uma imagem maior. Mesmo a cor parece ficar melhor equilibrada.
Todos os grandes astrofotógrafos hoje trabalham com o sistema entre f / 20 e f / 40, geralmente com 4000mm a 10000mm de distância focal, então acredito que algum motivo deve existir.
Eu sei que o ASI 290 não me traria mais ganho do que o ASI 224 que tem um pixel maior, mas apenas como exemplo, o ASI290 me permitiu capturar facilmente a divisão da Encke, mesmo quando vê não é o melhor, já com ASI 224 Eu raramente consegui capturá-lo bem, apenas nas noites de muito bom ver.
Eu acho que uma das verdades é que acabamos nos acostumando com uma certa configuração e é difícil de mudar porque ela muda todos os parâmetros de captura e processamento.
Cumprimentos!
 

Thanks Avani, I translated to English in Google as my Portuguese is not so good (well non existent..!!), so I hope I understand you correctly :icon_biggrin:.

Hello Geof!

I understand your settings and we already discussed a lot about this in other forums, including in astrobina itself I have a colleague who thinks exactly like you, I even thought I would have nothing to gain by using ASI 290 with pixels smaller than ASI 224.

I read about it, but what I can say is that practice does not always follow the theory. I am not referring to the scientific part that is perfectly correct, but rather the part that concerns the capture and processing of images.

At least for me, I find it very easy to capture and work the image captured with the Powermate 2X. It is difficult to explain, but it is several small details that favor the capture and processing of a larger image. Even the color seems to get better balanced.

All the big astrophotographers today work with the f / 20 to f / 40 system, usually 4000mm to 10000mm focal length, so I think some reason must exist.

I know that the ASI 290 would not bring me more gain than the ASI 224 which has a larger pixel, but just as an example, the ASI290 allowed me to easily capture Encke's division, even when I see it is not the best, already with ASI 224 I rarely managed to capture it well, only on very nice evenings.

I think one of the truths is that we end up getting used to a certain configuration and it is difficult to change because it changes all parameters of capture and processing.

Greetings!

I asked my planetary imaging tutor/mentor about it today and he said that regardless of the theory the proof is always in the practice and that it is often easier to work with larger image on the screen, so I think he agrees with you. He said F18 would be ideal for the ASI290 if a x1.5 Powermate is available, but I myself also usually image with x2 Powermate with C14 so F22 using an older ASI120.

By the way Avani, as others already said, your images are superb :icon_biggrin:

Regards, Geof

 

Link to comment
Share on other sites

21 minutes ago, astroavani said:

Yes!

 

Other than slightly stronger processing (not to my liking, I prefer processing that you've done on first image), this image to my eye contains the same level of detail to previous one?

Being far more experienced in planetary imaging what is your opinion? I'm not talking about ease of processing, or capture, that is personal preference (and not up for discussion, I personally find your processing marvelous and admire your images), I'm interested in only in level of detail. Do you think working on F/11 captures same amount of detail?

Link to comment
Share on other sites

3 hours ago, vlaiv said:

Other than slightly stronger processing (not to my liking, I prefer processing that you've done on first image), this image to my eye contains the same level of detail to previous one?

Being far more experienced in planetary imaging what is your opinion? I'm not talking about ease of processing, or capture, that is personal preference (and not up for discussion, I personally find your processing marvelous and admire your images), I'm interested in only in level of detail. Do you think working on F/11 captures same amount of detail?

Yes, undoubtedly, the details captured are the same and they are there!
Only in a larger image is it easier to perceive them, and precisely for that reason to make a meticulous processing.
What I like about the oversampling image is just what you said, that is, it seems to be smoother and consequently nice.

Link to comment
Share on other sites

Hi Avani and vlaiv,

I discussed this very interesting topic some more with my friend and we did some comparisons of our own, so I hope you (Avani) don't mind me playing with your images. Here is the comparison with the F11 image (on left) resized to same as F22 image (on right).

5a9945eeb5768_AvaniJupitercomparison.thumb.jpg.f59a97b836dcd03beec89c426b4a0e82.jpg

Perhaps the biggest give away is the subtle detail in the EZ Cyan festoon (see inserts) which is significantly better in the F22 image on the right than in the F11 image. To me the resized F11 image is more blocky and contrasty which as Avani says may in part be to it being more difficult to process the smaller F11 image, but whatever the reason, for me the final much smoother processed F22 image is easily the superior of the 2 images, so I think F22 wins in this practical experiment... :happy7:

Both of course are excellent images and I wish I could image Jupiter at 77 degrees altitude....!! For me this year it barely gets above 20 degrees at best, so it will be very difficult to capture much detail I think...!! Even worse, Mars and Saturn will reach only just over half that altitude from my location.... :hmh:.

Best regards, Geof

Link to comment
Share on other sites

6 minutes ago, geoflewis said:

Hi Avani and vlaiv,

I discussed this very interesting topic some more with my friend and we did some comparisons of our own, so I hope you (Avani) don't mind me playing with your images. Here is the comparison with the F11 image (on left) resized to same as F22 image (on right).

5a9945eeb5768_AvaniJupitercomparison.thumb.jpg.f59a97b836dcd03beec89c426b4a0e82.jpg

Perhaps the biggest give away is the subtle detail in the EZ Cyan festoon (see inserts) which is significantly better in the F22 image on the right than in the F11 image. To me the resized F11 image is more blocky and contrasty which as Avani says may in part be to it being more difficult to process the smaller F11 image, but whatever the reason, for me the final much smoother processed F22 image is easily the superior of the 2 images, so I think F22 wins in this practical experiment... :happy7:

Both of course are excellent images and I wish I could image Jupiter at 77 degrees altitude....!! For me this year it barely gets above 20 degrees at best, so it will be very difficult to capture much detail I think...!! Even worse, Mars and Saturn will reach only just over half that altitude from my location.... :hmh:.

Best regards, Geof

Not sure that comparison is fair. F/11 image is processed slightly differently (overdone compared to F/22) - but there is another significant factor. F/22 image is png - meaning no compression losses. F/11 image is jpeg and it shows compression artifacts (not so much in the image, but I also resampled it to x2 smaller and returned to original size and difference clearly shows jpeg compression artifacts - 8x8 block boundaries).

Ideally, if Avani is willing to share unprocessed 16 bit stack for each image we could do experiments on those to see what level of difference there is if any.

Link to comment
Share on other sites

3 minutes ago, vlaiv said:

Ideally, if Avani is willing to share unprocessed 16 bit stack for each image we could do experiments on those to see what level of difference there is if any.

I agree that the comparison is not completely fair, but I'm not sure if I could process either as well as Avani already did. I do know that when I process small images I resize them in PS to work with a larger image so that it is easier (for me anyway) to see the details being extracted, but that is usually after I already performed wavelets in Registax so still not the same process.

Link to comment
Share on other sites

3 minutes ago, geoflewis said:

I agree that the comparison is not completely fair, but I'm not sure if I could process either as well as Avani already did. I do know that when I process small images I resize them in PS to work with a larger image so that it is easier (for me anyway) to see the details being extracted, but that is usually after I already performed wavelets in Registax so still not the same process.

Yes, processing part is probably hardest thing to do "in a controlled setting". Registax does wavelet transform on a fixed size basis - there is no way to control actual level of detail in each of six sub bands - it is relative to actual sampling used - this is one of the reasons why people find processing easier with a certain setup. Still we could try out some combinations like wavelets on native resolution - then resample, or wavelets on resampled version (to equalize wavelet basis response). Then there is choice of resampling algorithm - find out which one does the least alterations and/or matches frequency restoration the best.

I find much joy in that - finding the best way to do things, but also understanding why it is the best way to do things.

Link to comment
Share on other sites

4 minutes ago, astroavani said:

Follows the output of AS! 2 with a sharpned of 50%.
Both images were TIF and were converted to PNG.
This has not been passed by Registax.

Thanks Avani, I will see what I can do, but I suspect not nearly as good as you :wink2:. Best regards, Geof

Link to comment
Share on other sites

2 hours ago, geoflewis said:

Thanks Avani, I will see what I can do, but I suspect not nearly as good as you :wink2:. Best regards, Geof

Well I tried using Registax wavelets and unsharp mask in PS (CS2), both separately and together, but I could not extract the same level of detail in the smaller F11 image as is possible with the larger F22 image. I will not post my versions as really they were no better than what Avani already achieved, so for me the resoltion in F11 is just not quite as good as F22, or at least if the detail is there then it is more difficult to process the smaller image. Maybe a number of other things affect the captured data, e.g. focus is slightly different, or subtle change in seeing, though of course the data captured in both images is superb, but for me it is just a bit superior at F22. I am, however, interested to see what vlaiv can achieve. It is a very worthwhile experiment and a lot of fun to play with such excellent data, so many thanks to Avani for sharing it.

Avani, a separate question for you if you do not mind please. I note that you used 50% sharpening in AS2, something that I never used, relying only on Registax and PS for sharpening, so in your opinion what does AS2 sharpening add to improve the data? I am very keen to improve my technique, if you do not mind sharing with me, so many thanks in advance.

Best regards, Geof

Link to comment
Share on other sites

Here is my result of F/11 image, unfortunately SGL won't allow me to access original 16 bit PNG Avani uploaded - when I download the image it is always saved as 8 bit. For this reason there is a bit of noise in the image due to loss of precision between 16 and 8 bit. Still, I believe F/11 version does indeed contain all the information that is in F/22 image.

result_2.thumb.png.a1eeb8123f027d239ce5a9ab21ce6c54.png

Color balance and the rest of processing is obviously different, but detail is there (in my view).

Link to comment
Share on other sites

6 hours ago, geoflewis said:

Well I tried using Registax wavelets and unsharp mask in PS (CS2), both separately and together, but I could not extract the same level of detail in the smaller F11 image as is possible with the larger F22 image. I will not post my versions as really they were no better than what Avani already achieved, so for me the resoltion in F11 is just not quite as good as F22, or at least if the detail is there then it is more difficult to process the smaller image. Maybe a number of other things affect the captured data, e.g. focus is slightly different, or subtle change in seeing, though of course the data captured in both images is superb, but for me it is just a bit superior at F22. I am, however, interested to see what vlaiv can achieve. It is a very worthwhile experiment and a lot of fun to play with such excellent data, so many thanks to Avani for sharing it.

Avani, a separate question for you if you do not mind please. I note that you used 50% sharpening in AS2, something that I never used, relying only on Registax and PS for sharpening, so in your opinion what does AS2 sharpening add to improve the data? I am very keen to improve my technique, if you do not mind sharing with me, so many thanks in advance.

Best regards, Geof

Hi Geof!
In fact in this catch I used a weak sharpened, usually I put in 35% that already comes out sharper. I think you know that the smaller the% the greater the sharpening.
I use either of the two output images to take to Registax, but I've noticed that there is a slight gain in detail using the image already sharpened by A! 2 but we have to be very careful not to overprocess.
I use the Registax layers to form a gentle slope from left to right, with Layer 1 around 25%, 2 in 20%, 3 in 15%, 4 in 10%, 5 in 5%, and in Layer 6 I I do not move.
I also use the filters only denote layers 1 and 2 at most 25%, all to prevent the image from looking painted.
I make a few more adjustments in Hystogran, contrast, brightness and RGB balance. Saved in the source folder and then open in Fitswork for deconvolution.
I use the Irfan View for some corrections of brightness contrast and color, change size, format, etc.
Currently after that I'm testing the Topaz Denoise which seems to improve the noise a bit without loss of detail.
In principle it is!
Best regards

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. By using this site, you agree to our Terms of Use.