Bad seeing, focus or something else?

[Anubis1311]

I recently had my first light with a skymax 127 and was quietly impressed with my first image given I didn’t use a Barlow.

So I went and bought a celestron X Cel 2x Barlow and subsequently shot some videos last night. The resulting images are far inferior to those I took a few days ago without the Barlow.

Was it the seeing? Is it a focus thing? From what I’ve read the 2x Barlow is fine with my FL of 1500mm and my asi120mc-s so I’m hoping it’s not that. 
 

Any suggestions would be greatly appreciated 

First image - no Barlow 

Second image - Celestron 2x X Cel Barlow

[vlaiv]

What were exposure lengths in these two recordings?

One of important aspects of planetary imaging is to get exposure length right without looking at the histogram (looking at the histogram and trying to get image bright enough with exposure length is wrong way to do it).

If you were going by histogram (often recommended for other types of photography) - then with barlow, image will be x4 dimmer as light is spread over x4 larger surface - which will then lead to x4 longer exposure to get the same histogram.

That is detrimental to lucky imaging approach - where you aim to freeze the seeing effects instead of them having motion blur effect if exposure is too long.

Ideal exposure length is about 5-6ms, but if you go to say 20ms or 30ms - you'll simply get motion blur in most cases except in the best seeing conditions.

My guess is that exposure length was to blame, and having no barlow is actually closer to critical sampling than having x2 barlow since your Mak is already at F/12 and critical sampling rate for ASI120 (3.75µm pixel size) is F/15. Going over F/15 simply does not make much sense and lowers SNR of recording (and that in turn causes host of other issues - like less precise stacking and less room for sharpening).

[Anubis1311]

22 minutes ago, vlaiv said:

What were exposure lengths in these two recordings?

One of important aspects of planetary imaging is to get exposure length right without looking at the histogram (looking at the histogram and trying to get image bright enough with exposure length is wrong way to do it).

If you were going by histogram (often recommended for other types of photography) - then with barlow, image will be x4 dimmer as light is spread over x4 larger surface - which will then lead to x4 longer exposure to get the same histogram.

That is detrimental to lucky imaging approach - where you aim to freeze the seeing effects instead of them having motion blur effect if exposure is too long.

Ideal exposure length is about 5-6ms, but if you go to say 20ms or 30ms - you'll simply get motion blur in most cases except in the best seeing conditions.

My guess is that exposure length was to blame, and having no barlow is actually closer to critical sampling than having x2 barlow since your Mak is already at F/12 and critical sampling rate for ASI120 (3.75µm pixel size) is F/15. Going over F/15 simply does not make much sense and lowers SNR of recording (and that in turn causes host of other issues - like less precise stacking and less room for sharpening).

Oh wow thanks for the detailed reply!

I think I had my exposure length set to about 50 and gain set to 50 (using video mode on AAP). When I dropped the exp down I could barely see it so I upped it to compensate. Should I up the gain to compensate for the dark image with such a low exp or do I deal with it in processing?

As for the Barlow have I made a mistake?  I do have the asi183mm pro but thought the asi12c-s was specifically designed for planetary. Would the asi183mm pro do a better job with the 2x Barlow. It’s mono which I’m sure has its own benefits however the field of view looked pretty small on astronomy.tools
 

sorry for all the questions 😂

 

 

Edited August 8, 2021 by Anubis1311

[vlaiv]

4 minutes ago, Anubis1311 said:

I think I had my exposure length set to about 50 and gain set to 50 (using video mode on AAP). When I dropped the exp down I could barely see it so I upped it to compensate. Should I up the gain to compensate for the dark image with such a low exp or do I deal with it in processing?

You should raise your gain to control the read noise.

Read noise depends somewhat on gain setting chosen - it usually drops with increased gain.

ZWO has graphs on their site related to each camera:

According to this, best gain is around 64, but you can also use 90-100 range. This only affects brightness of image in capture - which is really not important. Stacking makes all the difference here.

If you use AutoStakkert!3 for stacking (and you should), then you have this option:

which will give you properly bright stack in the end.

Boost your gain and lower your exposure time to something like 6ms to start with.

8 minutes ago, Anubis1311 said:

As for the Barlow have I made a mistake?

Not necessarily. Barlow is a versatile addition - both for imaging and observing. You might not use it with your current cameras and scope, but I'm sure it will come in handy in future.  There is no need to use it with your current setup for imaging.

10 minutes ago, Anubis1311 said:

I do have the asi183mm pro but thought the asi12c-s was specifically designed for planetary.

Any camera that has: Fast readout, small read noise and high quantum efficiency is good planetary imaging camera. I'm inclined to think that ASI183mm is better planetary camera when handled properly.

It has lower read noise, higher QE, and will also have faster readout if you connect it to USB3.0 port and choose appropriate ROI.

According to ZWO website it is well capable of supporting 6ms exposure to the full extent:

(1000ms / 6ms = 166.666FPS and at ROI 640x480 it can capture up to 169.92FPS so 166.66 should be doable).

Only thing that you'll need is filters and filter wheel, because it is mono camera. I prefer OSC for planetary as it is much simpler to work with, but in principle Mono+Filters will have small edge if handled properly (derotation if needed and so on).

14 minutes ago, Anubis1311 said:

Would the asi183mm pro do a better job with the 2x Barlow.

No - it would do even worse.

Certain pixel size dictates maximum F/ratio that should be used. Going with higher F/ratio - just dims the image (lower SNR) and does not record any additional detail - as there are none - that is maximum useful magnification.

ASI183 has 2.4µm pixel size - which has something like F/11 as critical f/ratio

ASI120 has 3.75µm pixel size which has something like F/15 as critical F/ratio

(when I say something like - that is because this depends on wavelength used and in my view, it is best to use wavelength around 500-550nm, although technically if you want to capture everything - you should go with 400nm, but blue part of spectrum is the most affected by seeing and it will be blurred away anyway, so I just go with middle of the spectrum - and peak of human visual sensitivity).

18 minutes ago, Anubis1311 said:

It’s mono which I’m sure has its own benefits however the field of view looked pretty small on astronomy.tools

That is what ROI is used for!

If you capture just central 640x480 (smaller capture file size and faster frame rate):

then you get FOV like this:

[Anubis1311]

2 hours ago, vlaiv said:

You should raise your gain to control the read noise.

Read noise depends somewhat on gain setting chosen - it usually drops with increased gain.

ZWO has graphs on their site related to each camera:

According to this, best gain is around 64, but you can also use 90-100 range. This only affects brightness of image in capture - which is really not important. Stacking makes all the difference here.

If you use AutoStakkert!3 for stacking (and you should), then you have this option:

which will give you properly bright stack in the end.

Boost your gain and lower your exposure time to something like 6ms to start with.

Not necessarily. Barlow is a versatile addition - both for imaging and observing. You might not use it with your current cameras and scope, but I'm sure it will come in handy in future.  There is no need to use it with your current setup for imaging.

Any camera that has: Fast readout, small read noise and high quantum efficiency is good planetary imaging camera. I'm inclined to think that ASI183mm is better planetary camera when handled properly.

It has lower read noise, higher QE, and will also have faster readout if you connect it to USB3.0 port and choose appropriate ROI.

According to ZWO website it is well capable of supporting 6ms exposure to the full extent:

(1000ms / 6ms = 166.666FPS and at ROI 640x480 it can capture up to 169.92FPS so 166.66 should be doable).

Only thing that you'll need is filters and filter wheel, because it is mono camera. I prefer OSC for planetary as it is much simpler to work with, but in principle Mono+Filters will have small edge if handled properly (derotation if needed and so on).

No - it would do even worse.

Certain pixel size dictates maximum F/ratio that should be used. Going with higher F/ratio - just dims the image (lower SNR) and does not record any additional detail - as there are none - that is maximum useful magnification.

ASI183 has 2.4µm pixel size - which has something like F/11 as critical f/ratio

ASI120 has 3.75µm pixel size which has something like F/15 as critical F/ratio

(when I say something like - that is because this depends on wavelength used and in my view, it is best to use wavelength around 500-550nm, although technically if you want to capture everything - you should go with 400nm, but blue part of spectrum is the most affected by seeing and it will be blurred away anyway, so I just go with middle of the spectrum - and peak of human visual sensitivity).

That is what ROI is used for!

If you capture just central 640x480 (smaller capture file size and faster frame rate):

then you get FOV like this:

Thank you so much for the effort you’ve put into helping me!

I didn’t notice when I looked at my asi183mm pro that it set the resolution to 5496x3672 which obviously gave a wider view! Another thing learned 😀 

I have an EFW loaded with RGB filters so could try it 🤔 But might see if I can get the hang of the asi120mc-s first! 

I’ll try again next time with a lower exp and ditch the Barlow as you suggest and hopefully this solves my blurring issue 👍🏻
 

thanks again for your help! 

[Anubis1311]

On 08/08/2021 at 11:29, vlaiv said:

What were exposure lengths in these two recordings?

One of important aspects of planetary imaging is to get exposure length right without looking at the histogram (looking at the histogram and trying to get image bright enough with exposure length is wrong way to do it).

If you were going by histogram (often recommended for other types of photography) - then with barlow, image will be x4 dimmer as light is spread over x4 larger surface - which will then lead to x4 longer exposure to get the same histogram.

That is detrimental to lucky imaging approach - where you aim to freeze the seeing effects instead of them having motion blur effect if exposure is too long.

Ideal exposure length is about 5-6ms, but if you go to say 20ms or 30ms - you'll simply get motion blur in most cases except in the best seeing conditions.

My guess is that exposure length was to blame, and having no barlow is actually closer to critical sampling than having x2 barlow since your Mak is already at F/12 and critical sampling rate for ASI120 (3.75µm pixel size) is F/15. Going over F/15 simply does not make much sense and lowers SNR of recording (and that in turn causes host of other issues - like less precise stacking and less room for sharpening).

Thank you for your advice. I adjusted the exposure times and we are back in business. Won’t win any awards but I’m super happy with it from my little 127 mak 😀 

 

Edited August 12, 2021 by Anubis1311

[vlaiv]

6 minutes ago, Anubis1311 said:

Thank you for your advice. I adjusted the exposure times and we are back in business. Won’t win any awards but I’m super happy with it from my little 127 mak

Why is your Jupiter gigantic in that image?

It is 950px in diameter, but your telescope is only capable of recording it to resolution of 119px at the moment (48.8" diameter, 0.41"/px).

This is what your image should look like:

And when you present it properly - it really looks like a excellent capture, but question is - why is it so large in your processing? What sort of barlow did you use and did you drizzle for some inexplicable reason?

 

[Anubis1311]

10 minutes ago, vlaiv said:

Why is your Jupiter gigantic in that image?

It is 950px in diameter, but your telescope is only capable of recording it to resolution of 119px at the moment (48.8" diameter, 0.41"/px).

This is what your image should look like:

And when you present it properly - it really looks like a excellent capture, but question is - why is it so large in your processing? What sort of barlow did you use and did you drizzle for some inexplicable reason?

 

Hmm not sure I follow. I did enlarge it in PS from 72dpi to 300dpi as it was a little pixelated. Enlarging it seems to depixelate it. Is that what’s done it?

I use my phone to upload to this forum so didn’t even know the image was large! 
 

Definitely didn’t drizzle. Not actually sure what that even does so leave it at the default setting which is no drizzle.

 

Edited August 12, 2021 by Anubis1311

[vlaiv]

23 minutes ago, Anubis1311 said:

Hmm not sure I follow. I did enlarge it in PS from 72dpi to 300dpi as it was a little pixelated. Enlarging it seems to depixelate it. Is that what’s done it?

I use my phone to upload to this forum so didn’t even know the image was large! 
 

Definitely didn’t drizzle. Not actually sure what that even does so leave it at the default setting which is no drizzle.

 

Not sure even where to start. I don't want to additionally confuse you, and in part - I'm also confused with some of the things you said.

When you say that image is pixelated, what exactly do you mean?

I often hear this - but in reality, "pixelation" of the image has nothing to do with image itself - it has to do with the way image is displayed and depends on device and software that displays the image.

On the other hand - if you do too much wavelet sharpening and there are some stacking artifacts - either because of low bit depth or due to compression or what not - you'll get effect similar to pixelation that one might confuse with "actual pixelation".

In any case, what I'm trying to say is following:

- Planetary images are best captured at critical sampling rate / critical resolution. This puts an limit to the size of planets in pixels that depends on aperture size of your telescope.

If you "decide" that you want larger image of planet - enlarging it will not bring any additional detail. Also, technique used for enlarging image is the thing that is responsible for "pixelation" and not the image itself.

I'm against software enlargement, and I let people examining the image to decide if they want to zoom in a bit more - that depends on their display and their vision. Image should be captured and presented at resolution that supports the level of detail in the image.

Here, look at image that you posted at 100% level versus same image that has been rescaled to critical sampling resolution:

Not sure what these look like at device that you are using to view the on, but I'm using computer screen with 96dpi at 70cm away - which means 1px is roughly 1.3' - or just slightly above 20/20 vision resolution which is one arc minute.

To me - bottom image is properly sampled - or shows no pixelation, and shows detail that is "compatible" to this resolution. If I start zooming it in - I'll see no more detail, and if I move a bit back and start to look at it from 2m away - I will be limited by sharpness of my eyesight - I'll start to loose detail.

Above image is just blurred rescaled version of bottom image - it does not provide more detail and if I zoom in bottom version to 800% (because they mismatch at exactly x8 in resolution), I'll get the same image as you above in terms of detail:

Yes, this is simply enlarged bottom image. But why enlarge it in software - we should let every user adjust the image to match their display device, and only make sure that we match level of detail in image to resolution / sampling rate of the image.

By the way, if you want to see level of detail that is better suited to this scale, then it would be this:

Does any of this make sense?

[Anubis1311]

5 minutes ago, vlaiv said:

Not sure even where to start. I don't want to additionally confuse you, and in part - I'm also confused with some of the things you said.

When you say that image is pixelated, what exactly do you mean?

I often hear this - but in reality, "pixelation" of the image has nothing to do with image itself - it has to do with the way image is displayed and depends on device and software that displays the image.

On the other hand - if you do too much wavelet sharpening and there are some stacking artifacts - either because of low bit depth or due to compression or what not - you'll get effect similar to pixelation that one might confuse with "actual pixelation".

In any case, what I'm trying to say is following:

- Planetary images are best captured at critical sampling rate / critical resolution. This puts an limit to the size of planets in pixels that depends on aperture size of your telescope.

If you "decide" that you want larger image of planet - enlarging it will not bring any additional detail. Also, technique used for enlarging image is the thing that is responsible for "pixelation" and not the image itself.

I'm against software enlargement, and I let people examining the image to decide if they want to zoom in a bit more - that depends on their display and their vision. Image should be captured and presented at resolution that supports the level of detail in the image.

Here, look at image that you posted at 100% level versus same image that has been rescaled to critical sampling resolution:

Not sure what these look like at device that you are using to view the on, but I'm using computer screen with 96dpi at 70cm away - which means 1px is roughly 1.3' - or just slightly above 20/20 vision resolution which is one arc minute.

To me - bottom image is properly sampled - or shows no pixelation, and shows detail that is "compatible" to this resolution. If I start zooming it in - I'll see no more detail, and if I move a bit back and start to look at it from 2m away - I will be limited by sharpness of my eyesight - I'll start to loose detail.

Above image is just blurred rescaled version of bottom image - it does not provide more detail and if I zoom in bottom version to 800% (because they mismatch at exactly x8 in resolution), I'll get the same image as you above in terms of detail:

Yes, this is simply enlarged bottom image. But why enlarge it in software - we should let every user adjust the image to match their display device, and only make sure that we match level of detail in image to resolution / sampling rate of the image.

By the way, if you want to see level of detail that is better suited to this scale, then it would be this:

Does any of this make sense?

Hi, actually I pretty much understood the most part of that. The reason I changed the resolution from 72dpi to 300dpi is that when viewing in photoshop the edges of the planet were like squares.
 

Almost as if you were trying to make a circle from hundreds of playing cards. If you zoom out it looks round but when you look closely at the edges you still see the playing cards. hope that makes sense 😂

Might be due to the sharpening?! I’ll have a look at it again in without enlarging it and see if it still looks that way. 
 

definitely wasn’t to try and make my image bigger though. It was literally to tidy the edges. It seems to have a cleaner edge at higher resolution but as you say that may be software related.

 

thanks again for your input 👍🏻

 

[vlaiv]

2 minutes ago, Anubis1311 said:

Hi, actually I pretty much understood the most part of that. The reason I changed the resolution from 72dpi to 300dpi is that when viewing in photoshop the edges of the planet were like squares.
 

Almost as if you were trying to make a circle from hundreds of playing cards. If you zoom out it looks round but when you look closely at the edges you still see the playing cards. hope that makes sense 😂

Might be due to the sharpening?! I’ll have a look at it again in without enlarging it and see if it still looks that way. 
 

definitely wasn’t to try and make my image bigger though. It was literally to tidy the edges. It seems to have a cleaner edge at higher resolution but as you say that may be software related.

 

thanks again for your input 👍🏻

 

Yes - that is classic case of "pixelation" - but there is no pixelation at all.

When you enlarge image beyond 100%, software needs to "make up" missing pixels - as there are more screen pixels than pixel in the image (image gets stretched and we need to put something in those "holes"). There are different ways to do it and it is referred as scaling interpolation (in this case up scaling as we are making image larger than original - opposed to down scaling which is making it smaller).

Choice of algorithm used for this is what creates pixelation - or rather one particular algorithm called nearest neighbor resampling is guilty of pixelation and idea that pixels are small squares. Pixels are not small squares! Neither image / mathematical pixels (which are just points without dimension) - nor camera pixels which are often little circles or rounded up squares rather than prefect squares:

This is what color pixels (and underlying circuitry) looks like under microscope.

Here, look at this:

That is small image (one that I rescaled and posted above) - zoomed in to 1600% - using nearest neighbor algorithm. This is IrfanView software for viewing images on Windows. It has option to turn on/off proper interpolation when zooming in:

Look at this, same image, same zoom level - but this option turned on:

No more "pixels"

It is due to zoom level and the way software handles zoom.

Photoshop seems not to have any sensible interpolation when zooming in - look at this answer:

https://community.adobe.com/t5/photoshop-ecosystem/visual-error-erro-visual/m-p/10837190

to quote:

Quote

Photoshop Zooming is done for fast performance not best image quality. At some zoom percentages image quality is quite poor.  You are only viewing you actual image when you view zoomed 100% your document actual pixels.  At any other zoom percentages the image you are viewing is a quickly scaled version of you actual image.  Interpolating when zooming would impact rendering time.

 

[Anubis1311]

11 minutes ago, vlaiv said:

Yes - that is classic case of "pixelation" - but there is no pixelation at all.

When you enlarge image beyond 100%, software needs to "make up" missing pixels - as there are more screen pixels than pixel in the image (image gets stretched and we need to put something in those "holes"). There are different ways to do it and it is referred as scaling interpolation (in this case up scaling as we are making image larger than original - opposed to down scaling which is making it smaller).

Choice of algorithm used for this is what creates pixelation - or rather one particular algorithm called nearest neighbor resampling is guilty of pixelation and idea that pixels are small squares. Pixels are not small squares! Neither image / mathematical pixels (which are just points without dimension) - nor camera pixels which are often little circles or rounded up squares rather than prefect squares:

This is what color pixels (and underlying circuitry) looks like under microscope.

Here, look at this:

That is small image (one that I rescaled and posted above) - zoomed in to 1600% - using nearest neighbor algorithm. This is IrfanView software for viewing images on Windows. It has option to turn on/off proper interpolation when zooming in:

Look at this, same image, same zoom level - but this option turned on:

No more "pixels"

It is due to zoom level and the way software handles zoom.

Photoshop seems not to have any sensible interpolation when zooming in - look at this answer:

https://community.adobe.com/t5/photoshop-ecosystem/visual-error-erro-visual/m-p/10837190

to quote:

 

That makes a lot of sense. So if I’d have continued editing at its native resolution the final image wouldn’t actually be pixelated even though it appeared so in photoshop! 
 

so I’m hoping to shoot Jupiter again tonight as my ADC came today. Hopefully I’ll have more data to play with tomorrow and I’ll leave it at its native resolution.

thank you so much! I’m learning loads and I appreciate the effort you’re going through to help me. 

[CraigT82]

15 minutes ago, vlaiv said:

Yes - that is classic case of "pixelation" - but there is no pixelation at all.

Great Post 👍

Never seen pixels under a microscope before, they don't look like I expected 

[powerlord]

to be fair, I think you could be confusing the op @vlaiv ? - your 96 dpi is fine for monitors, but can be tiny on a mobile screen for example - hell, a 100 pixel horizontal image would be miniscule viewed 100% on my huwaei p30 pro for example.

Now, mobile view on this forum is actually good, so if you view this thread on my phone, both original 96 and 300dpi image are EXACTLY the same size on it,  but not all apps work that way.

And many native apps, especially ones on windoze do a terrible job of resizing - so if you post someone a pic of your wonderful 100x100 pixel image - see it as tiny on your 1920x1080 screen, and use that app to zoom in, you very often get a pixelated mess, since all it does is a basic basic nearest neighbour resize.

Another good reason to enlarge is for printing, since software resizing, especially the latest algorithms in affinity and ps do a damn site better job than a printer driver will do.

 

 

[vlaiv]

6 minutes ago, CraigT82 said:

Great Post 👍

Never seen pixels under a microscope before, they don't look like I expected 

Yes, it is interesting, isn't it. It depends on what type of sensor it is - if it has micro lens or not and is it bayer matrix or not. There are few examples that google image search returns:

This one is particularly interesting as it shows manufacturing defects as well:

Some of those pixels will have lower QE. Another reason to use flats - even for planetary.

Computer screen pixels are interesting as well - also not something that can be easily guessed:

[vlaiv]

1 minute ago, powerlord said:

to be fair, I think you could be confusing the op @vlaiv ? - your 96 dpi is fine for monitors, but can be tiny on a mobile screen for example - hell, a 100 pixel horizontal image would be miniscule viewed 100% on my huwaei p30 pro for example.

Now, mobile view on this forum is actually good, so if you view this thread on my phone, both original 96 and 300dpi image are EXACTLY the same size on it,  but not all apps work that way.

And many native apps, especially ones on windoze do a terrible job of resizing - so if you post someone a pic of your wonderful 100x100 pixel image - see it as tiny on your 1920x1080 screen, and use that app to zoom in, you very often get a pixelated mess, since all it does is a basic basic nearest neighbour resize.

Another good reason to enlarge is for printing, since software resizing, especially the latest algorithms in affinity and ps do a damn site better job than a printer driver will do.

 

 

Yes, I was just thinking of that - "second part of equation".

We can control how we record and process image, but we can't control how it is being displayed / viewed.

Computers screens are matched rather good to human vision and viewing conditions. Most of us view computer screens at reasonable distance and 96dpi matches that.

Mobile devices on the other hand are somewhat "problematic". Mega pixel craze is apparently working in that area as well. Most of modern phones have display densities of around ~400dpi (or ppi to be precise in display terminology) with tendency of increasing.

That is simply too high to be match to ordinary people. If ~100ppi is sensibly matched to human vision at ~70cm, well then ~400ppi would require 70/4 = 17.5cm distance.

Most of us don't hold phone screen at 17.5cm distance to our eyes. In fact most people past age of 40 can't even focus that close without reading glasses.

 

[powerlord]

you need to meet my neices... 🙄

[vlaiv]

2 minutes ago, powerlord said:

you need to meet my neices... 🙄

Apparently up to age of 20 - even 10cm is fine, but I don't seem to remember holding anything that close to my face when I was that young - well with exception of threading a needle, yep that was done on such short distances

[AstroMuni]

2 minutes ago, vlaiv said:

I don't seem to remember holding anything that close to my face when I was that young

You didnt have phones in those days I am guessing 😉

[powerlord]

I clearly remember being able to focus to about 4-5cm  when I was 5 or 6. Now it's about 20cm without my contacts in. With them in (short sighted), I'm the one with my arms outstretched to read a book!

age is cruel 😞

 

[vlaiv]

4 minutes ago, AstroMuni said:

You didnt have phones in those days I am guessing 😉

We had the next best thing