102 F/7 or 80 F/7 for White light?

[Stu]

31 minutes ago, vlaiv said:

Well, here is simulation so we can see what it looks like. I took this high resolution image of granulation (best I could find and in fact it's titled: "Highest_resolution_photo_of_Sun_(NSF)_as_of_January_20,_2020"):

this is actually scaled down image - original is huge 7320px x 7320px - which is plenty of resolution for simulation, individual cells are huge and nicely resolved:

I took two perfect 80 and 100 mm scopes, added Baader Solar Continuum filter (540nm simulation) and produced respective images. In reality, view is going to be worse than this because of imperfect optics and seeing effects:

80mm:

100mm:

Both of these squares represent full image posted above (first one scaled down to fit the screen).

Although you can see the granulation texture - you really can't resolve single cells in either 80 or 100mm scope. Most cells are visually simply joined into larger blob that we see and think is individual cell - but it's not.

 

That does not represent accurately what I see under excellent conditions with my Tak, which are much sharper and clearer. No offence meant, but I’m happy for you to remain in your theoretical world, I will continue to enjoy the real views.

[Stu]

1 hour ago, michael.h.f.wilkinson said:

What I mean is that in an 80 mm scope, the smallest convection cells are not necessarily accurately imaged, because they are blurred by a point-spread function that somewhat bigger than their extent. I can distinguish the individual cells, and the bigger ones definitely stand out clearly. A four inch scope would show this a lot better. The problem with language is that the world "resolved" can be interpreted differently. Are the individual grains resolved by an 80 mm scope? According to the Rayleigh criterion they are. Can we accurately resolve the sizes and shapes of 2" grains with a 2" radius of the Airy disc? Not really.

I understand what you are saying, but I can see the surrounding dark cell walls and the bright inner parts to the cells. This is at x200. What am I seeing? Much like Vlaiv’s first image above but much smaller scale. It is definitely not like the second green image which is just blurry and indistinct. The reality is much sharper and more striking.

[michael.h.f.wilkinson]

35 minutes ago, Stu said:

I understand what you are saying, but I can see the surrounding dark cell walls and the bright inner parts to the cells. This is at x200. What am I seeing? Much like Vlaiv’s first image above but much smaller scale. It is definitely not like the second green image which is just blurry and indistinct. The reality is much sharper and more striking.

I also see a sharper image than what vlaiv posted as representative of the view of an 80mm. The main issue here is that I am not sure his image reproduces the dynamic range of the live image that well. The input image has been processed, and might not represent contrast faithfully. Higher contrast always results in higher apparent resolution.

[vlaiv]

6 minutes ago, michael.h.f.wilkinson said:

I also see a sharper image than what vlaiv posted as representative of the view of an 80mm. The main issue here is that I am not sure his image reproduces the dynamic range of the live image that well. The input image has been processed, and might not represent contrast faithfully. Higher contrast always results in higher apparent resolution.

Do keep in mind that image was produced at 540nm - or rather simulated at that light - as Baader Continuum filter is centered at that wavelength.

At 540nm, critical sampling frequency for 80mm scope is 0.7"/px - that means that typical cell is represented with only 3 pixels. In those three pixels across one needs to put both walls of the cell and interior. I just don't see how can we resolve typical cell at that wavelength to clearly see interior and walls.

Here is what 0.7"/px looks like regardless of resolution - and that is sharper than the optics will produce:

Here we have comparison of original image and one sampled at 0.7"/px and then enlarged (Lanczos resampling) to match the size of right image.

Even without effects of aperture - we can't really say that we see cells and walls. Add to that blurring by telescope optics (perfect one - better even than the all mighty Tak ) and you'll get this:

Here is full procedure for creating the image:

I took original Jpeg image, converted it to 32bit / monochromatic version, did inverse gamma of 2.2 to bring it to linear (hopefully whoever made an image followed sRGB standard), and measured random cell size. It measured 270px across so we now have 1/135"/px base resolution for our base image.

I generated Airy disk pattern and calculated critical sampling rate for 540nm and convolved linear image with airy disk and scaled it to critical sampling rate (no need to make it bigger - but we can always enlarge it if needed without loss of detail as there is no detail past critical sampling rate).

I applied green LUT to further simulate 540nm filter and I did forward gamma of 2.2. Image converted to 8bit and saved as jpeg (high quality settings).

@Stu

Please don't think that I'm negating your experience. There are several things that could be wrong with my simulation. I could have made wrong measurement of a cell - took particularly large cell that is actually 3" instead of 2" in size (although I did aim for average one and measured it in shorter direction just to be sure). I could further apply theory in a wrong way - or simply made error in calculation.

That is why it is important to have repeatable results when doing simulations - similarly to experiments. I would be much happier if someone else also did simulation so we can compare results.

There is also already mentioned question of dynamic range / contrast and monitor calibration, pixel size and viewing distance.

[michael.h.f.wilkinson]

3 minutes ago, vlaiv said:

Do keep in mind that image was produced at 540nm - or rather simulated at that light - as Baader Continuum filter is centered at that wavelength.

At 540nm, critical sampling frequency for 80mm scope is 0.7"/px - that means that typical cell is represented with only 3 pixels. In those three pixels across one needs to put both walls of the cell and interior. I just don't see how can we resolve typical cell at that wavelength to clearly see interior and walls.

Here is what 0.7"/px looks like regardless of resolution - and that is sharper than the optics will produce:

Here we have comparison of original image and one sampled at 0.7"/px and then enlarged (Lanczos resampling) to match the size of right image.

Even without effects of aperture - we can't really say that we see cells and walls. Add to that blurring by telescope optics (perfect one - better even than the all mighty Tak ) and you'll get this:

Here is full procedure for creating the image:

I took original Jpeg image, converted it to 32bit / monochromatic version, did inverse gamma of 2.2 to bring it to linear (hopefully whoever made an image followed sRGB standard), and measured random cell size. It measured 270px across so we now have 1/135"/px base resolution for our base image.

I generated Airy disk pattern and calculated critical sampling rate for 540nm and convolved linear image with airy disk and scaled it to critical sampling rate (no need to make it bigger - but we can always enlarge it if needed without loss of detail as there is no detail past critical sampling rate).

I applied green LUT to further simulate 540nm filter and I did forward gamma of 2.2. Image converted to 8bit and saved as jpeg (high quality settings).

@Stu

Please don't think that I'm negating your experience. There are several things that could be wrong with my simulation. I could have made wrong measurement of a cell - took particularly large cell that is actually 3" instead of 2" in size (although I did aim for average one and measured it in shorter direction just to be sure). I could further apply theory in a wrong way - or simply made error in calculation.

That is why it is important to have repeatable results when doing simulations - similarly to experiments. I would be much happier if someone else also did simulation so we can compare results.

There is also already mentioned question of dynamic range / contrast and monitor calibration, pixel size and viewing distance.

But then I mage at roughly 0.5" per pixel and get this

Which is pretty close to the texture I see in the image I took. (I added blur, incidentally)

 

[vlaiv]

6 minutes ago, michael.h.f.wilkinson said:

But then I mage at roughly 0.5" per pixel and get this

Is that sharpened version?

If I take the image you posted from the beginning of the thread and compare that to 80mm scope enlarged to same resolution, we get this:

vs

Mind you, second one is not wavelet sharpened / deconvolved and is simulated at 540 and not full spectrum.

That looks pretty similar to me.

 

Edited June 15, 2021 by vlaiv
Had another copy of image

[michael.h.f.wilkinson]

1 minute ago, vlaiv said:

Is that sharpened version?

If I take the image you posted from the beginning of the thread and compare that to 80mm scope enlarged to same resolution, we get this:

vs

Mind you, second one is not wavelet sharpened / deconvolved and is simulated at 540 and not full spectrum.

That looks pretty similar to me.

 

 

No sharpening applied, just blurring and resampling of the high-res image. That was probably contrast stretched. The dark colour you added to the simulated image muddies the comparison. My original image was taken at 540 nm, slightly deconvolved and sharpened, but no real contrast stretch. I will run our connected granulometry software on this and see if the morphological pattern spectra are similar, that should allow assessment of the similarity of the textures in a quantitative way.

I think we agree that the granulation is visible at 80 mm aperture, but that the shapes of the grains themselves are not well resolved.

[vlaiv]

22 minutes ago, michael.h.f.wilkinson said:

I think we agree that the granulation is visible at 80 mm aperture, but that the shapes of the grains themselves are not well resolved.

Indeed.

I think that OP has enough material to at least somewhat asses differences between 100mm and 80mm scopes for WL observation - both simulated and first hand accounts.

[Stu]

2 hours ago, vlaiv said:

Even without effects of aperture - we can't really say that we see cells and walls. Add to that blurring by telescope optics (perfect one - better even than the all mighty Tak ) and you'll get this:

 

2 hours ago, vlaiv said:

My trouble is that this doesn’t represent what I see, and as said, I don’t agree that you can’t see cells and walls in a 100mm scope; I do see this. When conditions are excellent then high power can be used and you do see the detail of the cells. Are they tiny? Yes! But can you see them? Yes! They are sharper, brighter, more vivid although smaller than your simulation.

The way I see it is that the Double Double stars are 2.2” and 2.4” separation and are a trivial split at high power, so being able to resolve detail at this kind of level is quite possible.

I checked back on some old posts and @Merlin66 whose opinion I value a lot has said that 4” and above will show granulation properly at high power. I used to be of the mistaken opinion that I was seeing granulation at low power, but I now understand that this is more ‘macro granulation’, not the cells themselves which need high power to show themselves.

Seeing is king though, and whilst I see granulation on most sessions, it is relatively infrequent that I get really cracking seeing and high powers are stable enough for the really ‘wow’ views. Today for example is really quite muted.

I’ve just done a quick experiment by putting my Baader CoolWedge onto my Telementor to see how it compares. The conditions are not great so I need a better day to do it properly, but in simple terms the 63mm showed the basic umbra/penumbra and some relatively faint faculae. Only very subtle hints of any surface texture, barely anything really. The 100mm showed some detail and structure in the Sunspot, granulation  around the spot and the faculae were much easier to see and more detail. There was some evidence of granulation over the main rest of the surface but very muted compared with times of good seeing. I’ll try this again another time.

Sorry to bang on about this, I guess I get a little ‘challenged’ by being told I can’t see something because of theory, vs the real world experience of what is practically achievable. I’ve spent a long time trying different kit combinations and now I have found something that does work very well, so I know I’m getting maximum quality and am delighted with the views. Of course, a larger aperture may show more, but would be used less, although I generally found that the Tak showed me as much as a 120ED.

…… and breathe 😉😉🤣🤣

[vlaiv]

8 minutes ago, Stu said:

Sorry to bang on about this, I guess I get a little ‘challenged’ by being told I can’t see something because of theory, vs the real world experience of what is practically achievable.

Don't worry, I perfectly understand where you are coming from and I do believe that you are seeing what you are seeing.

This is not the first time people say that they are seeing more at the eyepiece than theory predicts should be possible. Well - not theory, but rather my interpretation of theory / my simulation. I have no doubt that theory is correct (otherwise it would not be scientific theory) - and it is often our misuse of it that makes bad predictions.

In any case, if practice and theory/simulation strongly diverge, then there are couple of possible explanations.

- My understanding of theory is incomplete / my application of theory is outside of its domain of validity or I simply made a mistake applying the theory (wrong simulation parameters or just error in simulation process).

- There are additional parameters that we neglected / omitted that are significant enough to make the difference

- There are other factors that result in perception of what might not be physical reality. When I say this - I don't mean that you are imagining things, I just simply mean that eye-brain system is very complex and does a lot of things "behind curtains" to enable us to see the way we see and some of that might have particular effect in this case. That is the reason we have optical illusions for example, or the reason we never see photon noise (although we should - we are sensitive enough to detect light at that level - however our brain does noise suppression).

In any case, I feel that it is beneficial to pursue this discrepancy further, and I already have couple of ideas of how to go about it (not for this particular case - but to test out resolution of actual optics on object such as Jupiter without influence of atmosphere - to see/record how does different quality optics render image).

[vlaiv]

Here is another sim

This time 10" scope capturing granulation:

Left is excellent capture by @AbsolutelyN using 10" scope and ASI178mm camera and right is simulation of granulation from above image. I'm not sure if I matched pixel scale properly since I'm missing pixel scale for both images - I just measured cells and assigned 2" to measurement. I also tried to match contrast brightness.

[AbsolutelyN]

1 hour ago, vlaiv said:

Here is another sim

This time 10" scope capturing granulation:

Left is excellent capture by @AbsolutelyN using 10" scope and ASI178mm camera and right is simulation of granulation from above image. I'm not sure if I matched pixel scale properly since I'm missing pixel scale for both images - I just measured cells and assigned 2" to measurement. I also tried to match contrast brightness.

It's actually an 8" scope - I had to stop the aperture down to the size constraint of an a4 sheet of baader solar film. Taken with a 3x barlow so 3600mm. Interesting comparison - needs good seeing though. 

[vlaiv]

45 minutes ago, AbsolutelyN said:

It's actually an 8" scope - I had to stop the aperture down to the size constraint of an a4 sheet of baader solar film. Taken with a 3x barlow so 3600mm. Interesting comparison - needs good seeing though. 

There is probably one more variable that is off - at least according to AstroBin info for that animated version - it says SII filter was used (probably to tame the seeing). That is different wavelength than 540nm used for simulation above.