3085 replies to this topic

[Psion]

This is picture from OSC camera, no filters used.

[kgb]

This is picture from OSC camera, no filters used.

UV/IR filter or a UV/IR coated protection window count as a filter.

Edited by kgb, 15 January 2024 - 04:06 PM.

[GSBass]

Vaonis is pricy… I’ve saved money where I could, used third party filters… I really don’t feel I gave up anything by doing that and it has allowed me to use them on multiple scopes… I have had no desire to buy their proprietary filters…. I’m actually in the robot hobby pretty cheap, I was an early bird so I got my vespera for just a grand, paid 165 for my l’enhance through alibba and I think I paid full price for my baader neodymium… I also spent 200 for an acratech leveler and 30 dollars for some walking stick feet to put on the standard tripod. But that’s been pretty much my entire investment, unless you count buying the dwarf and hestia

Indeed.  It’s clear that the major difference between the S50 and the Vespera line is the app.  I would be more inclined to jump on the V wagon if they would have included the filters, which tacks another $300+ on the cost. 

[GSBass]

This is my latest image from the classic, generally my stars look pretty nice https://x.com/artemi...1XgzO5JXF9NadZA

[Psion]

UV/IR filter or a UV/IR coated protection window count as a filter.

It's common for everyone to shoot with filters, especially in bad skies, but I don't imagine filters will make clear circles around stars; a good-quality filter doesn't do that. Unfortunately, the filter before the chip, especially the OIII, is always a problem, unlike the filter before the lens.

 

A note on the image. When Celestron publishes any images, they should at least do a White Balance or, ideally, a photometric colour calibration.

Attached Thumbnails

• 

[GSBass]

Here is another recent one in Taurus showing the extend of halo on bright stars with classic https://x.com/artemi...1XgzO5JXF9NadZA

[kgb]

It's common for everyone to shoot with filters, especially in bad skies, but I don't imagine filters will make clear circles around stars; a good-quality filter doesn't do that. Unfortunately, the filter before the chip, especially the OIII, is always a problem, unlike the filter before the lens.

A note on the image. When Celestron publishes any images, they should at least do a White Balance or, ideally, a photometric colour calibration.

Okay, but I was responding to your post in which you said the RASA shows CA as if it that were equivalent to the pic you showed from the Vespera. The RASA configuration, which is not fully known and is based only upon the picture you posted, clearly shows reflections around high magnitude stars whereas the Vespera clearly shows the unfocused light, characteristic of CA, around lower and higher magnitude stars.

[Starmix]

The speed of the Origin intrigues me the most. Coming from a Hyperstar system and being strictly EAA, I want the images as fast as I can get them. Also the robotic nature is important too especially when it’s cold out. 
 

I do see Vaonis has having the advantage as far as software though, which helps with their overall standing in the smart scope field. If they could produce a killer scope similar to the RASA for reasonable money, they would kill the game, obviously it would not be as portable, but you can’t have everything.

[GSBass]

It would surprise me if they did a reflector, about the most I think we will ever see from them is a 3 or 4” refractor system, one of the reasons is colmination, it seems to be very important to the ceo to keep simplicity the most important factor. It’s also the reason their competition will have much more manual control. I think it was hard for him to even allow darks etc on the pro, he wanted that all to be behind the scenes but for whatever reason they decided not to automate that process…, as a vespera owner it will feel awkward doing the extra stuff… all we have had to do was level and turn it on

The speed of the Origin intrigues me the most. Coming from a Hyperstar system and being strictly EAA, I want the images as fast as I can get them. Also the robotic nature is important too especially when it’s cold out. 
 

I do see Vaonis has having the advantage as far as software though, which helps with their overall standing in the smart scope field. If they could produce a killer scope similar to the RASA for reasonable money, they would kill the game, obviously it would not be as portable, but you can’t have everything.

[jprideaux]

It would surprise me if they did a reflector, about the most I think we will ever see from them is a 3 or 4” refractor system, one of the reasons is colmination, it seems to be very important to the ceo to keep simplicity the most important factor. It’s also the reason their competition will have much more manual control. I think it was hard for him to even allow darks etc on the pro, he wanted that all to be behind the scenes but for whatever reason they decided not to automate that process…, as a vespera owner it will feel awkward doing the extra stuff… all we have had to do was level and turn it on

Yes it has been a bit of a struggle to get Vaonis to include manual features. Their initial vision was a purely autonomous system where the user does very little and they only added manual control slowly over time. You can’t even use the Vespera/Stellina to take a daytime bird picture but you can with Dwarf2/SeeStar/Origin.

I do agree that Vaonis is the leader right now in software (in the robotic world) for imaging. I would not be surprised if the Vespera-pro even does quite well compared to the Celestron Origin for imaging. The Origin may be quite a bit better in getting images to a certain quality fast. I don’t think we have really seen the best of what the Origin can do yet. We have just seen short exposures so far.

I’m hoping both the SeeStar and Origin come out with mosaic modes soon. The Origin would be disappointing if it were always to just have its native FoV with the INX178 (similar FoV to the SeeStar).

[Psion]

I would not be surprised if the Vespera-pro even does quite well compared to the Celestron Origin for imaging. The Origin may be quite a bit better in getting images to a certain quality fast. I don’t think we have really seen the best of what the Origin can do yet. We have just seen short exposures so far.

Here is a table that shows how much signal a pixel produces on the camera, which determines the Signal-to-Noise ratio and overall image quality.

Attached Thumbnails

• 

[anurag1212]

ouch :'(

im starting to think i should cancel my pro preorder and wait for stellina v2 which vaonis will invariably announce 31 days after the pro ships. 

 

Here is a table that shows how much signal a pixel produces on the camera, which determines the Signal-to-Noise ratio and overall image quality.

[ColSanderz]

I think the graph is a bit misleading as it's a function of aperture vs pixel pitch.

 

It doesn't mean the pro will be incapable of great photos, only that, given identical gain, more time on target will be required to achieve a similar signal-to-noise ratio.

 

It will be interesting to see how Vaonis overcomes this, even to some degree. 

 

It does make me question my pro pre-order as well since I imagine primarily from bortle 8 and sometimes 5-6.

 

Oh well, we have some time to cancel. I too am hopeful for a stellina v2. It may be worth it for me to just stick it out with passengers for the time being

 

ouch :'(

im starting to think i should cancel my pro preorder and wait for stellina v2 which vaonis will invariably announce 31 days after the pro ships. 

[jprideaux]

I think the graph is a bit misleading as it's a function of aperture vs pixel pitch.

 

It doesn't mean the pro will be incapable of great photos, only that, given identical gain, more time on target will be required to achieve a similar signal-to-noise ratio.

 

It will be interesting to see how Vaonis overcomes this, even to some degree. 

 

It does make me question my pro pre-order as well since I imagine primarily from bortle 8 and sometimes 5-6.

 

Oh well, we have some time to cancel. I too am hopeful for a stellina v2. It may be worth it for me to just stick it out with passengers for the time being

The other factor not included is how well the relative scopes will do with frame-rejections.  If they are rejecting a lot of frames, then the effective "aperture vs pixel pitch" will go down.  

We already know that the Vespera/Passenger/Vespera2 can do quite well if you are able to spend enough time on the target.  I suspect that the Vespera-pro will do even better but with the F5 and even smaller pixels, you will need to spend even more time om the target for those really nice images.

[Psion]

I think the graph is a bit misleading as it's a function of aperture vs pixel pitch.

The table is a function of lens diameter, focal length, pixel size and quantum efficiency.

 

Small pixels are more sensitive to guiding and quality of optics.

[GSBass]

Yes, most of my keepers are very long exposures…. But that does seem to be the case for the advanced folks too no matter how efficient the equipment…. My exposures are short compared to theirs 

The other factor not included is how well the relative scopes will do with frame-rejections.  If they are rejecting a lot of frames, then the effective "aperture vs pixel pitch" will go down.  

We already know that the Vespera/Passenger/Vespera2 can do quite well if you are able to spend enough time on the target.  I suspect that the Vespera-pro will do even better but with the F5 and even smaller pixels, you will need to spend even more time om the target for those really nice images.

[ColSanderz]

Was simplifying my explanation a bit Point being that for the Pro, signal-to-noise will be its biggest "enemy" with regards to exposure time. 

 

Since I live in a pretty light poluted area, have to shoot for hours on end to get something I'm happy with. Someday I'll convince my better half to move further out lol.

The table is a function of lens diameter, focal length, pixel size and quantum efficiency.

 

Small pixels are more sensitive to guiding and quality of optics.

[WillR]

It's hard to compare, but when I process the Vespera image, I have to mask the blue halo around the bright stars a lot. Here's an unedited crop of two images from about 10 minutes of exposure on both instruments.

Thanks for posting that. The Vespera has a bit less noise, but the Seestar does better on dim stars. But most people wouldn’t see the difference, certainly on the small screen of a phone.

 

Nothing to justify the price difference (for me), especially for EAA. And the extreme portability of the Seestar is a big plus. This topic on the Origin may convince me to get the Seestar.

 

These are cropped images. Horsehead? What are the uncropped FOVs?

Edited by WillR, 16 January 2024 - 10:14 AM.

[Psion]

Understand  Of course, a lot depends on the image processing, i.e. the software. I think that eVscope, with proper image processing, should do much more than it shows.

[smiller]

Here is a table that shows how much signal a pixel produces on the camera, which determines the Signal-to-Noise ratio and overall image quality.

I’m afraid this simply isn’t correct.   SNR of an image not just determine by the light flux on a single pixel.  This would only be true if your system was completely dominated by read noise.   This may have been true 20 years ago in the early days of CCD cameras, but it is not at all true today.   And even then, it would not have been fully true and certainly not true if you were in light polluted skies.

 

With the modern CMOS cameras in these systems the read noise is so low that exposures under 10 seconds is possible with read noise contributing only a small percent to the total noise stack, with the vast majority still being the sky background noise.

 

The fact that your spreadsheet dosn’t have a dependency on the sky background light flux, the actual gain the camera is using (which strongly determines read noise), nor the filter being used makes your table incomplete as some of these factors dominate the calculation.

 

In order to calculate the fraction of noise contributed to a stack by the light flux on the pixel, you need ALL of the following:

 

1) Pixel size (which I assume you included)

2) Telescope f-ratio (which I assume you included)

3) Camera read noise (depends on camera and gain setting). (This is a big factor)

4) Filter used: Broadband or narrowband filter bandpass. (This is a huge factor)

5) Sky background photon flux (This is a huge factor)

6) Total light gathering power of the system:  this is a combination of aperture, F-ratio and sensor area and this is a huge factor

7j And a small factor: Camera quantum efficiency (Small because these cameras are similar and this can be ignored)

 

If you want to learn more, watch this video and use Dr. Robin Glover’s tables and if you make some assumptions about the gain the telescopes are using and the sky conditions, you can arrive at some correct numbers:
 

https://youtu.be/3RH...3W3B9duVonGzcUl

 

As I’ve done a fair number of read noise contribution calculations for myself and others, I’ll tell you right now that unless you are in very dark skies (Bortle 4 or better) that for broadband targets read noise is likely to be a fairly small contributor to the overall noise stack for all of these telescopes.   But in dark skies or with a dual narrowband filter, there will probably be some material differences, but tiny compared to the differences you show.

Edited by smiller, 16 January 2024 - 10:41 AM.

[tarbat]

If you want to learn more, watch this video and use Dr. Robin Glover’s tables and if you make some assumptions about the gain the telescopes are using and the sky conditions, you can arrive at some correct numbers:

I've already done these calculations for the Seestar, for two usage cases:

• UV/IR filter
• Dual-band Filter

 

This shows that without the dual-band filter, to hit the 5% mark recommended by Robin, only Bortle 4 or better will benefit from a 20s or 30s sub-exposure over the standard 10s sub-exposure.  With the dual-band filter, estimated at 50nm, you can hit that 5% mark at Bortle 6 or better with a 30s exposure. Everything then depends on the accuracy of tracking over a 30s sub.

[Psion]

I'll explain. This is a simple comparison of how many photons hit a given area of the chip, in this case, a pixel, and how much of that the camera generates in signal at a given quantum efficiency. Here, read noise is assumed to be zero and dark currents are also assumed to be zero. Of course, other factors come into play, such as the use of filters, photon sky noise, binning, etc. The calculation is certainly not intended to be an exact result, it is a simple comparison. However, for example, Vespera PRO cannot, in principle, have a better nebulosity signal per pixel in the same unit time as Vespera Classic or Passengers.

[smiller]

I'll explain. This is a simple comparison of how many photons hit a given area of the chip, in this case, a pixel, and how much of that the camera generates in signal at a given quantum efficiency. Here, read noise is assumed to be zero and dark currents are also assumed to be zero. Of course, other factors come into play, such as the use of filters, photon sky noise, binning, etc. The calculation is certainly not intended to be an exact result, it is a simple comparison. However, for example, Vespera PRO cannot, in principle, have a better nebulosity signal per pixel in the same unit time as Vespera Classic or Passengers.

The problem is that this is not only not a precise result. It’s not a useful result, because the large factors are not being considered. You’re only considering the generally unimportant factors.

 

In an ideal world where there’s zero read noise, it doesn’t matter at all how big your pixels are or how much light hits each pixel because each pixel is now a perfect photon counter, and it doesn’t matter how you divide them all up… the overall image SNR result is exactly the same because you are down to sky background noise and target shot noise.  So when in your simplify example, the only telescope factor in SNR per unit of sky captured is down to aperture and, for narrowband targets, filter bandpass.

 

watch that Robin Glover video and you may then understand.  
 

Tarbat gets the factors right in his post above and he shows how it’s generally a small factor.  Notice how Tarbat needs to include all sorts of factors, such as assumptions about the camera gain (the read noise), assumptions about what filter is used and the bandpass of the filter, assumptions about how dark the sky is, assumptions about the exposure time.  These are critical!
 

It would be useful for others to spot check Tarbat’s results but his process is spot on.

 

What Tarbat can do is calculate his results for each telescope and then you will know how they differ in read noise contribution for all of the situations he has tables and graphs for.   Then we'll get an idea of how these scopes differ in this regard although we'll need one more thing:  How the aperture, f-ratio, and sensor size impacts total photon gathering power, which shows how fast SNR improves over time and how much of the sky is captured at once.   This is where an large aperture F2.2 scope will really start to shine (especially if it didn't have such a teeny tiny sensor).

Edited by smiller, 16 January 2024 - 10:56 AM.

[Psion]

So when in your simplify example, the only factor in SNR per unit of sky captured is aperture.

 

Once more. The table is a function of lens diameter, focal length, pixel size and quantum efficiency.

 

Please post errors or comments or whatever on the cloudynights.com thread:
https://www.cloudyni...due-calculator/

Edited by Psion, 16 January 2024 - 10:50 AM.

[smiller]

Once more. The table is a function of lens diameter, focal length, pixel size and quantum efficiency.

 

Please post errors or comments or whatever on the cloudynights.com thread:
https://www.cloudyni...due-calculator/

I understand that you calculated an optical factor of the system and that was likely correct, so no need to correct this fundamental calculation.  The issue was your statement "which determines the Signal-to-Noise ratio and overall image quality.".    This is not correct unless camera read noise is totally dominating the equation.    If you just said, I'm calculating etendue and left it at that, with no broad implications of what it meant, it would have been fine.