ASI 1600 vs ASI 2600

[Rodd]

10 minutes ago, vlaiv said:

It works for any signal and noise - no matter how bright the sky is

Not true--if there is no signal to pick up becuase the target is fainter than the sky--teh camera cant determine what is signal and what is noise.  That is why no matter low long you image a galaxy 13 billion light years away with a 4" refractor, you will not capture it--and I refer to a cluster of galaxies as a whole so the target is not beneath the dawes limit of teh scope.   I refer mostly to broad band.  Go to a bortle 8 sky and try to image a very very faint reflection nebula--maybe the IFN with a lum filter.  I can't even take a 30 sec lum sub in my sky becuase it turns out bright gray (not quite white but pretty bright

Edited December 22, 2021 by Rodd

[vlaiv]

6 minutes ago, Rodd said:

Not true--if there is no signal to pick up becuase the target is fainter than the sky--teh camera cant determine what is signal and what is noise.  That is why no matter low long you image a galaxy 13 billion light years away with a 4" refractor, you will not capture it--and I refer to a cluster of galaxies as a whole so the target is not beneath the dawes limit of teh scope.   I refer mostly to broad band.  Go to a bortle 8 sky and try to image a very very faint reflection nebula--maybe the IFN with a lum filter.  I can't even take a 30 sec lum sub in my sky becuase it turns out bright gray (not quite white but pretty bright

No point in arguing "true / not true" - light is additive in nature.

It does not matter if target is fainter then the sky - in fact - almost all targets that we image are fainter then the sky. Sky has brightness of 22 mag / arc second squared even at darkest locations on the earth.

We image targets that are fainter than that regularly.

Say your sky produces 100000 photons and your target produces 1 photon - that is 100001 photons captured by camera. That one photon won't be lost - it will be captured. Even if there is less than one photon per exposure - like 0.1 photon per exposure. That means that every tenth exposure (on average) will capture that one photon - when you average those exposures - you'll get 1/10 (one out of ten) = 0.1

again - given enough time, you can image however faint target you want, even from bortle 8 sky. Problem is that most people don't image for 1000+ hours - in fact many don't image form more than 10 hours - but you can pick up IFN from light polluted skies as well.

 

[david_taurus83]

Regarding the proposed 91% QE. Isn't this peak around the green/OIII wavelength? The more important Ha QE is more like 60%? I think this point gets lost in all the hype.

 

[Xilman]

34 minutes ago, Rodd said:

I dont agree with this--if the signal is below the sky fog limit, no amount of doubling will increase signal.  This is only true for dim targets that are below teh sky fog threshold

I disagree.

Let us assume that the signal is 1% of the background and the noise is proportional to sqrt(I), where I is the detector response in photons at each pixel. Suppose for simplicity that all the signal appears in a single pixel and all the surrounding pixels detect only background..

If I = 101 at the position of a signal,, the background is 100 and so the signal is 1.  The noise in the background is 10 and that in the signal-detecting pixel is 10.05.  The signal is swamped by the noise by a factor of 200 and so is indetectable.

If I=1,010, 000 the noise in the signal is sqrt(1010000) = 1005. Subtract the background as estimated from the neighbouring pixels and get the result 10,000 plus/minus RMS (1000+1005).  The noise from the signal is sqrt ((1005^2+1000^2)/2) = 1002.5, giving a signal detection with a SNR of 10000/1002.5 = 9.97.

Detecting objects which are 1000 times fainter than the sky is routine. I've managed it for a factor of 100 without any difficulty whatsoever. I can provide images on request.

All this assumes you know nothing about the signal and its time-dependent behaviour, otherwise you can do much better. If you don't believe me, take a look at measurements of sea-level rises which can be measured to better than 1mm accuracy, despite 10m waves superimposed on 10m high tides.

 

Edited December 22, 2021 by Xilman
CLARIFICATION

[Rodd]

15 minutes ago, vlaiv said:

No point in arguing "true / not true" - light is additive in nature.

It does not matter if target is fainter then the sky - in fact - almost all targets that we image are fainter then the sky. Sky has brightness of 22 mag / arc second squared even at darkest locations on the earth.

We image targets that are fainter than that regularly.

Say your sky produces 100000 photons and your target produces 1 photon - that is 100001 photons captured by camera. That one photon won't be lost - it will be captured. Even if there is less than one photon per exposure - like 0.1 photon per exposure. That means that every tenth exposure (on average) will capture that one photon - when you average those exposures - you'll get 1/10 (one out of ten) = 0.1

again - given enough time, you can image however faint target you want, even from bortle 8 sky. Problem is that most people don't image for 1000+ hours - in fact many don't image form more than 10 hours - but you can pick up IFN from light polluted skies as well.

 

You are right--I am not teh one to argue.  But Craig Stark is and I got what I am saying from reading one of his articles.  In it he said that if teh signal is below teh sky fog limit, the camera will not recognize it as signal.  LP is not noise.  It is signal.  Bad signal.  But signal.  

[Rodd]

10 minutes ago, Xilman said:

I disagree.

Let us assume that the signal is 1% of the background and the noise is proportional to sqrt(I), where I is the detector response in photons at each pixel. Suppose for simplicity that all the signal appears in a single pixel and all the surrounding pixels detect only background..

If I = 101 at the position of a signal,, the background is 100 and so the signal is 1.  The noise in the background is 10 and that in the signal-detecting pixel is 10.05.  The signal is swamped by the noise by a factor of 200 and so is indetectable.

If I=1,010, 000 the noise in the background is sqrt(1010000) = 1005. Subtract the background as estimated from the neighbouring pixels and get the result 10,000 plus/minus 1005.  The noise from the signal is sqrt ((1005^2+1000^2)/2) = 1002.5, giving a signal detection with a SNR of 10000/1002.5 = 9.97.

Detecting objects which are 1000 times fainter than the sky is routine. I've managed it for a factor of 100 without any difficulty whatsoever. I can provide images on request.

All this assumes you know nothing about the signal and its time-dependent behaviour, otherwise you can do much better. If you don't believe me, take a look at measurements of sea-level rises which can be measured to better than 1mm accuracy, despite 10m waves superimposed on 10m high tides.

 

Show me an image of vaint IFN captured from Bortle 8 with broad band filters (no LP filters or other tricks).  Or, maybe there are targets even fainter than IFN.  Bortle 8 and an image of....can't think of one (I haven't imaged it!)

[Rodd]

21 minutes ago, vlaiv said:

It does not matter if target is fainter then the sk

 

13 minutes ago, Xilman said:

Detecting objects which are 1000 times fainter than the sky is routine

This is why it is impossible to image Sirius B with a 4" scope (or any scope some times depending on position).  Sirius A is just too bright.  It drowns out Sirius B.

[Rodd]

17 minutes ago, david_taurus83 said:

Regarding the proposed 91% QE. Isn't this peak around the green/OIII wavelength? The more important Ha QE is more like 60%? I think this point gets lost in all the hype.

True--but the ASI 1600 efficiency for Ha is even lower, so teh 2600 is still much more efficient, no?

[Xilman]

4 minutes ago, Rodd said:

Show me an image of vaint IFN captured from Bortle 8 with broad band filters (no LP filters or other tricks).  Or, maybe there are targets even fainter than IFN.  Bortle 8 and an image of....can't think of one (I haven't imaged it!)

Here is an image of Caliban, also known as Uranus XVI, I took a while back. It was at magnitude 22.2 at the time of capture and was taken in Bortle 4 skies --- approximately 21.0 mag per square arcsecond. The seeing was around 2-3 arcseconds, so the image of the satellite is 4-9 square arcseconds.Let's be particularly generous and assume particularly good seeing --- so only 1.5 magnitudes brighter background per resolution element. The sky background is thus magnitude 19.5, or 2.7 magnitudes brighter than the satellite --- a factor of  20 times. The stacked images accumulated just under 1.5 hours. I could easily have exposed for much longer to get a better SNR but I did not need to do so.

 

[Rodd]

Another factor in that has not been discussed is FWC.  the ASI 1600 has 20K--which means that I have to shoot 10sec subs (at high gain) to recover blown star cores.  With a 50L FWC, I may not need to do this--which will save about 20 minutes per stack (Thats how much 10sec subs I take per stack).  So teh camera will be even faster

[Rodd]

10 minutes ago, Xilman said:

Here is an image of Caliban, also known as Uranus XVI,

That is not an extended target and its inverted.  Not too convincing if flipped I'd wager (as an image--scientifically its fine.  But that is not what I do)-but it is not an extended target.  That is why I said IFN.  It is only visible becuase you know where it is.  It looks like the rest of the noise in the image.  How do we know it isnt like the other blobs in the frame and just happens to be located at that spot.

Edited December 22, 2021 by Rodd

[Xilman]

17 minutes ago, Rodd said:

 

This is why it is impossible to image Sirius B with a 4" scope (or any scope some times depending on position).  Sirius A is just too bright.  It drowns out Sirius B.

Sirius B is vastly brighter than the sky.

It is most certainly possible to image the Pup with a 4" scope. You need to deal with scattered light from seeing and instrumental distortions, not sky glow. The standard approaches are to use Lucky imaging, possibly in conjunction with a coronograph and/or a diffraction grating to place the companion between the diffraction spikes.

[Xilman]

2 minutes ago, Rodd said:

That is not an extended target and its inverted.  Not too convincing if flipped I'd wager (as an image--scientifically its fine.  But that is not what I do)-but it is not an extended target.  That is why I said IFN.

Here is the flipped image. As you should expect, it shows exactly the same as the black-on-white version.

 

[Rodd]

3 minutes ago, Xilman said:

Here is the flipped image. As you should expect, it shows exactly the same as the black-on-white version.

 

Not too convincing--looks like the other blobs in teh image

[Rodd]

5 minutes ago, Xilman said:

Sirius B is vastly brighter than the sky.

It is most certainly possible to image the Pup with a 4" scope. You need to deal with scattered light from seeing and instrumental distortions, not sky glow. The standard approaches are to use Lucky imaging, possibly in conjunction with a coronograph and/or a diffraction grating to place the companion between the diffraction spikes.

My point is Sirius B is brighter not sky glow.  I thought that would be obvious.  Have to switch gears a bit and be maleable.  In this example Sirius A is the "sky glow" that is interfering with Sirius B.  Show me an image taken with a 4" scope that does not make use of a special filter or diaphram bloacking mechanism.  Just straight imaging on a sensor of Sirius B.

[Xilman]

15 minutes ago, Rodd said:

Not too convincing--looks like the other blobs in teh image

So?

Exactly the same is true of the blobs in the other image.

That particular blob is within an arcsecond of the predicted position and is within 0.2 magnitudes of the predicted brightness. No other objects are visible within that range on other images which go to comparable depth and no known asteroids are anywhere near that position at that time.

 

Convinces most people.

Edited December 22, 2021 by Xilman
s/of/on/

[vlaiv]

1 minute ago, Rodd said:

Not too convincing--looks like the other blobs in teh image

I honestly fail to see why do you need convincing?

Darkest natural sky on earth is magnitude 22.

Craig Stark that you call upon measured brightness of M51 here:

https://clarkvision.com/astro/surface-brightness-profiles/introduction.html

It goes below 22 magnitude - in that crappy little image. Look at any other deep M51 image - tidal tail is easily magnitude 26-27 that is full 5 magnitudes - hence x100 fainter than darkest skies on earth.

Yet we image M51 and many other fainter galaxies with ease.

Just because object is fainter then sky - does not mean it can't be imaged. It is not how it works.

[Xilman]

2 minutes ago, Rodd said:

My point is Sirius B is brighter not sky glow.  I thought that would be obvious.  Have to switch gears a bit and be maleable.  In this example Sirius A is the "sky glow" that is interfering with Sirius B.  Show me an image taken with a 4" scope that does not make use of a special filter or diaphram bloacking mechanism.  Just straight imaging on a sensor of Sirius B.

I do not have a 4" scope and so am unable to satisfy your request. Perhaps others here could do so.

[Xilman]

3 minutes ago, Xilman said:

I do not have a 4" scope and so am unable to satisfy your request. Perhaps others here could do so.

Actually, my guide scope is a 100mm refractor fitted with a SX Lodestar2, so I was lying ---- or, to be charitable, deluded.

I may give it a try when I return to La Palma in the spring. I see no particular reason why a few hundred 10ms exposures should not show the companion in a few images. The separation should be several pxels (more if I fit a Barlow) and Sirius B is so bright it would be an easy binocular object if it were further from Sirius A.

 

 

[Rodd]

1 hour ago, Xilman said:

Actually, my guide scope is a 100mm refractor fitted with a SX Lodestar2, so I was lying ---- or, to be charitable, deluded.

I may give it a try when I return to La Palma in the spring. I see no particular reason why a few hundred 10ms exposures should not show the companion in a few images. The separation should be several pxels (more if I fit a Barlow) and Sirius B is so bright it would be an easy binocular object if it were further from Sirius A.

 

 

Why lapalma. Why not from our soupy skies?

[Rodd]

1 hour ago, vlaiv said:

I honestly fail to see why do you need convincing?

Darkest natural sky on earth is magnitude 22.

Craig Stark that you call upon measured brightness of M51 here:

https://clarkvision.com/astro/surface-brightness-profiles/introduction.html

It goes below 22 magnitude - in that crappy little image. Look at any other deep M51 image - tidal tail is easily magnitude 26-27 that is full 5 magnitudes - hence x100 fainter than darkest skies on earth.

Yet we image M51 and many other fainter galaxies with ease.

Just because object is fainter then sky - does not mean it can't be imaged. It is not how it works.

Show me ifn taken from bortle 8/9 using lum filter. That is all the convincing I will need

[Rodd]

1 hour ago, Xilman said:

So?

Exactly the same is true of the blobs in the other image.

That particular blob is within an arcsecond of the predicted position and is within 0.2 magnitudes of the predicted brightness. No other objects are visible within that range on other images which go to comparable depth and no known asteroids are anywhere near that position at that time.

 

Convinces most people.

Could be an accident. It might be a blob coincidentally in the right spot

[ONIKKINEN]

17 minutes ago, Rodd said:

Show me ifn taken from bortle 8/9 using lum filter. That is all the convincing I will need

Why dont you try taking a picture like that yourself? Nothing impossible about that, just takes very long to get the SNR of the faint IFN to produce a nice looking image.

[Rodd]

39 minutes ago, ONIKKINEN said:

Why dont you try taking a picture like that yourself? Nothing impossible about that, just takes very long to get the SNR of the faint IFN to produce a nice looking image.

I don’t agree. I can’t take a 30sec lum sub without blowing the frame on most nights. If I take subs of 10 seconds often they are hard to align due to few and very blurry stars due to lack of signal 

[ONIKKINEN]

17 minutes ago, Rodd said:

I don’t agree. I can’t take a 30sec lum sub without blowing the frame on most nights. If I take subs of 10 seconds often they are hard to align due to few and very blurry stars due to lack of signal 

Well that i dont believe, there will always be starcores visible in any kind of condiitons. I image mostly from B6-8 type skies (depending on Moon, transparency, direction of imaging etc) exclusively with 30s subs and there are always stars. Also imaging with an 8 inch F4.2 newtonian and an OSC camera. Also, i have an in progress project where i try to image IFN from the afore mentioned conditions. At 5 hours in i can separate the IFN from the background but it is obviously very noisy and requires at least double or triple the exposure to reach a decent result. You can do the math yourself how long that would take for a hard Bortle 8 location and figure out why people dont do that. But not impossible.

Honestly i think you're trolling here. You have been given evidence left and right and it just seems that you want to move the goalposts so outrageously to your favour that nobody can actually prove their point to you. I mean you can see extremely faint signal in your own profile picture (assuming you took it yourself) that would be impossible from your arguments.