18 replies to this topic

[Hunterhall10]

I just made the switch from a DSLR to a ZWO ASI 533 and run it with ASIAIR. I used to be able to use the old DSLR trick of "histogram 1/4 to 1/3 of the way from the left" but the ASIAIR histogram is a bit more complicated. I did a bunch of research and ended up with the formula from Jon Rista from another post on here and watched the Dr. Robin Glover Presentation often referenced.

 

In the image below, I put in my camera's info into the formula and it is giving me some results that I want to confirm. When I took my 533 out for the first imaging session this past weekend of M101 I was doing exposures of anywhere from 60 sec and 180sec and got average ADUs of 3900 and 6600 respectively. With the formula though it appears I should have been in the 300-400 range. Is that accurate? I feel like that would lead to very very short sub exposures? My final image ended up with stars that have some weird Cyan or black artifacts around the edges of the star so is my overexposure causing that?

 

For reference I am using a ZWO ASI 533MC pro (Gain 100) , Canon EF 70-200 F/2.8 L ii is USM, Optolong UV/IR cut filter, on a star adventurer gti mount in a Bortle 3-4 zone.

Attached Thumbnails

• 

[imtl]

Need more information and show us the actual numbers you put in your calculation.

 

Are you sure you are not mixing up the ADC bit for different numbers you are plugging in?

[900SL]

For my 533MC Pro using AsiAir pro, the default offset is 70 (bias)

 

So bias ADU should be around 2800 ADU (70 offset, x 40 to convert to 16 bit)) and at gain 100 for swamp factor of 5, I need (5 * 1.5e)^2 = 56 which is then converted to 16 bit so 4 * 56 = 224 ADU above the bias to swamp read noise

 

Add 224 to 2800 and exposure median should be above 3024. I usually set an exposure then check ADU and histogram. I'm happy with mean ADU at 3200

 

More here, including a spreadsheet:

 

https://stargazerslo...st-lens/page/2/

[Hunterhall10]

For my 533MC Pro using AsiAir pro, the default offset is 70 (bias)

 

So bias ADU should be around 2800 ADU (70 offset, x 40 to convert to 16 bit)) and at gain 100 for swamp factor of 5, I need (5 * 1.5e)^2 = 56 which is then converted to 16 bit so 4 * 56 = 224 ADU above the bias to swamp read noise

 

Add 224 to 2800 and exposure median should be above 3024. I usually set an exposure then check ADU and histogram. I'm happy with mean ADU at 3200

 

More here, including a spreadsheet:

 

https://stargazerslo...st-lens/page/2/

 

 

Need more information and show us the actual numbers you put in your calculation.

 

Are you sure you are not mixing up the ADC bit for different numbers you are plugging in?

Ahh I believe i see my errors, I was not converting to 16 bit. My initial calculation was (10*((1.5^2)/1)+70)*((2^16)/(2^14))=370 which i feel was also an incorrect formula I got from a different youtube video. however I also see the forumla in your calculation ((5 * 1.5e)^2 = 56) is different from Jon Rister's (Nread^2 * Swamp). is that correct because it appears to give different results. I appreciate the info though and if it works for you with the same camera it will work for me haha

[imtl]

Ahh I believe i see my errors, I was not converting to 16 bit. My initial calculation was (10*((1.5^2)/1)+70)*((2^16)/(2^14))=370 which i feel was also an incorrect formula I got from a different youtube video. however I also see the forumla in your calculation ((5 * 1.5e)^2 = 56) is different from Jon Rister's (Nread^2 * Swamp). is that correct because it appears to give different results. I appreciate the info though and if it works for you with the same camera it will work for me haha

A few things. First, Jon's formula is the correct one. Also, you need to convert the offset into ADU in 14 bits. For the 533MC I think the factor is 10.

 

So, [(10*1.5^2)/1+70*10]*(2^16)/(2^14)=2890 ADU in 16 bit.

[Hunterhall10]

A few things. First, Jon's formula is the correct one. Also, you need to convert the offset into ADU in 14 bits. For the 533MC I think the factor is 10.

 

So, [(10*1.5^2)/1+70*10]*(2^16)/(2^14)=2890 ADU in 16 bit.

Understood, so I should set my sub exposure to where the mean ADU is around 2890 according to this in order to be properly exposed? thank you again for your help on this. Sometimes I just need it to be explained like I am 5 lol.

[imtl]

Understood, so I should set my sub exposure to where the mean ADU is around 2890 according to this in order to be properly exposed? thank you again for your help on this. Sometimes I just need it to be explained like I am 5 lol.

You don't have to be too maticulus about it. I don't know what numbers the ASIair spits out but my guess it will spit out in 16 bit (on the app I mean). So yes, you should aim for around 2900 ADU for a swamp factor of 10. For narrow band you don't really have to go for swamp factor of 10. Lower could work as well. 

[Hunterhall10]

You don't have to be too maticulus about it. I don't know what numbers the ASIair spits out but my guess it will spit out in 16 bit (on the app I mean). So yes, you should aim for around 2900 ADU for a swamp factor of 10. For narrow band you don't really have to go for swamp factor of 10. Lower could work as well. 

Got it and you are correct that is how ASIAIR shows it. I don't shoot in narrowband.....yet. I just had a feeling I was way overexposed after reading and watching a bunch of things today so this gives me a much better starting point next time out. Hopefully it can fix these stars too. if not just another problem to solve Thank you so much again for your help!!!!!

Attached Thumbnails

• 

[imtl]

Got it and you are correct that is how ASIAIR shows it. I don't shoot in narrowband.....yet. I just had a feeling I was way overexposed after reading and watching a bunch of things today so this gives me a much better starting point next time out. Hopefully it can fix these stars too. if not just another problem to solve Thank you so much again for your help!!!!!

The Chromatic abberations you are experiencing are probably from the quality of the lens you are using... A good lens for day light photography is not the same as for AP. I happen to have this lens, it's mediocre even for daylight...

[Hunterhall10]

The Chromatic abberations you are experiencing are probably from the quality of the lens you are using... A good lens for day light photography is not the same as for AP. I happen to have this lens, it's mediocre even for daylight...

Ahhhh gotcha, Thank you!  so what i am hearing is now its time for a dedicated astro scope!

[BucketDave]

I use Bobzeq25's rule for setting exposure now:

Expose until you have a few hundred pixels saturated. Much less exposure and you aren't using the sensors dynamic range. Much more exposure (to reduce the effect of read noise) and you'll saturate too many stars (with consequent loss of colour).

100 gain works best for the 533. I typically use 300s exposures in narrowband and 30-120s in broadband.

Forget the histogram rule, it doesn't work for astrocams!

Edited by BucketDave, 15 April 2024 - 05:25 PM.

[Hunterhall10]

I use Bobzeq25's rule for setting exposure now:

Expose until you have a few hundred pixels saturated. Much less exposure and you aren't using the sensors dynamic range. Much more exposure (to reduce the effect of read noise) and you'll saturate too many stars (with consequent loss of colour).

100 gain works best for the 533. I typically use 300s exposures in narrowband and 30-120s in broadband.

Forget the histogram rule, it doesn't work for astrocams!

What do you mean by "saturated' like visible in linear form? or is there a statistic that shows how many pixels are saturated?

[pedxing]

Saturated as in having the max value of the sensor's avalable range (65535 for a 16-bit camera).

[900SL]

 

From the thread I linked earlier

 

 

[900SL]

A few things. First, Jon's formula is the correct one. Also, you need to convert the offset into ADU in 14 bits. For the 533MC I think the factor is 10.

So, [(10*1.5^2)/1+70*10]*(2^16)/(2^14)=2890 ADU in 16 bit.

I'm fairly certain the read noise formula I quoted is correct. For a swamp factor of 5 with a 533MC pro at 100 or unity gain:

You have 1.5e of read noise. This means that you need 7.5e (x5) of LP noise or square that to get signal - 56.25e for signal, ie (5*1.5) squared.

Convert this to 16 bit and add to bias ADU for a target ADU. I usually shoot a little higher to allow for the signal from the object affecting the average ADU

Edited by 900SL, 16 April 2024 - 01:50 AM.

[imtl]

I'm fairly certain the read noise formula I quoted is correct. For a swamp factor of 5 with a 533MC pro at 100 or unity gain:

You have 1.5e of read noise. This means that you need 7.5e (x5) of LP noise or square that to get signal - 56.25e for signal, ie (5*1.5) squared.

Convert this to 16 bit and add to bias ADU for a target ADU. I usually shoot a little higher to allow for the signal from the object affecting the average ADU

It is not. The correct term is swamp*RN^2. Look at the line between the tables and the graphs in the picture that you added.

 

"Swamping factor of 3.16 is equivelent to 10*RN^2 swamping formula on CN forum".

 

For some reason, the person that put this spreadsheet together decided it would be easier to write it like that. But it is confusing people. RN is added in quadrature to other signals. So target sky background needs to be swamp*RN^2.

Edited by imtl, 16 April 2024 - 02:13 AM.

[Robert7980]

Understood, so I should set my sub exposure to where the mean ADU is around 2890 according to this in order to be properly exposed? thank you again for your help on this. Sometimes I just need it to be explained like I am 5 lol.

By “properly exposed” you do realize the swamp factor term in the equation is nearly completely arbitrary… It’s just helping provide a minimum exposure time that sets the average of the frame above the read noise, it’s not actually telling you much about how you’re actually exposed, you can still be underexposed or overexposed depending on what the target actually is. 

 

It’s the same as losing shadows or highlights in daylight photography even if the mathematical camera metering is saying you’re good… Sometimes you aren’t… Stacking makes it more complicated 

 

Shooting for 2890 doesn’t mean one of those nuclear orbs isn’t blowing things out, and it doesn’t mean you’re pushing deep into the ultra dim areas very quickly either…

 

I think it’s important to understand what the math is and isn’t telling you. It’s just a simplified (in some cases way oversimplified) way to give a general idea where your average stack noise will be with regard to background (sorta - because the object is included in the average) levels and read noise, which is more applicable to imaging with very heavy light pollution giving the shortest times you can get away with… So they aren’t “optimal” as there’s no such thing, there’s too little and too much and the gap between those is sometimes huge, like under a second to over 20 minutes just depending on the situation… 

 

‘So there’s no upper limit to capture ultra dim things (real cameras have practical limits) and the brightest area of the target sets when you’re going to clip high… neither of those things are really considered in Dr Glover’s math. In some cases you’ll actually need multiple sets of exposures to bracket enough dynamic range to capture things like the Orion Nebula… Actual space gets ridiculously dark right next to extremely bright things pretty often, so having a range of settings in your quiver is pretty useful… 

 

Just throwing a point of view out there so you aren’t held hostage in a little box of mathematical constraints… 

 

One practical point is maybe a few blue stars are now white, but you don’t have 36 quintillion subs on your hard drive, eh not so easy to find the optimum… Might not matter today, but 200TB later you realize those blue stars are expensive lol… 

Edited by Robert7980, 16 April 2024 - 02:32 AM.

[imtl]

By “properly exposed” you do realize the swamp factor term in the equation is nearly completely arbitrary… It’s just helping provide a minimum exposure time that sets the average of the frame above the read noise, it’s not actually telling you much about how you’re actually exposed, you can still be underexposed or overexposed depending on what the target actually is.

It’s the same as losing shadows or highlights in daylight photography even if the mathematical camera metering is saying you’re good… Sometimes you aren’t… Stacking makes it more complicated

Shooting for 2890 doesn’t mean one of those nuclear orbs isn’t blowing things out, and it doesn’t mean you’re pushing deep into the ultra dim areas very quickly either…

I think it’s important to understand what the math is and isn’t telling you. It’s just a simplified (in some cases way oversimplified) way to give a general idea where your average stack noise will be with regard to background (sorta - because the object is included in the average) levels and read noise, which is more applicable to imaging with very heavy light pollution giving the shortest times you can get away with… So they aren’t “optimal” as there’s no such thing, there’s too little and too much and the gap between those is sometimes huge, like under a second to over 20 minutes just depending on the situation…

‘So there’s no upper limit to capture ultra dim things (real cameras have practical limits) and the brightest area of the target sets when you’re going to clip high… neither of those things are really considered in Dr Glover’s math. In some cases you’ll actually need multiple sets of exposures to bracket enough dynamic range to capture things like the Orion Nebula… Actual space gets ridiculously dark right next to extremely bright things pretty often, so having a range of settings in your quiver is pretty useful…

Just throwing a point of view out there so you aren’t held hostage in a little box of mathematical constraints…

One practical point is maybe a few blue stars are now white, but you don’t have 36 quintillion subs on your hard drive, eh not so easy to find the optimum… Might not matter today, but 200TB later you realize those blue stars are expensive lol…

Hear hear.

[900SL]

By “properly exposed” you do realize the swamp factor term in the equation is nearly completely arbitrary… It’s just helping provide a minimum exposure time that sets the average of the frame above the read noise, it’s not actually telling you much about how you’re actually exposed, you can still be underexposed or overexposed depending on what the target actually is.

It’s the same as losing shadows or highlights in daylight photography even if the mathematical camera metering is saying you’re good… Sometimes you aren’t… Stacking makes it more complicated

Shooting for 2890 doesn’t mean one of those nuclear orbs isn’t blowing things out, and it doesn’t mean you’re pushing deep into the ultra dim areas very quickly either…

I think it’s important to understand what the math is and isn’t telling you. It’s just a simplified (in some cases way oversimplified) way to give a general idea where your average stack noise will be with regard to background (sorta - because the object is included in the average) levels and read noise, which is more applicable to imaging with very heavy light pollution giving the shortest times you can get away with… So they aren’t “optimal” as there’s no such thing, there’s too little and too much and the gap between those is sometimes huge, like under a second to over 20 minutes just depending on the situation…

‘So there’s no upper limit to capture ultra dim things (real cameras have practical limits) and the brightest area of the target sets when you’re going to clip high… neither of those things are really considered in Dr Glover’s math. In some cases you’ll actually need multiple sets of exposures to bracket enough dynamic range to capture things like the Orion Nebula… Actual space gets ridiculously dark right next to extremely bright things pretty often, so having a range of settings in your quiver is pretty useful…

Just throwing a point of view out there so you aren’t held hostage in a little box of mathematical constraints…

One practical point is maybe a few blue stars are now white, but you don’t have 36 quintillion subs on your hard drive, eh not so easy to find the optimum… Might not matter today, but 200TB later you realize those blue stars are expensive lol…

Indeed. I use the mean ADU as a ballpark figure. It is affected by altitude, LP gradient and target brightness.

Some software allows the user to sample a small region of dark sky background. I don't think this is possible with the AsiAir.

I typically shoot for a mean ADU around 3000, with the 533MC at 100 gain. I may then adjust the exposure to suit, if it is either too long or too short. Also depends on whether I am using dual NB filter.