How to image globular clusters with DSLR without burning stars colors?

[kalasinman]

Due to the Sony sensor, the D100/D7000 have very low read noise at all ISOs. Unity gain is below ISO200. I'm still integrating a lot of new kit and software, but my goal is to get guiding good enough to run 5-10 minute or longer subs at ISO 400 or below in time for next galaxy season.

[faulksy]

Ahahaha that's the best comment by far! 

True indeed..but I like photography so I actually would anyway want to do all the work to produce a final pic to hang somewhere But I envy your Dob surely

only messing, but you no that. great sense of humor, we like that here

great capture buddy. but dont tell anyone i said that, as im visual only and shouldnt be in this section

[rofus]

Due to the Sony sensor, the D100/D7000 have very low read noise at all ISOs. Unity gain is below ISO200. I'm still integrating a lot of new kit and software, but my goal is to get guiding good enough to run 5-10 minute or longer subs at ISO 400 or below in time for next galaxy season.

Good luck

I find that 800iso with drizzling/dithering is good, more is not, and yes 640iso or lower even better when possible.

It took me a good amount of work and I'm an IT Systems Administrator...I run on my Macbook a Parallels virtual machine with Windows XP dedicated only and uniquely to astrophotography with all the software versions matching. Big improvement on my EQ5 was done updating the firmwares of motors and handset to latest versions, huge improvement with motors in particular.

Also ditching PHD2 because somehow with same parameters it doesn't work as well as PHD1.

I have a heavy Startravel 150 on my EQ5, plus camera/battery grip, and Orion 200mm guidescope and Starshooter guider...couple of kilos over mount recommended load (for visual!), but it doesn't matter as long as all is well balanced and scope is well aligned through the polarscope (no drifting needed) and the mount is stable and in plain.

With that system I don't even align to stars, no need, I use Astrotortilla with EQMod so every object is dead spot centered, and depending on the position I can track up to 20 minutes (see other posts of mine with screenshots of perfectly flat PHD graphs) with my 750mm focal...I will get soon to experiment with 1250mm and will let you know, up to 1500mm should be enough with the guidescope I have

As I said it took long time of experimentation on the software side, but once done I setup all in 15 minutes and it goes without faults of any sort night after night since months. Good luck with your preparation

[dph1nm]

Not sure what camera you consider, but for any digital camera, for obvious reasons I'm not explaining here, a 10s pic at 6400 ISO will be always more noisy than the same 10s pic at 100 ISO.

That is not 'intrinsic', is just real life simple example of two same length pics (same exposure at higher ISO = more noisy data). Then obviously depending on the quality of the camera, sensor, temperature etc. this difference can be more or less obvious etc. but here I'm talking from a practical point of view

The noise from an astrophysical source (including skyfog) is proportional to the square root of the number of photo-electrons detected by the camera. This Is not dependent on ISO, so the true noise is the same at all ISO. All ISO does is in effect is multiply the numbers by a constant - it does not increase the sensitivity of your camera to light.  To put it another way, the signal-to-noise ratio is the same at ISO100 as it is at ISO128000.  Now there are some sources of noise intrinsic to the camera which may be ISO dependent (e.g. read-noise), but these may actually be lower at high ISO in some cameras.

There was a web page I saw once where someone had taken some astro shots with the same exposure but at all different ISOs (100,200,400,800,1600) and multiplied them by 16,8,4,2,1 respectively., in order to show that they were essentially indistinguishable.

NigelM

[kalasinman]

The noise from an astrophysical source (including skyfog) is proportional to the square root of the number of photo-electrons detected by the camera. This Is not dependent on ISO, so the true noise is the same at all ISO. All ISO does is in effect is multiply the numbers by a constant - it does not increase the sensitivity of your camera to light.  To put it another way, the signal-to-noise ratio is the same at ISO100 as it is at ISO128000.  Now there are some sources of noise intrinsic to the camera which may be ISO dependent (e.g. read-noise), but these may actually be lower at high ISO in some cameras.

There was a web page I saw once where someone had taken some astro shots with the same exposure but at all different ISOs (100,200,400,800,1600) and multiplied them by 16,8,4,2,1 respectively., in order to show that they were essentially indistinguishable.

NigelM

Please show us 2 otherwise identical images one taken at ISO 100 and one at ISO 1228000 so we can verify that what you say is true, that the noise levels will be indistinguishable.

"real noise"?? So read noise and thermal noise aren't "real noise" based on what standard? To me, if they show up in an image, they are real enough for me.

[rofus]

Please show us 2 otherwise identical images one taken at ISO 100 and one at ISO 1228000 so we can verify that what you say is true, that the noise levels will be indistinguishable.

"real noise"?? So read noise and thermal noise aren't "real noise" based on what standard? To me, if they show up in an image, they are real enough for me.

I agree completely but I don't enter the debate.

What Nigel is speaking about is all technical true stuff, but it just doesn't matter in reality as it doesn't relate to 'real life' usage/case, where higher iso/shorter exposures is 90% of the time (and equipment) worse than lower iso/longer exposures, and then combining this with the number of exposures (and drizzle/dithering) things change as we all know etc.

For me is like saying that a more powerful engine doesn't make a car going faster, because is not 'necessarily' true and not an 'intrinsic' feature of a more powerful engine, because it depends (obviously) how heavy is the car, how much you push it, etc. etc. Yeah all correct but the fact is that give two identical cars in identical conditions, and anyway in most of the other cases (real life), a more powerful engine WILL make a car faster. We all know the theory, we use it if needed, but 90% of the time it doesn't matter to real life (even if true obviously).

[rofus]

As an example...forget such a clean image like the one below (took it combining 4 different exposure times for the various zones all at iso800) with any higher iso, I think Nigel as well cannot take one as clean at high iso (clean but still with details because no denoise used as there was no noise to start with because of the low ISO/low noise).

I underline this only because for newbies can be confusing, and the rule is always valid: keep your ISO as low as possible and your exposures count (how many pics you take) as high as possible to deal with noise 

[kalasinman]

...and don't forget to DITHER !

[rofus]

...and don't forget to DITHER !

Always, absolutely!

...but that is a more advanced topic for newcomers usually..but definitely that helps HUGELY achieving more resolution and less noise at the same time, crazy not doing it indeed

[chrisvdberge]

Just to chime in on the ISO levels and noise discussion: 

During my stay in Namibia I tried several settings a few times and compared the frames with PixInsight to measure the SNR. Sometimes visually the frames seemed identical, but the measurement did show a difference. 

Turned out subs of 10 min at ISO400 would be better than 5min at ISO800. Surprising to me as I thought the unity gain of a D600 would be around 800. Apparently not.

More surprising: for my D7000 a 10min sub at ISO400 is better than a 10min sub at ISO800.

I guess my point is that the theoretics about all the different kinds of noise and differences in each camera (model)  are so complicated it's best to just try out for yourself what works best. 

[rofus]

Excellent point and observations, they're pretty much similar to the ones I did with my D90.

I agree that practically trying without adhering too much to any theory is the best way. I also got some similar tests on Canon D5 and a modded Canon d400 (if I remember correctly), and I got always the same thing (that just confirms a simple theory that a warmer chip and/or higher gain will make inherent more noise): go always with the maximum amount of ISO that gives you the less possible noise, and that's usually between ISO400 and ISO800 for most DSLRs, and can be up to 1200 for the professional ones (Canon D5 etc.).

Clear skies to everyone...next week first light of my new Altair Astro 6" Ritchey Chretien astrograph! Can't wait..is going to be a huge step up!

[dph1nm]

I underline this only because for newbies can be confusing, and the rule is always valid: keep your ISO as low as possible and your exposures count (how many pics you take) as high as possible to deal with noise

Telling someone to use the lowest ISO possible is not necessarily optimal for astro work. As others have said, you probably want to be around unity gain (which can be between ISO800 and 1600 on some older cameras). Also, if you are using a Canon and are read-noise limited you need to be at high ISO for the best results.

NigelM

[kalasinman]

Unity gain under ISO 100 for my "old" D5100