ISO v sub length v histogram peak v light pollution

[Physicist13]

17 hours ago, Thalestris24 said:

Hi

I was going by http://www.photonstophotos.net/Charts/ReadNoise_e.htm#Canon%20EOS%206D and http://www.dxomark.com/Cameras/Canon/EOS-1100D---Measurements (for the 1100d) but maybe they're not applicable to AP? I wish there was more and relevant info out there.

Louise

thanks! let me look... this is counterintuitive and I dont see the physics behind it yet....its implying that the noise increases as amp gain decreases??.........

[Thalestris24]

2 hours ago, Physicist13 said:

thanks! let me look... this is counterintuitive and I dont see the physics behind it yet....its implying that the noise increases as amp gain decreases??.........

I don't understand it either! I'd like to know more details about how they do the tests. Clearly, though, the tests aren't done in an AP situation - would you necessarily get the same results under long exposure/low light? If I ever get some clear night skies again I might do some comparative tests of my own

Louise

[StuartJPP]

4 hours ago, wxsatuser said:

The 6D is a remarkable camera at higher ISOs and does really well at short exposures.

The Ha responce is very good out of the box.

You mean I modified mine for nothing? I had to clench for two hours whilst doing it

[Physicist13]

22 minutes ago, Thalestris24 said:

I don't understand it either! I'd like to know more details about how they do the tests. Clearly, though, the tests aren't done in an AP situation - would you necessarily get the same results under long exposure/low light? If I ever get some clear night skies again I might do some comparative tests of my own

Louise

I'm trying to find out..they mention "international standards" but I see no ref to any ISO standard.  Results looks v strange to me! their 18%SNR (???) drops with ISO at the 3dB/oct rate?!! ...def not AP type conditions.  I always rec following Craig Starks method (a physicist ;)) ) 

you can noise meas from Bias frames and darks etc... keep you busy on cloudy nights!

yes do the experiment!!  

[Thalestris24]

2 hours ago, Physicist13 said:

I'm trying to find out..they mention "international standards" but I see no ref to any ISO standard.  Results looks v strange to me! their 18%SNR (???) drops with ISO at the 3dB/oct rate?!! ...def not AP type conditions.  I always rec following Craig Starks method (a physicist ;)) ) 

you can noise meas from Bias frames and darks etc... keep you busy on cloudy nights!

yes do the experiment!!  

I don't have any tools to measure noise. I was thinking of just doing an unscientific visual comparison of calibrated lights at different iso. At the end of the day I'm only interested in what actual images look like. Tiny differences in noise levels don't matter if my eyes can't detect them! As it happens there might be a short clear spell between 6 and 7pm tonight - though that might change!

Louise

[RSM]

Hi. Lots of great thoughts in here about how the different variables can be altered to give different results. I recently started guiding and have found that typically a set of 3 minute subs at ISO 800 works for me. I'm 25 miles as the crow flies from central London so there's lots of light pollution around. Histogram on a dark night is typically mid range. I don't use an LP filter but process in PI which does a good job of removing LP and vignetting. I've had some good success with these settings for orion, the trapezium is a bit overexposed but lovely detail in other areas. Have taken some shorter exposures too and plan to look at layering soon. Like someone else mentioned, try and dither too. That seems to help massively in keeping noise down allowing you to stretch the good photons you've captured that little bit more. Good luck with your next few attempts. For me, that's one of the great parts of learning how to get good images. Richard. 

[wxsatuser]

11 hours ago, StuartJPP said:

You mean I modified mine for nothing? I had to clench for two hours whilst doing it

I wanted to mod mine but did'nt have the guts to do it.

Have'nt had anytime to do long exposures but I was rather surprised what it picked up on short 30sec subs.

7x 30secs no real attempt to process.

[Physicist13]

really excellent info here! http://www.clarkvision.com/index.html 

P

[Physicist13]

On 2/8/2016 at 15:25, dph1nm said:

Not true for Canons. The read noise (in e-) is much higher at lower ISOs.

NIgelM

have a look at http://www.clarkvision.com/articles/digital.sensor.performance.summary/#read_noise explains things very well! 

Pat

[Ken82]

Oh at least something , going by that graph my 6d is quite good

[dph1nm]

3 hours ago, Physicist13 said:

have a look at http://www.clarkvision.com/articles/digital.sensor.performance.summary/#read_noise explains things very well! 

Pat

Yes, see table 4a which shows how read noise (in e-) is much higher at low ISO. I would note mind you that much of the data on that page refers to very old models of camera.

NigelM

[Physicist13]

its not really read noise as Clarke explains..its really ADC "noise" . I'm prob being over semantic due to my job!