#1  
Old 19-05-2013, 03:38 PM
rogerco's Avatar
rogerco (Roger)
Roger

rogerco is offline
 
Join Date: Sep 2007
Location: Woodford,NSW,Australia
Posts: 388
Chip gain vs DSLR ISO setting

Just started reading Charles Bracken's book "The Deep-Sky Imaging Primer", very interesting read. Question:- early on he talks about the relationship between gain and dynamic range (which I would relate to number of grey tones) and suggests that as you increase ISO (increasing gain) you decrease the dynamic range. In an example he gives a theoretical DSLR where 100 ISO equates to zero gain and 200 ISO indicates a 2x gain.

Now I have a Canon 600D and read through the manual but nowhere can I find anything that discusses gain or any indication that 100 ISO might be zero gain. Do I just assume that because 100 is the lowest ISO setting that that is zero gain?

Anyone with more knowledge than me (just about everyone!) know the answer to this.
Reply With Quote
  #2  
Old 19-05-2013, 04:00 PM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,877
Hi Roger,

Interesting as I was just studying up on this myself recently.
There is a definite relationship between maximum dynamic range (showing shadows and highlights in an image) and ISO.

For most modern cameras I have seen graphs for maximum dynamic range is usually the lowest ISO it will do - typically ISO100.

Then it may stay stable for a little while, in the case of a Nikon D800E its much the same until ISO800 and then dynamic gain starts to fall.

You can see it in some images.

In nightscapes though you want the maximum ISO before ISO starts to deteriorate. That introduces a few measures. DXOMark uses an ISO rating where signal to noise ratio is at a certain level that is considered excellent image quality. Once ISO is raised above a certain point this signal to noise ratio falls and you now have less than excellent image quality. You can look up your camera and you will see an ISO rating.

See this graph. You can click on various camera models on the right to see the curve for your camera (I didn't see 60D though).

http://home.comcast.net/~NikonD70/Ch...adow.htm#D800e

It looks like ISO1600 is the optimum night sky ISO and higher than that you are simply losing dynamic range. Its much the same if you exposed at ISO1600 and then used exposure compensation in post processing to boost the image. Higher ISO simply multiplies the signal by a factor. Its not a totally simple relationship as manufacturers noise reduction enters into it as well.

So for example the Nikon D800E has an optimum ISO of 1600. This has been born out by my own testing. ISO1600 to 3200 seems good with this camera. ISO6400 is workable but you can see the data starting to stretch thin and dynamic range falling. By ISO12800 its becoming quite clear. DXOMark rates the D800E to ISO2910. So these graphs and DXO are not far apart in agreeing what is ideal.

That then means that the ideal exposure for your camera is what is required to get a bright exposure at that ISO and the lowest F ratio your lens can do without aberrations. So that means a tracked camera and a good lens is important. You can get away with ISO6400 panoramas but ideal is longer and stopped down and lower ISO for lowest noise, best signal and well exposed image that will take some processing without breaking down.

Greg.
Reply With Quote
  #3  
Old 21-05-2013, 10:22 AM
rustigsmed's Avatar
rustigsmed (Russell)
Registered User

rustigsmed is offline
 
Join Date: Mar 2012
Location: Mornington Peninsula, Australia
Posts: 3,950
Interesting stuff Roger,

Although i'm not answering your question in anyway i thought i'd add the following. I've played around alot with iso settings on the 600d (i've got one too) and always paid close attention to what results others have got with the same camera and to me from a purely subjective view it looks like either iso 400 or iso 800 is the go for this camera. iso 1600 is stretching it a bit far, but I'd be keen to see some scientific data to back it up.

cheers
Reply With Quote
  #4  
Old 21-05-2013, 10:49 AM
wasyoungonce's Avatar
wasyoungonce (Brendan)
Certified Village Idiot

wasyoungonce is offline
 
Join Date: Jul 2006
Location: Mexico city (Melb), Australia
Posts: 2,336
Latest article by Craig Stark on long exposures, see CN here. Shows the gain given for xsi and more. Craig has done a lot of articles on DSLRs and how to measure your individual camera performance, here.

However, basically the conclusion is that Canon is re-scaling image data before writing image files. The most interesting point is that ISO400 may be the best settings after all with wider dynamic range and less image scaling due to lower noise.

Anyway...FWIW.
Reply With Quote
  #5  
Old 21-05-2013, 04:28 PM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,877
Certainly too high an ISO comes at a cost of lost dynamic range.

That point would vary for different cameras and manufacturers.
It would be worth knowing where that point is.

But other factors come into it - like tracking or no tracking, dark sky or not dark sky.

One thing that is important is you can't change ISO afterwards in RAW. What a shame as it would make all this unnecessary.

Greg.
Reply With Quote
  #6  
Old 21-05-2013, 06:04 PM
rogerco's Avatar
rogerco (Roger)
Roger

rogerco is offline
 
Join Date: Sep 2007
Location: Woodford,NSW,Australia
Posts: 388
Thanks for all the comments, especially Greg for the detailed reply. I had read elsewhere that 400 - 800 - 1600 was the most likely range but Bracken's book made me wonder about zero or unity gain.

The book is very well written by the way, just starting to get into the processing chapters.


God, just watching the tornado damage in the USA, that was one big twister.
Reply With Quote
  #7  
Old 23-05-2013, 08:19 PM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,877
Quote:
Originally Posted by rogerco View Post
Thanks for all the comments, especially Greg for the detailed reply. I had read elsewhere that 400 - 800 - 1600 was the most likely range but Bracken's book made me wonder about zero or unity gain.

The book is very well written by the way, just starting to get into the processing chapters.


God, just watching the tornado damage in the USA, that was one big twister.

DXO Mark has a measure for highest ISO with no image quality loss for different cameras. Its worth a check for your camera.
Here it is here:

http://www.dxomark.com/index.php/Cam...(brand3)/Canon

As you can see it varies a lot by brand and model. For 600D it is ISO793 but for Nikon D800E is it 2979, for Canon 6D it is 2340, 5D2 ISO1815, 5D3 ISO2293, 1100D ISO755, 7D ISO854, Sony Nex 6 ISO1018, Nikon D600 ISO 2980 Sony Nex 5R ISO910, Sony Nex 7 ISO1016. Fuji XE1 would be similar to Nex 6, perhaps a bit higher IS01250 or so (not listed by DXOMark).

So there really can be a different best ISO for each camera.

Greg.

Last edited by gregbradley; 25-05-2013 at 09:42 PM.
Reply With Quote
Reply

Bookmarks

Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +10. The time is now 11:46 PM.

Powered by vBulletin Version 3.8.7 | Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Advertisement
Testar
Advertisement
Bintel
Advertisement