Avoiding Saturating an Image

[russellhq]

I've read a lot about not saturating the image but not how you can measure the saturation.

So how do I go about measuring saturation and telling if my image is saturated?

Also, on a similar note, how do I measure the amount of light pollution in an image and how do I know if the level of light pollution is too high?

Knowing how to do this I can then tailor the ISO and exposure length to the particular subject.

[Rustmonkey]

If you shoot in RAW then saturation shouldn't be a problem.....can be managed in photoshop later.

you'll see the LP on an image probably as an orange glow....I guess it's trial and error and trying to find as dark as site as is possible...not easy in the Midlands.

[Rustmonkey]

RAW format effectively turns off all the cameras onboard adjustments allowing the user to play with them later. Feel free to correct me if I'm wrong but that is my understanding.

[russellhq]

Thanks rustmonkey but maybe I'm using the wrong terminology. I'm talking about over-exposing stars e.g. in the core of a globular cluster or the core in the orion nebula.

Is there a quantitative measure I can take, e.g. a reading from the histogram or is it a qualitative measure i.e. try an exposure and if it looks good to the eye then carry on?

[jgs001]

That's a tricky one, and not one I've an answer too. You also need to be careful in processing to avoid blowing the stars completely... it's easy to do when pulling the background faint stuff out. When I did my M42 image last year, I deliberately shot a series of short (30 second exposures I think) to capture the trap, and carefully layered that onto the main image. I think that's probably the best option to extend the dynamic range. As even shooting in RAW, it's still easy to over expose the brightest parts of an image.

[Velvet]

what camera are you using for the DSO's. Normally ISO800 is what i use and then as many subs as possible.

The rest is left to photoshop with the curves and levels.

Velvet

[russellhq]

I'm using an unmodded Canon 1100D. The trouble is, when I expose for too long, some parts of the image cannot be recovered in Photoshop because they have become over-exposed.

Also, because I'm not working under dark skies, sky glow starts to creep in the longer I expose. But I don't know where the cut-off is where the sky glow is too much and the image can't be recovered.

[jgs001]

I'm not sure on the skyglow question myself... And I'm not sure the histogram display on the camera is able to show specifics on the saturation levels for pin points... I believe it's averaged. Whether one of the laptop control applications, like APT, is able to analyse the saturation levels at that point, I've not looked... I probably ought to

[sologuitarist61]

Most DSLR's also have a function in the menu (some times in the 'modes') for how the image is processed in-camera - I usually set it to 'natural' and that cuts that bit of the in-camera processing to almost nil - that lets you continue using JPEGS and saves any problems you may later get in processing RAW files, so programs just don't like 'em

[LukeSkywatcher]

I'm using an unmodded Canon 1100D. The trouble is, when I expose for too long, some parts of the image cannot be recovered in Photoshop because they have become over-exposed.

Also, because I'm not working under dark skies, sky glow starts to creep in the longer I expose. But I don't know where the cut-off is where the sky glow is too much and the image can't be recovered.

It really is down to trial and error from individual locations as to how long of an exposure you can take before skyglow becomes an issue.

For me (in my location) the perfect settings for my Canon 450D are: ISO @800, 8-10 seconds per exposure on the BULB setting. Because the exposures are so short,i have to take dozens if not hundreds and then i use free software (DSS or Registax) to stack them all into one single exposure.

Example: 10 exposure of 8 seconds each when stacked = 80 seconds (1 min 20s) of exposure and skyglow is nowhere to be seen. I end up with a nice black sky and white stars (i like taking widefield images of constellations etc).

Hope this helps.

Summary:

If skyglow is a worry for you (it is with most of us) just take many many shorter exposures and stack them. The stacking software is self explaining and nothing complcated in the least.

[themos]

I use IRIS to examine .CR2 files. Use the File -> Load a Raw File menu and then use View -> Histogram. IRIS is tricky to operate, though. But it does show a proper histogram which will tell you how oversaturated your image is.

[russellhq]

I use IRIS to examine .CR2 files. Use the File -> Load a Raw File menu and then use View -> Histogram. IRIS is tricky to operate, though. But it does show a proper histogram which will tell you how oversaturated your image is.

What do you look for in the histogram to measures how over saturated the image is?

[peroni]

Thanks rustmonkey but maybe I'm using the wrong terminology. I'm talking about over-exposing stars e.g. in the core of a globular cluster or the core in the orion nebula.

Is there a quantitative measure I can take, e.g. a reading from the histogram or is it a qualitative measure i.e. try an exposure and if it looks good to the eye then carry on?

I think reading the histogram is the right thing to do. Check this link for some extra info. about Canon histograms. (I have a Canon 1000D)

http://support-sg.canon-asia.com/contents/SG/EN/8200229500.html

I've only taken shots in on a camera so can't say what the histograms look like through telescopes but I think if you shoot in RAW and check the histogram, over exposure shows up as the histogram 'bell' shape getting cut off.

[themos]

yes, you find the largest values in your image (IRIS shows you these with the "stat" command) and then you look in the histogram around these values and make sure that they are not "bunched up" right against the highest possible value for your camera. It's best to shoot a deliberately overexposed RAW and have a look at its histogram so you know how that looks.