Advice on exposure time

[Shibby]

I know there are quite a few thread on this already, but now I've started guiding, I was hoping for some advice on it with my particular situation. It's not always easy to relate existing discussions to your own data...

Here's one of the best scoring (according to DSS) subs from my last imaging session. It's 300s, ISO 800 taken with my 450D + SW LPF. Please ignore the colour balance.

The "sky background" is 4.77%.

It seems quite bright already. How bad is this LP for 5 minutes (at F5)? How much longer subs should I be aiming for, or would you say I'm already becoming limited by sky background?

Thanks!

[Tim]

Hi Lewis,

I tend to find that if the histogram of the raw untouched picture shows the bulk of the data much more than 1/3 of the total ADU value of the camera chip, that the images become harder to process. Keep the data spike in the left hand third of the histogram and you should be ok.

Dont forget that sky conditions alter all the time, so what works well one night might be rubbish the next.

Cheers

Tim

[Shibby]

Thanks for the advice Tim. So, up to that point it's always worth taking as long exposures as possible (assuming not over-exposing the target)?

Rather embarrassingly, I'm having trouble actually determining the histogram peak location... every piece of software I open it in seems to place it somewhere different!! I think most of them are applying white balance adjustment.

The raw images look a lot darker in DSS (with the brightness slider half-way) than any other tool I use.

Perhaps I should put the images on the camera and look at the histogram on there?

[Tim]

Perhaps I should put the images on the camera and look at the histogram on there?

Sorry Lewis, yes that was what I meant, check the histogram on the camera on the first couple of exposures and adjust times accordingly.

I believe DSS can rate the images on the fly too, but have never tried that. I usually reject backgrounds with over 5%, but each target varies. Hard to put a rule on it.

Cheers

Tim

[MartinB]

Thanks for the advice Tim. So, up to that point it's always worth taking as long exposures as possible (assuming not over-exposing the target)?

No, not always. You want to have enough subs to allow statistical rejection stacking algorithms to work i.e a minimum of 10 subs. Not usually a problem with broad band but does become an issue using narrowband filters which require much longer exposures.

Also, the longer your subs the greater your chance of loosing them - stray clouds, gusts etc.

As a minimum you should be aiming for the read noise in the image to be a negligible part of the overall noise <10%. Exposing beyond this produces only small benefits in overall signal to noise ratio. If you know the characteristics of your chip here is a very good calculator Welcome to CCDWare

Craig Stark has written some excellent articles on noise and expsoure times backed up with real live data (look at 2009 articles)

Articles & Reviews