Ccd vs canon dlsr imaging sensitivity

[pjphilli]

Hi

Any suggestions for measuring the relative sensitivities of CCD and
Canon DLSRs cameras assuming pixels have the same size?

Cheers Peter

[Peter Ward]

You first need to record the equivalent flux. (I've had previous debates about this ;) )

Also depends if you want colour...

Bayer matrix chips do this with one shot. It takes three (filtered) shots to do the same with a monochrome CCD.

You might want to look here as a start.....

http://www.atscope.com.au/BRO/tutorials/dslr.html

[Phil Hart]

I did this test awhile ago:

http://philhart.com/content/qhy9-vs-astro40d

Phil

[Peter Ward]

Sorry Phil but I suspect you analysis is flawed. You are comparing 23 units of flux vs 15 (and even that isn't correct...but I'm lazy and couldn't be bothered with the math :) )

In terms of time to get data, then certainly things get a little muddy.

But in terms of spectral flux hitting the sensor its pretty straight forward.

The performance of CCD's, particularly with low flux, is significantly superior....as is their read noise.

An easy acid test is narrow band.

Expose, say 5 minutes, with an SII filter with a DSLR, then do the same with a mono CCD.

London to a brick you'll have usable signal with the CCD, but the DSLR will be mainly noise.

[pjphilli]

Thanks for these interesting replies. I had a feeling I was asking a difficult question but it is a bit surprising that a standard method has not been worked out previously. To get an approximate answer with my gear I think I will image a feint object and adjust the relative exposure times until I get a similar quality image. The difference in exposure times would then give me an approximate answer to my question? Cheers Peter

[Phil Hart]

do you under-estimate me that much Peter? http://www.iceinspace.com.au/forum/../vbiis/images/smilies/winking70.gif

i don't claim the test to be the final answer and it was done very early in my CCD days. but like a good little scientist, i've put all the details there so anyone can see exactly what i did and make their own interpretation.

but equivalent imaging time was certainly my main interest (i didn't get the bonus "50% more clear dark sky package" when i bought my CCD :lol:). so at the time i wanted to see what i got on the new CCD vs the cooled SLR that i was familiar with in the same amount of time, since a few hours on an object at a dark sky site was all i was lucky to get. (i do quite a lot of comparative testing of my gear at times, different ISO, binning, settings etc.. it's one of the ways i learn).

one of the common misconceptions i hear from people comparing mono-CCD to DSLR is that SLR imaging is quicker because you don't have to take images through separate filters. but as anybody who's used a nice sensitive CCD knows, you can get good (enough) binned colour data pretty damn quick and still easily match DSLR resolution. so my primary aim of the test was to show that CCD will outperform DSLR even in the same (short) time.

my surprise was how small the difference was in that case. narrow-band obviously different and clearly the much greater strength of mono CCD.

having said all that, i thought DSLRs were pretty good from a read-noise point of view, but nowhere near as good in thermal noise. quantum efficiency also lower but that's partly an RGB vs mono argument rather than CCD vs DSLR technology.

Phil

Edit.. P.S. why is that winking emoticons rarely work for me? hope i don't come across as too serious! :nerd: :prey: :whistle:

[gregbradley]

I prefer your DSLR image in that comparo, I guess because its 14 megapixels verus 8?

I'd guess that the difference would become more pronounced with dimme objects.

Also of course in narrowband where the Bayer matrix would be a liability.

Another test perhaps more even would be cooled DSLR versus one shot colour CCD. That may be very similar.

Greg.

[Peter Ward]

Not at all Phil :)

I think we are on the same page.

In terms of "time" vs "quality of result" when you have a good deal of flux, I'd agree the difference is not great.

In terms of S/N from your comments I say we'd both agree that CCD technology, particularly when you don't have many photons to play with, can be easily shown to be "more sensitive" (a no-brainer really :) )

....which I think was the question being posed.

[Phil Hart]

Indeed we are. And I often wonder whether my comparisons at f3 are fair to the CCD. I believe that at f8 or something the DSLR would really start to struggle in the comparison (even without moving to narrowband).

My main aim in originally publishing the test was to help others understand that just because you have to shoot through several filters, doesn't mean that LRGB imaging 'requires' more exposure time. So hopefully somebody will find it useful one day!

Phil