Shooting in the Dark at iso 6400.

Just for a challenge, I went to the Osprey nest at 7:30 this evening in dark, mostly overcast conditions just to test the autofocus of the D800 and
also to see if I could get a decent flight shot in really dark conditions. It doesn't appear that dark but this is at ISO 6400.
The whites are a bit blown, almost impossible to get right shooting against a dark BG in light like this.
Just an experiment, not great IQ, I would not shoot in this light again, but I thought I post it to show
what these new generation camera's are capable of.
This is also a 50% crop.
NR on BG only, none on the Osprey.

1/1000, f5, iso6400, 500mm (300f2.8 w/ 1.7 tc), HH, D800



DSC_1860.jpg

Hi Dan, I wonder why you say the following: "The whites are a bit blown, almost impossible to get right shooting against a dark BG in light like this." ?
If you expose manually according to the whites you should not have any problem.

Pretty impressive at such very high ISO Dan, not too shabby really

Hi Dan - you're really enjoying that D800. Impressive results too.

TFS,
Rachel

The bird is amazingly noise-free for that ISO. Nice bg, too. I might have wished for a cleaner head profile, but I realize this was an experiment to add to your great Osprey files.

looks good for ISO 6400 and a 50% crop. BG is nice and the pose is good too with good eye contact. there is a slight magenta cast on the whites.

I wish the beak had not merged with the far wing. I agree with Ofer head whites seem too hot. you should be able to nail the whites with manual exposure. just constantly check your histogram and make sure you dial down the exposure as needed. Osprey whites is indeed a tricky subject but I'm sure you can nail it eventually.

Pretty cool for an ISO of 6400.
TFS
Gail

Thanks for the comments on this, I probably should have posted it to the photography discussion forum as the 6400 iso IQ is not really great.
Ofer and Arash, I used manual settings and I pushed it way to the right on purpose as I thought that exposing for the whites would have
caused the noise to get really bad on the dark part of the bird and the dark BG, so I overexposed.
Maybe not correct?, but in this situation I thought that was the way to go.

Hey Dan I would not overexpose with any camera. Saturated whites cannot be recovered but noise can be removed with good processing and that is where D800 shines.

Dan,

I agree with your push to the right as much as possible. It is easy to go to far with an osprey but you do not want to be too dark especially at that ISO. I think it looks remarkably good.

Hi Dan, the way I see things - blown whites send the shot to the bin whereas you can deal with noise in most cases. I never check the histogram - I only check the highlights alert and open the aperture one third of a stop below blowing the highlights. Works perfectly well, no issues with exposure or noise. Having said that I never shoot birds in very high ISO as I don't see the point.

Hi Dan, pretty good IQ for your high ISO, and I would be happy with this result. I like the BG, and agree, there is a slight magenta cast.

Nice one Dan!
I agree with the magenta cast - but this is easy to fix.

Quote:

---

Originally Posted by Ofer Levy

I never check the histogram - I only check the highlights alert and open the aperture one third of a stop below blowing the highlights.

---

Hi Ofer,

I can't argue with your results as I have never seen anything but perfectly exposed images from you but it can be a bit dangerous to rely solely on the highlight alerts, especially with Canon cameras.

Canon's highlight alerts are based on luminance or "perceived brightness" which takes into account the sensitivity of the human vision vision system to various colors.
The luminance of a pixel is computed from its RGB values using a weighted formula. The most often quoted formula is (0.3 * R) + (0.59 * G) + (0.11 * B).

Because the formula is so heavily weighted toward green, you can, for example, have lots of pixels with the red channel blown yet their luminance will not reach the point where highlight alerts will be triggered.
This can be a real issue when photographing subjects in very warm light or subjects with intense reds. Your camera's single color histogram (based on luminance) can look fine and you can have no blinkies at all
but when you look at the histogram for the individual RGB channels, you may see that there are clearly pixels with blown red channels.

Take any image into Photoshop, add a patch of pure red, and then choose "Luminosity" from the Channel drop down in the histogram window and you will likely not see any spikes on the right side.
However, if you enable the "All Channels" view, you will see an obvious spike in the red channel.

Nikon provides a "workaround" to this issue on most of their DSLRs with their "Highlights" display mode which will give you blinkies when any ONE of the individual color channels is blown.
Hopefully, Canon will eventually follow suit, but as far as I can tell from the 1DX manual, nothing has been changed in this regard.
So, for now, Canon users should set their default Histogram type to RGB (rather than Brightness) and keep an eye on the red channel in warm light or when photographing subjects with intense reds.

I also use the highlight alert as opposed to histogram. It has been working well for me. The RGB histogram in Canon cameras is too small, if only a small area is blow it will not show in the histogram so it's not very useful IMO.

Q for Mike, do you have a reference from Canon that indicates they use such formula or is it just something that has been said on the net?

Quote:

---

Originally Posted by arash_hazeghi

Q for Mike, do you have a reference from Canon that indicates they use such formula or is it just something that has been said on the net?

---

Hi Arash,

I first came across the concept of Luminance (also referred to as Luminosity or Brightness) vs. Individual Color Channel histograms in the Cambridge in Colour tutorials on the subject :

http://www.cambridgeincolour.com/tut...istograms1.htm
http://www.cambridgeincolour.com/tut...istograms2.htm

Then I found this article and performed the tests described long ago on my 1D Mk II which verified the results :

http://www.brisk.org.uk/photog/hiblow1.html

But the real proof is that I have seen it in the field on numerous occasions with the Mk II, Mk III, and now the Mk IV (e.g. with white bird at sunset), where the Brightness histogram had no indication of any values anywhere near the right side and there were no blinkies, yet an examination of the individual RGB histograms showed that the Red channel was clearly and badly blown.

You can see an example of such a situation on page 2 of this Canon web page :

http://www.learn.usa.canon.com/resou...stograms.shtml

by clicking on the phrase "there are instances where it can be useful to view a graph of all three color channels"

Hi Mike, thank you for the excellent information. I wasn't aware of this limitation of the Canon system although I have noticed this issue in the field so acted accordingly - when shooting in very warm light and/or birds with red parts I am more careful and expose a bit less than what the highlights alert allows me.

Thanks Mike too bad the blinkies are only based on weighted luminance. I hope they can provide a full screen RGB histogram like Nikon cameras.