Highlight Headroom of D5100

Started Aug 23, 2012 | Discussions thread
OP MichaelEchos Regular Member • Posts: 210
Re: Nope

Mako2011 wrote:

MichaelEchos wrote:

Graystar wrote:

apaflo wrote:

Graystar wrote:

MichaelEchos wrote:

Anyone could tell me the highlight headroom from D5100 we could get in normal conditions (which includes normal white balance) I could get by spot metering the brightest part of the seen?

The answer is 2.7 stops and it's the same for all cameras. White balance doesn't matter, but the color of the light does. With light that has a lot of red, there's a slight risk of clipping.

All cameras from about 2007 on are calibrated the same way...which is to 12.7% gray. You get nearly 3 stops of highlight room over the reference. So you can spot meter a tone, set EC to 2.7, and get the largest signal possible from that tone without clipping.

Please read the cite that panos_m provided from rawdigger.com, and also read the references they provided.

Nikon, Canon and Sekonic use a calibration constant of 12.5% for reflected light, which provides 3.15 EV between middle gray and 100% white. Kenko and Pentax allow for about 0.4 EV less, giving 2.75 EV from middle gray to white.

Sekonic uses a "K" of 12.5, as stated on their website. As far as I'm aware, neither Canon nor Nikon provide a "K" value because "K" isn't used in the calibration of in-camera meters. All the information about metering that is referenced on that page is hearsay, as none of it can be traced to reliable sources. Much of the information on the Wiki page doesn't have any citations.

Also, if you're calibrated to 12.5% reflectance, then you only have 3 stops to 100%...not 3.15. It's simple math. The "3.15" seems to come from a completely unrelated calculation on the wiki page that's being misapplied...a calculation on the calibration constants of handheld meters, using arbitrary values. If the camera had 3.15 stops from its metered reflectance to saturation, then the metered reflectance would be sRGB 94, 94, 94. But it isn't. If you take a picture of a neutral surface and minimally process the image (demosaicing and WB only) then the sRGB values of the metered area will average 100, 100, 100...which is 12.7%, and give just under three stops of highlight headroom from the reference luminance. This is supported by that fact that on Nikon cameras, photographing a bright white at +3 EC will get clipping in the green channel.


What about different white balances?

I think I'll spot meter at brightest part of the scene and add 3 1/3 stop EC because the spot for D5100 is quite large, causing it to cover quite an amount of area outside the part I'm trying to cover.

Won't work. Try it, do as suggested and measure what you get from metering and shooting the neutral surface. Go ahead and go out and do as you suggest and bump EC by 3 1/3. You will find, graystar is correct. I only know that as I did exactly as he suggested and found it to be true in practice.

What if I cover a larger part where there is darker areas in the spot?

Post (hide subjects) Posted by
MOD Mako2011
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow