Tag Archives: claibration

How We Judge Exposure Looking At an Image And The Importance Of ViewFinder Contrast.

This came out of a discussion about viewfinder brightness where the compliant was that the viewfinder on the FX9 was too bright when compared side by side with another monitor. It got me into really thinking about how we judge exposure when purely looking at a monitor or viewfinder image.

To start with I think it’s important to thing understand a couple of things:

1: Our perception of how bright a light source is depends on the ambient light levels. A candle in a dark room looks really bright, but outside on a sunny day it is not perceived as being so bright. But of course we all know that the light being emitted by that candle is exactly the same in both situations.

2: Between the middle grey of a grey card and the white of a white card there are about 2.5 stops. Faces and skin tones fall roughly half way between middle grey and white. Taking that a step further between what most people will perceive as black, something like a black card, black shirt and a white card there are around 5 to 6 stops and faces will always be roughly 3/4 of the way up that brightness range at somewhere around about 4 stops above black . It doesn’t matter whether that’s outside on a dazzlingly bright day in the desert in the middle East or on a dull overcast winters day in the UK, those relative levels never change.

Now think about this:

If you look at a picture on a screen and the face is significantly brighter than middle grey and much closer to white than middle grey what will you think? To most it will almost certainly appear over exposed because we know that in the real world a face sits roughly 3/4 of the way up the relative brightness range and roughly half way between middle gray and white.

What about if the face is much darker than white and close to middle grey? Then it will generally look under exposed as relative to black, white and middle grey the face is too dark.

The key point here is that we make these exposure judgments based on where faces and other similar things are relative to black and white. We don’t know the actual intensity of the white, but we do know how bright a face should be relative to white and black.

This is why it’s possible to make an accurate exposure assessment using a 100 Nit monitor or a 1000 Nit daylight viewable monitor. Provided the contrast range of the monitor is correct and black looks black, middle grey is in the middle and white looks white then skin tones will be 3/4 of the way up from black and 1/4 down from white when the image is correctly exposed.

But here’s the rub: If you put the 100 Nit monitor next to the 1000 Nit monitor and look at both at the same time, the two will look very, very different. Indoors in a dim room the 1000 Nit monitor will be dazzlingly bright, meanwhile outside on a sunny day the 100 Nit monitor will be barely viewable. So which is right?

The answer is they both are. Indoors, with controlled light levels or when covered with a hood or loupe then the 100 Nit monitor might be preferable. In a grading suite with controlled lighting you would normally use a monitor with white at 100 nits. But outside on a sunny day with no shade or hood the 1000 Nit monitor might be preferable because the 100 nit monitor will be too dim to be of any use.

Think of this another way: Take both monitors into a dark room and take a photo of each monitor with your phone.  The phone’s camera will adjust it’s exposure so both will look the same and the end result will be two photos where the screens will look the same. Our eyes have iris’s just like a cameras and do exactly the same thing, adjust so that the brightness is with the range our eyes can deal with. So the actual brightness is only of concern relative to the ambient light levels.

This presents a challenge to designers of viewfinders that can be used both with or without a loupe or shade such as the LCD viewfinder on the FX9 that which be used both with the loupe/magnifier and without it. How bright should you make it? Not so bright it’s dazzling when using the loupe but bright enough to be useful on a sunny day without the loupe.

The actual brightness isn’t critical (beyond whether it’s bright enough to be seen or not) provided the perceived contrast is right.

When setting up a monitor or viewfinder it’s the adjustment of the black level and black pedestal which alters the contrast of the image (the control of which is confusingly called the brightness control). This “brightness” control is the critical one because if the brightness adjustment raises the blacks by too much then you make the shadows and mids brighter relative to white and less contrasty, so you will tend to expose lower in an attempt to have good contrast and a normal looking mid range. Exposing brighter makes the mids look excessively bright relative to where white is and the black screen surround is.

If the brightness is set too low it pulls the blacks and mids down then you will tend to over expose in an attempt to see details and textures in the shadows and to make the mids normal.

It’s all about the monitor or viewfinders contrast and where everything stits between the darkest and brightest parts pf the image. The peak brightness (equally confusingly set by the contrast control) is largely irrelevant because our perception of how bright this is depends entirely on the ambient light level, just don’t over drive the display.

We don’t look at a VF and think – “Ah that face is 100 nits”.  We think – “that face is 3/4 of the way up between black and white” because that’s exactly how we see faces in all kinds of light conditions – relative levels – not specific brightness.

So far I have been discussing SDR (standard dynamic range) viewfinders. Thankfully I have yet to see an HDR viewfinder because an HDR viewfinder could actually make judging exposure more difficult as “white” such as a white card isn’t very bright in the world of HDR and an HDR viewfinder would have a far greater contrast range than just the 5 or 6 stops of an SDR finder. The viewfinders peak brightness could well be 10 times or more brighter than the white of a white card. So that complicates things as first you need to judge and asses where white is within a very big brightness range. But I guess I’ll cross that bridge when it comes along.