can't understand - why did this photographer meter off the sky?

[BZChi7d]

I've just read something in Bryan Peterson's Understanding Exposure the next scenario:

- he was photo'ing a landscape and wanted to get everything sharp.
- he point the camera to the sky above to get his reading with the aperture set to 2.8 i think.
the camera says that a (let's say) 1/200s is needed to get the corect exposure
- he wants to get everything sharp remember ? so he calculates that from 2.8 aperture to 16 aperture there are like 5-6 stops so he sets his aperture to f16 and then lenghten the time accordingly.
- he shoots the landscape.

now -why, while taking the light reading from the sky, didn't he use the f16 aperture right from the beginning ? would he have got a different results regarding shutter speed ?

please, excuse my bad english.

radu

[carregwen]

Radu

What page in the book? and what shot?

[Steaphany]

I feel it's just his style of working and sounds like he just started out with his camera set to Auto.

Yes, if you start out in Aperture Priority mode, select f/16, point to the sky to determine the exposure, frame, and shoot, you will get the same result.

BTW, you can get a better sense of how he works by looking at his youtube videos:

http://www.youtube.com/user/ppsop2009

For a professional photographer and instructor of his standing, these are quite amateur. Watch, he'll provide good tips and info, just don't get sea sick from the video camera's motion or distracted by the clutter of the surrounding area where the video's being shot.

[David]

Steaphany - I agree that on first thought you ought to get the same exposure reading at f16 as you would calculate from f2.8. However, in practice, that may not occur. Whether it did or not would depend on how linear the response of the evaluative metering was on the camera. It is an interesting point and I shall have a go experimenting the next time I get a chance.

Cheers

David

[Steaphany]

David,

I agree, but what manufacturer would risk their reputation by allowing products that far "out of spec" to be on the market ?

By definition, the shutter speed and aperture are supposed to track for a given ISO and sensed light level.

If a camera's internal meter or a separate light meter behaved that poorly when new and right out of the box or after just a short time of use, considering how fast new models are brought to market, again the manufacturer's reputation would suffer and they better make good on warranty repair or replacement.

[Dave Humphries]

I haven't seen the book(?); but from what's written here, are we not overlooking that fact that the exposure will be about 2 stops under if metered from the sky and including foreground?

(I learnt that from Colin )

and that's in addition to the other bit that didn't make sense that you have all discussed.

Confused,

[arith]

I haven't seen the book(?); but from what's written here, are we not overlooking that fact that the exposure will be about 2 stops under if metered from the sky and including foreground?

(I learnt that from Colin )

and that's in addition to the other bit that didn't make sense that you have all discussed.

Confused,

I'm confused because I thought he would just get great sky. I look at the sky to determine the brightest and the adjust up to that according if there are a lot of white objects or not, and other things.

In high contrast I always bracket anyway.

[Dave Humphries]

because I thought he would just get great sky

Me too, but it said

he was photo'ing a landscape and wanted to get everything sharp

not "skyscape", so I assumed there would be foreground in shot and therefore, I'd have thought one might aim to expose for that

The requirement for everything to be sharp implies more depth than a skyscape, which is essentially all at infinity

[Colin Southern]

now -why, while taking the light reading from the sky, didn't he use the f16 aperture right from the beginning ? would he have got a different results regarding shutter speed ?

If you stop down the aperture by 5 stops and then decrease the shutterspeed by 5 stops then - exposure wise - you end up in the same place. eg 1/200 @ F2.8 gives the same exposure as 1/6th @ F16.

With low-light landscape though metering at F16 can bite you in the bum becuase most cameras can't meter for a shutterspeed of longer than 30 seconds. So - for example - if you're metering shows an exposure of 15 seconds is required at F2.8 then at F16 it would need 8 minutes - but - the camera can't do an 8 minute exposure automatically so it would give you a 30 second exposure that would be under-exposed by 4 stops (ie almost completely black frame).

Out of interest, you can also use a similar trick for metering using the ISO instead of aperture.

[Colin Southern]

I'm confused because I thought he would just get great sky. I look at the sky to determine the brightest and the adjust up to that according if there are a lot of white objects or not, and other things.

Metering of the sky is fine - but there are still quite a few variables:

- Spot or Evaluative/matrix metering

- How you want the sky to appear

- How much dynamic range the scene requires -v- how much you can capture

- The eveness of the lighting in the sky

If you want to capture all tones in the sky as efficiently as possible then you can spot meter the brightest part of the sky (which will return a shutterspeed to expose the sky as a mid-tone) but you then need to up-shift the expose by around 2 to 3 stops to expose it as a highlight (2 for safety, 3 if you're being agressive and/or have highlight tone priority turned on).

Having just said all that though, having a sky that's exposed correctly technically, doesn't necessarily mena that it will look the best visually; all colours wash out the more you expose them - with blue if you under-expose it by a couple of stops it'll look far more saturated - which is often desireable.

In practice it comes down to a number of competing variables, and which one wins out. A lot of my stuff is high-contrast - I can't have blown highlights (large areas of "white ink" on the canvas don't look cool), but if I stop down too far I'm going to have a lot of shadow noise when I did detail out with the fill light control in ACR.

So ...

... In summary, my personal technique is to spot-meter the brightest part of the sky - up shift it by 2 stops (plus x more if I'm using filters) and it's usually pretty right - but that's different to a normal dynamic range exposure where I'm wanting a saturated looking sky.

Note: the reason I do it my way is that most of my landscape exposures are multi-minute and thus I use a lightmeter and set the camera on manual.

[Colin Southern]

I haven't seen the book(?); but from what's written here, are we not overlooking that fact that the exposure will be about 2 stops under if metered from the sky and including foreground?

(I learnt that from Colin )

Yes, No, Maybe

Assuming it's evenly lit all you can assume is that it'll be exposed as a MIDTONE - whether that's what you want or not depends on the shot (more about that above). The technique that I described that you're probably thinking of is spot-metering of the BRIGHTEST part of the sky (well brightest part of the entire scene actually, but that equates to the sky 9999 out of 10,000 times, hence the "generalization") and then up-shifting the exposure a couple of stops.

As a "case in point" (going back to exposure control / sky saturation), try taking a shot of someone in the foreground with clear sky in the background - then go manual and decrease the exposure by 3 stops and see what it does to the sky. THEN pick up the under-exposure of the foreground subject with a flash and "viola" you've turned a daytime shot into a moody "end of day" / "twilight" shot (with great sky saturation!)

[BZChi7d]

you guys didn't understand something. and i didn't at first either!

first of all, you can get meter reading from the sky even though you don't actually include the sky in the composition - it's all about 18% reflectance. You can meter blue sky, green grass or an actual grey card - they are all close to 18% reflectance and based on the the camera calculates the incident light. i have read this (that green and blue fall on the same category, a certain blue and green to be precise) here on cambridgetutorials. He uses center-weighted 99% of the time (he mentioned spot metering just a couple of times, eg. a machine in a field of coal).

I understand his using the sky to get his meter readin. What i don't understand is his f2.8 when eventually using f16 and the required math in between.

i'll scan the page tomorrow and let you guys have a look!

[Dave Humphries]

Hmmm, I think we were missing far too much context here from what you have said now Radu - for example, the sky I was assuming was one with bright white clouds in, if clear blue sky, then yes, it may be far closer to 18% and not require the 2 stops offset I mentioned. Presumably you can see the picture and I cannot.

A little knowledge is a dangerous thing ...
(meaning we all dived in without the whole story)

[BZChi7d]

full size here

second picture (down) explanation on the left

[Colin Southern]

In a nutshell ...

- Metering off the sky is a valid enough technique (if you want to expose it as a mid-tone) (snow will often throw metering off), but

- Adjusting the aperture / stopping down seems pointless to me - I'd have just metered the scene at the final aperture (F16) in the first place.

It's not a very very old book by any change is it? I'm just wondering if he wasn't able to meter long exposures directly for some reason?