Using a light meter on the job

mangaloreaviators wrote in post #5725307
Ok Doc. Does anybody agree with Doc ?

Yup. Sure do.

The meters I have used over the years have been far simpler to use - far more accurate than guessing and less stuff to mess with.

The only thing the chart might be good for, as far as I am concerned, is determining which of two meters is wrong if there's a significant difference in the readings on them.

Ok Skip.

Will test the chart in the next few days as it has worked for me in some situation. Let me do a detailed analysis.

Analyze this while you're at it: Need an exposure crutch?

Analysis will not cure the problem of being crippled as compared to the speed of a good modern meter.
You chart could sure be handy for the right person in the right situation.

In Tim's defense, his apparent need to 'tweak' the reading from an incident light meter is because he wants simply to exploit his electronic sensor to the fullest, and 'shoot to the right' of the histogram. Shooting to the right means that you are NOT shooting to the inherent brightness of the light falling on the scene, but shooting to the brightness of the light reflected back from the scene. The incident meter reads the inherent light falling on the scene and ignores the brightness of the objects being photographed.

Plain and simple.

Wilt wrote in post #5733365
In Tim's defense, his apparent need to 'tweak' the reading from an incident light meter is because he wants simply to exploit his electronic sensor to the fullest, and 'shoot to the right' of the histogram. Shooting to the right means that you are NOT shooting to the inherent brightness of the light falling on the scene, but shooting to the brightness of the light reflected back from the scene. The incident meter reads the inherent light falling on the scene and ignores the brightness of the objects being photographed.

Plain and simple.

That's not what the ISO standards say.

DocFrankenstein wrote in post #5733378
That's not what the ISO standards say.

Care to elaborate what the ISO standards do say?

I can take an incident reading and it exactly mimics what a gray card in front of my 40D spot meter tells me to shoot at. The readings are the same because the subject brightness matches the 18% gray which is assumed by an incident meter reading the light intensity. I might decide (or not) to shoot to the right with either reading, ending with the same result after peeking at the histogram as recorded.

Wilt wrote in post #5733389
Care to elaborate what the ISO standards do say?

Right after you support your "it's that simple" statement with primary literature.

Wilt, I agree that this is one of the things that I love about an incident meter..... either old fashioned selenium powered analog needle type, or new battery powered digital.

It's all about the OVERALL QUANTITY of light falling on a scene. It's not about how much or how little is being reflected back by a reflected reading off a bride's white dress, or the guy's black tux.

I did a wedding a couple of weeks ago - horrible bright noon sun right out in the open, white dress, canopy shade, black suits.... metering nightmare. I knew I wanted a starting ISO of 200 so that part was easy, along with 1/200 to freeze any action and allow a high flash synch. So I set things to TV. I’d also taken an incident reading at a spot that would match “center stage” just in case I couldn’t get a camera meter reading that I’d be happy with.

Right away the camera's metering system began jumping around as it saw the dress, the suits, the under-canopy shade... it was responding instantly to all the different environments. This was no good. I immediately switched to Manual on the camera.

I could have simply used the Sunny 16 rule, but I'm like the guy who wears suspenders with a belt for added security. For this reason I always have my 1970's Sekonic L-398 hand held meter with the incident dome. No batteries, always reliable.

Once I got the initial SUNLIT reading with the Sekonic, I knew everything would be constant from that point on.

If I zoomed into the shade under the canopy it was just a minor matter to open up a couple of stops. If I wanted to change ISO which I did a couple of times, it was easy enough to recalculate shutter/aperture in my head since I always go in doubling multiples of my ISO. I'm a simple person, and this helps keep my exposures simple.

Here's the shot I was talking about. I was basically metering for the bright sun, but ACR helped me nicely recover the RAW highlights and pump a little exposure back into the shadows. There was bit of PP for this quick 4x6 proof, but it got the print in the ballpark. If she orders an enlargement then I’ll do some finer post processing:

Wilt wrote in post #5733389
Care to elaborate what the ISO standards do say?

I can take an incident reading and it exactly mimics what a gray card in front of my 40D spot meter tells me to shoot at. The readings are the same because the subject brightness matches the 18% gray which is assumed by an incident meter reading the light intensity. I might decide (or not) to shoot to the right with either reading, ending with the same result after peeking at the histogram as recorded.

HOSTED PHOTO
please log in to view hosted photos in full size.

DocFrankenstein wrote in post #5733410
Right after you support your "it's that simple" statement with primary literature.

This is quite interesting, I didn't know that there was disagreement on the shooting to the right concept. As I understand it, because the sensor is logarithmic, but the digital output is linear the last stop holds 50% of the total information, so by shooting to the right you capture the maximum data.

But, I don't know of any actual analysis to show how much this improves the picture after you've essentially re-exposed down in post to get the proper exposure/brightness.

It would be easier to essentially shoot for the middle, which is normal proper exposure is, isn't it?

Stinger wrote in post #5733809
This is quite interesting, I didn't know that there was disagreement on the shooting to the right concept. As I understand it, because the sensor is logarithmic, but the digital output is linear the last stop holds 50% of the total information, so by shooting to the right you capture the maximum data.

But, I don't know of any actual analysis to show how much this improves the picture after you've essentially re-exposed down in post to get the proper exposure/brightness.

It would be easier to essentially shoot for the middle, which is normal proper exposure is, isn't it?

I don't push to the right for two reasons:
1) The displays are not big enough to show the histograms accurately
2) The camera shows a JPEG histogram, which is completely different from a RAW histogram
3) There's hardly any room for error with digital, because you lose info in separate channels if you overexpose it.
4) I don't like messing with exposure for more than one shot.

Both the meter and the sensor have been very carefully calibrated. If you know where to point the meter to get the exact effect, you just can't get any more precise than just plucking what the meter tells you right into the camera.

90% of the images I've seen from people who "shoot to the right" end up with blown channels.

Interesting question Stinger... even though I've used an incident meter for a few decades, I never really went much into the math of it. I just look at the cause and effect for my results and "tweaked" accordingly.

I wonder if an incident reading is essentially "shooting for the middle?"

But in reality I was shooting a bit for the highlights. Even though I did a fairly good job of saving the bride's highlights, I know from experience that some detail was lost. So when I did these actually exposures, I DID use my incident reading, but I also fudged a bit to help save those bright pixels.

Hmmmmm..... I don't know; what do you think Doc?

Stinger wrote in post #5733809
.....It would be easier to essentially shoot for the middle, which is normal proper exposure is, isn't it?

I tend to use the light meter mainly in the church, I take a reading then do a test shot. Often the RGB histogram doesn't reach the right side, so I increase exposure by a stop. Later I sometimes get blown channels, but easily recoverable (especially with the 40D HTP on).

sapearl wrote in post #5733979
Interesting question Stinger... even though I've used an incident meter for a few decades, I never really went much into the math of it. I just look at the cause and effect for my results and "tweaked" accordingly.

I wonder if an incident reading is essentially "shooting for the middle?"

But in reality I was shooting a bit for the highlights. Even though I did a fairly good job of saving the bride's highlights, I know from experience that some detail was lost. So when I did these actually exposures, I DID use my incident reading, but I also fudged a bit to help save those bright pixels.

Hmmmmm..... I don't know; what do you think Doc?

Middle... or 18% or 12% gray is the film term and simply is an anachronism in the digital era. That's what the reflective meters were calibrated for by the german standards of 50 years ago.

With film you have two "shoulders" - the highlights and the shadows. There's a (somewhat) linear portion of the film with the "grey" area in the middle, and that's what most cameras aimed for.

That convention was made simply because you can never tell just how much highlight detail you get with film until you know what film it is and how it was developed.

The incident meters are calibrated differently. There's no "gray" - they work on saturation. Here's the standard:
When you take an incident reading of an object with 100% reflectance, it registers on the film as 90% grey on the linear scale.

Well, basically to guarantee that the highlights are preserved, you'd have to point the meter at the sun - not at the camera. Then the 100% white bridal dress will register as 90% grey on the linear scale.

Note that the scale is linear, so instead of losing a "whole stop" of data you'd only have abour 401 unused pixels. But not a single pixel would've been blown AND unless you added some fill (20% of ambient minimum) and re-metered, you'd have pretty dark faces.