Lux/EV, light meters, camera point meters, questions

[Mr Whippy]

I'm currently trying to measure the intensity of things, from highlights on surfaces from the sun, the sky on an overcast day, to the clouds, the blue sky on a clear day, and the sun itself.

I have a few methods of doing this.

Light meter (dome type thingy, so wide angle scene intensity)
DSLR (d70s) (+ two equations)
iPhone + Light Meter app


The problem is, they are all giving wildly different numbers.

I trust the light meter the most since it is a purpose made device, but the SLR may not be doing a wide angle metering with the same rationalised method as the light meter, so I can't get numbers looking the same there, but also the equation I use from cam settings > Lux causes a big discrepancy.

The iPhone is point meter only, and it appears to use the 269*f.sq / iso * shutter speed = Lux

I have an alternative method for getting Lux which is 50*f.sq / iso * shutter speed

Clearly one gives much bigger numbers than the other, but the difference appears to be about 2 stops, which is what I seem to find is the amount of stops to get from metered 18% grey or around that, to white... so maybe the SLR/iPhone need to compensate to white to get the real intensity, to match the light meter!?

Is this maybe where the discrepancy is coming from? Two stops up on the iPhone app seems to get me near the proper light meter readings, and + 2.0 on the SLR gets me near the light meter readings too... (with the 269 coefficient equation)


I guess I have kinda answered my own question, so I'm mainly checking that my logic of 2 tops up from ~ 18% grey gets me white, which is probably what my light meter is exposing for when measuring Lux!?

Many thanks for any help/advice!

Dave

[Simpo Two]

The disparity may well be due to the area that's being metered and how the electronics grinds it up. For example a D70 can meter using either matrix, centre weighted or spot, and IIRC the area of the last two can be adjusted in the menu.

The devices you're using may also have differing sensitivities to wavelength, eg a blue sky.

And yes, I'd say that +2.0 EV generally gets you from 18% grey to white - I dial this in automatically when photographing white objects, sometimes more.

Anything more technical and you'll have to wait for somebody else to turn up!

[Elderly]

Mr Whippy said:

I'm currently trying to measure the intensity of things, from highlights on surfaces from the sun, the sky on an overcast day, to the clouds, the blue sky on a clear day, and the sun itself.

I have a few methods of doing this.

Light meter (dome type thingy, so wide angle scene intensity)
DSLR (d70s) (+ two equations)


Dave

I'm not sure that I understand what you are trying to achieve
(that's my problem!).

You mention a "dome type thingy"; can I assume that with 'the dome', you are measuring incident light? With the DSLR you will be measuring reflected light.

Is this where your discepancy is?

[Simpo Two]

EFA:

Simpo Two said:

Anything more technical and you'll have to wait for Elderly or LongQ to turn up!

[Mr Whippy]

Elderly said:

Mr Whippy said:

I'm currently trying to measure the intensity of things, from highlights on surfaces from the sun, the sky on an overcast day, to the clouds, the blue sky on a clear day, and the sun itself.

I have a few methods of doing this.

Light meter (dome type thingy, so wide angle scene intensity)
DSLR (d70s) (+ two equations)


Dave

I'm not sure that I understand what you are trying to achieve
(that's my problem!).

You mention a "dome type thingy"; can I assume that with 'the dome', you are measuring incident light? With the DSLR you will be measuring reflected light.

Is this where your discepancy is?

Yes, the dome thing is an 'incident' light meter. It was £35 from Maplin.

I'm not quite sure what practical difference there is between incident and reflected measurement for what I am trying to do, which I will try explain better now.

Imagine the moon on a totally clear night. If I were to isolate just the moon, what lux is it? Now imagine the moon is reflected on a cars paintwork, how many lux are being reflected off the car at 90deg (perpendicular to the car), and at 45deg, and perhaps at a glancing angle (I guess it'll err towards a mirror)?
The incident light meter is clearly going to struggle here, since it appears to be influenced over a very wide field of view (probably approaching 180deg)


The problem is, even in an environment such as a room lit with a single light bulb, pointing the incident meter at the wall fairly close up (but not shadowing the metering area significantly), the lux level of light bouncing off the wall is quite different to that suggested by the equations given by using settings determined by my DSLR and iPhone metering app with Lux readout (when not correcting for a cameras default need to exposure for ~ 15% or so grey)
If I imagine I need to expose for white, then suddenly moving 2 stops on these devices give numbers similar to what the incident meter is giving in the relatively controlled 'wall' test.


My light meter purchase was to try check my SLR and iPhone results, however it's ended up with me having more questions and problems

I'm mainly wondering if, to determine lux, I need to expose for white with the SLR on the object in question, rather than grey... or if the 2 stop movement used to match the proper light meter readings is just fluke and there is something else going on here... ie, me totally mis-understanding what is going on


Hmmm

Dave

Edited by Mr Whippy on Wednesday 6th October 20:31

[Elderly]

Mr Whippy said:

........ the difference appears to be about 2 stops, which is what I seem to find is the amount of stops to get from metered 18% grey or around that, to white...

Dave

When measuring exposure for taking photographs, (I get impression you are carrying out some kind of physics experiment ) your neutral test card only gives a reference area of known reflectance when it's held straight up and down close to and in front of the subject, and facing half way between the camera and the main light.
That reflectance is 18% for the grey side (as you allude to) and 90% for the white side.

I wonder if that's around 2 stops?

[Mr Whippy]

Hmmmm.

I'll do some more testing, and see what happens with my numbers.


It's not really for a physics experiment. There is a car simulation/game that I enjoy creating content for. It's in constant development, and recently it implemented a full HDR environment/material system... so knowing how intense different things are under different lighting conditions is important for it to look real.

I think the hardest bit to get right is the mapping of the HDR info to a limited dynamic range display, that also feels like the human eye view, rather than a more limited dynamic range camera exposure. BUT, you have to start at the start and get it right to get the steps down the line correct


Thanks

Dave

[Major Bloodnok]

Mr Whippy said:

Elderly said:

Mr Whippy said:

I'm currently trying to measure the intensity of things, from highlights on surfaces from the sun, the sky on an overcast day, to the clouds, the blue sky on a clear day, and the sun itself.

I have a few methods of doing this.

Light meter (dome type thingy, so wide angle scene intensity)
DSLR (d70s) (+ two equations)


Dave

I'm not sure that I understand what you are trying to achieve
(that's my problem!).

You mention a "dome type thingy"; can I assume that with 'the dome', you are measuring incident light? With the DSLR you will be measuring reflected light.

Is this where your discepancy is?

Yes, the dome thing is an 'incident' light meter. It was £35 from Maplin.

I'm not quite sure what practical difference there is between incident and reflected measurement for what I am trying to do, which I will try explain better now.

Imagine the moon on a totally clear night. If I were to isolate just the moon, what lux is it? Now imagine the moon is reflected on a cars paintwork, how many lux are being reflected off the car at 90deg (perpendicular to the car), and at 45deg, and perhaps at a glancing angle (I guess it'll err towards a mirror)?
The incident light meter is clearly going to struggle here, since it appears to be influenced over a very wide field of view (probably approaching 180deg)


The problem is, even in an environment such as a room lit with a single light bulb, pointing the incident meter at the wall fairly close up (but not shadowing the metering area significantly), the lux level of light bouncing off the wall is quite different to that suggested by the equations given by using settings determined by my DSLR and iPhone metering app with Lux readout (when not correcting for a cameras default need to exposure for ~ 15% or so grey)
If I imagine I need to expose for white, then suddenly moving 2 stops on these devices give numbers similar to what the incident meter is giving in the relatively controlled 'wall' test.


My light meter purchase was to try check my SLR and iPhone results, however it's ended up with me having more questions and problems

I'm mainly wondering if, to determine lux, I need to expose for white with the SLR on the object in question, rather than grey... or if the 2 stop movement used to match the proper light meter readings is just fluke and there is something else going on here... ie, me totally mis-understanding what is going on


Hmmm

Dave

Edited by Mr Whippy on Wednesday 6th October 20:31

I'm not sure why you're using an incident light meter for these measurements. An incident light meter is designed to measure only the light falling on a scene - as such, it is independent of the reflectivity of any objects in the scene - basically it's designed to give you an exposure that would be correct for an 18% grey card at that position under those lighting conditions. Also, there are two types of incident meter: the 'flat' type that you point at the light source, and the dome type that you point toward the camera (in both cases, readings are taken from the subject position). Pointing them in any other direction will give a false reading.

From your moon description, it sounds as if you need a narrow-angle (1 degree) spot meter. Take a reading from the moon, then take one from the highlight and note the difference (actually, even that's a bit iffy, since the moon is only half a degree wide). AFAICS, this is the only method that will give you what you want - everything else will be affected by other things in the scene, and won't give reliable results.