The Inverse Square law and photography

[Matt]

I take it this is 6db on the A scale? I think that that would follow the ISL.
Too technical for me, I know that those particular terms have relevance in specialist lighting and scientific applications, but they're outside of my own knowledge. But, regardless of terminology, it seems that the interference to the theoretical ISL figures in acoustics is similar to that of light - unwanted reflections.

Of course, in light, reflections aren't always unwanted - it's just that they interfere (sometimes dramatically) with the control of light

its all wave theory, light waves just operate on a different frequency range, I cant see that it would be that different.
In acoustics, we want some reflections and when designing a room we build the room to take into account the nearfield and far field interactions with the listening position, the difference though is that acoustically we can calculate very accurately what reflections there will be at a given point in the room, im not sure that you can do that as easily with light - i suppose you could do the same thing as we do for rooms, each surface will have a certain degree of light reflectance in a given frequency, so diffuse reflections, scattering, absorption etc will all be an attribute of a given surface. when designing a studio, you would pick the surfaces that will give you the best light, and adding things like reflectors will help to control the light in much the same way as you do with diffusors and absorbers in acoustics.

 

[Garry Edwards]

Yes, we do do that anyway. There are always going to be people who either paint the ceiling and walls white because they don't know any better, or because they like the room to look cheerful, or because they don't have any choice because the room has to be used as a living room too, but dedicated studios are usually either dark grey or preferably black painted with high ceilings, to avoid unwanted reflections. Reflectors are then added as required, for specific purposes.

People who don't have any choice about white/light walls ceiling can use black flags to reduce unwanted reflections.

 

[arad85]

Re: light and sound waves... :nono: You're mixing two different things. Sound wavelength is measured in metres, so any change in intensity (which can be large) due to room interactions is based on the additive properties of pressure waves and your position in the room relative to the fundamental frequencies of the room. Light, measured in nanometres, bounced off a surface will only add to the light at any point as the human has no ability to perceive or control movements at the nanometre scale. (i.e. the variations in intensity due to the wavelength are proportional to the wavelength)

Sorry if I'm off topic/missing the point, but it was something I saw raised in the posts above...

 

[paul williams]

Re: light and sound waves... :nono: You're mixing two different things. Sound wavelength is measured in metres, so any change in intensity (which can be large) due to room interactions is based on the additive properties of pressure waves and your position in the room relative to the fundamental frequencies of the room. Light, measured in nanometres, bounced off a surface will only add to the light at any point as the human has no ability to perceive or control movements at the nanometre scale. (i.e. the variations in intensity due to the wavelength are proportional to the wavelength)

Sorry if I'm off topic/missing the point, but it was something I saw raised in the posts above...

Yes, come on now, don't start adding another even more complex topic to one this simple country boy has trouble following already



Paul

 

[Garry Edwards]

Re: light and sound waves... :nono: You're mixing two different things. Sound wavelength is measured in metres, so any change in intensity (which can be large) due to room interactions is based on the additive properties of pressure waves and your position in the room relative to the fundamental frequencies of the room. Light, measured in nanometres, bounced off a surface will only add to the light at any point as the human has no ability to perceive or control movements at the nanometre scale. (i.e. the variations in intensity due to the wavelength are proportional to the wavelength)

Sorry if I'm off topic/missing the point, but it was something I saw raised in the posts above...

Fair comment, and I agree that we're wandering off topic, but there are similarities, as I remember from the days when my learning was theoretical...
Because we humans can only hear sounds within a limited range, so when frequencies change and go outside this range the sound may as well not exist as far as we are concerned because we can't hear it - or something like that.

Well, that applies to light too, We can only see a very limited range of wavelengths, and if light is outside of them it may as well not exist as far as we are concerned.

So, sticking strictly with light reflections, my main point in all this, to simply it as far as possible, is that reflected light, plus light from other sources, can contaminate the result of the ISL to a very high degree

 

[arad85]

Well, that applies to light too, We can only see a very limited range of wavelengths, and if light is outside of them it may as well not exist as far as we are concerned.

That's true. But it may not be true for the sensor in the light meter!

So, sticking strictly with light reflections, my main point in all this, to simply it as far as possible, is that reflected light, plus light from other sources, can contaminate the result of the ISL to a very high degree

It can contaminate it, yes, but then that should be obvious from the effects of using reflectors. My point was that as far as the equations go, you can just be additive for light - you don't need to worry about how the phase of the signal changes depending on your position like you do with acoustics.

 

[Triggaaar]

Blimey, this thread's gone off on one.

But I'm not relying on just the large softbox to illustrate my point. Tests done in exactly the same testing environment at the same time with a reflector showed the following results:
At 0.5 m f/22 (benchmark)
At 1 m f/11 d6, an error of .4 stops
At 2 m f/5.6
At 4 m f/4 d3, an error of 1.3 stops

I'm still intrigued by these measurements. What size is the reflector, something like 20cm - and does it produce a very narrow beam of light?

 

[Garry Edwards]

Yes, a fairly narrow beam of light, the diameter is 29cm

 

[Triggaaar]

I'm not suggesting that our lights accurately follow the inverse square law, but those measurements seem very odd to me, and I can't explain them without assuming you've got reflections all over the place. Do those measurements not surprise you Garry?

 

[Garry Edwards]

No, they don't surprise me - fairly typical I'd say.

My studio is fairly large and within the measured area the ceiling is painted black, as is one wall. The other wall is painted white, but is some distance away and the beam of light will have missed nearly all of it. Reflections off of the wall that the light actually reached aren't relevant - far too far away and dark paint. There is an area of white painted ceiling but it's far beyond any of the measuring points, and there is no apparent reflection route back to the flashmeter sensor. It's a commercial studio, used for largely still life photography, so reflections are tightly controlled.

Edit: Obviously, with the softbox test, reflections will have played a part because some light will have been reflected from the white painted wall at the side, but far less than in most studios.

 

[Matt]

Re: light and sound waves... :nono: You're mixing two different things. Sound wavelength is measured in metres, so any change in intensity (which can be large) due to room interactions is based on the additive properties of pressure waves and your position in the room relative to the fundamental frequencies of the room. Light, measured in nanometres, bounced off a surface will only add to the light at any point as the human has no ability to perceive or control movements at the nanometre scale. (i.e. the variations in intensity due to the wavelength are proportional to the wavelength)

Sorry if I'm off topic/missing the point, but it was something I saw raised in the posts above...

i wasnt saying sound and light = the same, im using my knowledge of sound and ISL to try and rationalise why there is a difference in what is predicted and what is measured...the same principles of near field and far field seem to be interfering with the light in the same way as it does with sound, or is it something different that is happening.

Point source could be regarded as Line source in some instances also (strip light anyone?) so there ARE simularities, we just need to understand what the lighting equivelants are..does that make more sense?

 

[arad85]

.the same principles of near field and far field seem to be interfering with the light in the same way as it does with sound, or is it something different that is happening.

Something different is happening. Near/far field for audio is metres. Near/far for light is nanometres.

 

[Matt]

T

It can contaminate it, yes, but then that should be obvious from the effects of using reflectors. My point was that as far as the equations go, you can just be additive for light - you don't need to worry about how the phase of the signal changes depending on your position like you do with acoustics.

i havent mentioned phase - thats even more complicated, im just talking about why ISL doesnt work in acoustics and trying to tie that with ISL with light

 

[Matt]

Something different is happening. Near/far field for audio is metres. Near/far for light is nanometres.

but what is happening? ambient levels of light?

 

[Garry Edwards]

but what is happening? ambient levels of light?

Ambient lighting certainly skews the figures where continuous light is used as the light source, but nearly everyone uses flash, and unless either an unusually long shutter speed is used or the ambient light is at an extremely high level (e.g. sunlight streaming through a window) then it doesn't matter.

My own tests were carried out in my studio, with no natural light, just fairly low level fluorescent overhead lighting, and the shutter speed set on the flash meter was 1/25oth, so ambient light will have been irrelevant.

Leaving aside the skewed results from large light sources such as softboxes, the biggest single factor in producing results very different from the expected ISL results is the pollution caused by other photographic lights and the pollution caused by reflected light assuming that the source light isn't controlled to such an extent that the light isn't allowed to quadruple its coverage every time it doubles its distance.

 

[HoppyUK]

Isn't the big difference between light and sound the fact that light travels through the atmosphere virtually unimpeded (unless you work in a fog) whereas sound actually has to move it? It's 'mechanical' in that sense, light is electro-magnetic radiation.

And further, high frequency sounds are absorbed and reduced quickly by this and it's only the lower frequencies that have substantial energy and can get through at any significant distance.

And, back on topic, it still doesn't change the fact that Garry is making spurious point about the ISL - it works with light, and it works with studio lights and hot-shoe guns. The fact that it's effect is reduced slightly at close range with a softbox - like 25%, quarter of a stop, which is next to nothing in exposure terms - is worth noting, but nothing more.

Or to put that another way, how is that knowledge going to materially affect the way we work? Light fall-off over distance is always massive whatever light source you use (lasers and fresnels apart) and it's the ISL which is driving that. Furthermore, in the close working environment of a studio, very small changes of distance have big effects - if you move a light back just a foot or two you can easily lose a stop, and that does matter. Before we all had multi-adustable power flash heads, or tungsten back in the day, that's how exposure was adjusted.

ISL is a fundamental underlying principle of all studio and artifical lighting. To even suggest that it in some way the ISL doesn't apply to photography, or even that it's impact is ever anything less than very substantial, is not helpful IMO. To use compounded figures, as Garry is doing to prop up a lost cause, and then mis-quote linked videos is frankly cuckoo.

Yours in love and peace

 

[Garry Edwards]

Richard,

it still doesn't change the fact that Garry is making spurious point about the ISL - it works with light, and it works with studio lights and hot-shoe guns.

I haven't made an spurious claims. What I have done is to carry out some tests and suggest that other people may want to carry out their own tests, and not to just assume that the ISL will apply to whatever type of lighting they are using, regardless of the type of light or the environment.

The fact that it's effect is reduced slightly at close range with a softbox - like 25%, quarter of a stop, which is next to nothing in exposure terms - is worth noting, but nothing more.

My particular test with the softbox I used showed a divergence, not of 25%, but of far more. My first reading was taken at 0.5 meter - fairly typical for many studio lighting situations, and it produced a reading of f/22. At 1 meter the reading *should* have been f/11 but in fact was f/16 d1, a loss of only 0.9 stop instead of the *expected* 2 stops, at 2 meters the reading was f/8 d4, a loss of only 2.6 stops instead of the *expected* 4 stops, and at 4 metres the reading was f/4 d8, a loss of only 5.2 stops instead of the *expected* 6 stops - what's so difficult to understand about that?

And this was in my own studio, studios vary widely in size and decoration, I think even you will agree that a small, white painted studio would be expected to show much greater deviation than my larger, black painted studio!

To use compounded figures, as Garry is doing to prop up a lost cause, and then mis-quote linked videos is frankly cuckoo.

Not compounded figures, just the figures that the ISL says should apply. And that video clearly referred to the ISL, in that very section. Why mention it if he didn't think it was relevant?

Yours in love and peace

Yeah, right - I put my avatar up as a joke, but maybe some people do like to snipe at me - and miss

 

[Matt]

Isn't the big difference between light and sound the fact that light travels through the atmosphere virtually unimpeded (unless you work in a fog) whereas sound actually has to move it? It's 'mechanical' in that sense, light is electro-magnetic radiation.

And further, high frequency sounds are absorbed and reduced quickly by this and it's only the lower frequencies that have substantial energy and can get through at any significant distance.

Yours in love and peace

ok as i didnt make my point very well..remove all references to acoustics and start again

ISL doesnt work as it should do as other interactions and factors interfer with it - intensity of source is reduced as per ISL, BUT external factors interfer. soundwaves are absorbed by the action of friction as they move through the air, light is not, so light bouncing around the place IS going to impact the light reading at a given distance

 

[HoppyUK]

Garry, my point is about relevance. The ISL is often quoted because, with very few exceptions, it applies - and in practical terms it actually applies very closely to the absolute law. Ignore it at your peril (which I know you don't, as you cite it yourself often enough).

1) You are testing at 0.5m - how many people on here are really using a softbox at that distance?
2) When I do that test with a 70x70 softbox, even from 0.5m to 1m, I get a loss of 1.5 stops - much less reduction that you are claiming. Worth noting, but in no way undermining the fundamental application of the ISL.
3) Your figures for greater distances are compounded, so any deviation from absolute is not only increased, but multiplied - it's just manipulating the figures to prop up your argument. And comparing distances like 0.5m and then 4m is just not relevant in practise. What's the point of that? Nobody does it.
4) When I do the same test with a hot shoe gun, I get absolute parity with the ISL.
5) Another poster also gets parity with a snoot.
6) The guy in the video you linked to - watch it again. What he says is a) true, and b) nothing to do with exposure (as I already explained). What's your point? Where's the relevance?

Sorry if you feel victimised, but you don't need a fancy avator to get shot at

 

[arad85]

@Hoppy: light/sound. Yes, they propagate differently, but still are modelled as waves, so can have the same maths applied.

Of course the ISL works. But it only truly works for point sources. A softbox at 0.5 and then 1m is hardly a point source. It is partially collimated, so you wouldn't expect it to act anything like a point source at close distances. The further away from it you get, the more it acts like a point source, so getting closer and closer to the true ISL. I think that's what Garry is saying - notice the further away the softbox is, the closer to the ISL expectations you are.

Are we all actually agreeing and just saying the same things just differently here?

 

[Garry Edwards]

1) You are testing at 0.5m - how many people on here are really using a softbox at that distance?

Well, whether or not you use a softbox at that distance, a lot of people do - especially for product photography. In fact they are frequently used much closer than that. It may be that a lot of people don't use softboxes at that distance, but that doesn't make my test irrelevant.

Your figures for greater distances are compounded, so any deviation from absolute is not only increased, but multiplied - it's just manipulating the figures to prop up your argument. And comparing distances like 0.5m and then 4m is just not relevant in practise. What's the point of that? Nobody does it.

No, the figures are NOT compounded, the first reading is taken as the reference and all the other readings are relative to that. If the figures were only relative to the last reading (not the first) then they would be misleading, and I have no wish to mislead people.

Sorry if you feel victimised, but you don't need a fancy avator to get shot at

Other people have either agreed or disagreed, which is fine. Your own responses seem to me to be much more personal - not that it bothers me if you choose to question my motives, because I know that my motive here has been to ask people to question and test for themselves, you seem to be the only regular contributor who seems to be critical of my motives - which you don't know.

Anyway, "ignore member" button now clicked, so that's the end of it as far as I'm concerned. I will however respond to any questions/discussion points from other people who raise them.

 

[HoppyUK]

@Hoppy: light/sound. Yes, they propagate differently, but still are modelled as waves, so can have the same maths applied.

Of course the ISL works. But it only truly works for point sources. A softbox at 0.5 and then 1m is hardly a point source. It is partially collimated, so you wouldn't expect it to act anything like a point source at close distances. The further away from it you get, the more it acts like a point source, so getting closer and closer to the true ISL. I think that's what Garry is saying - notice the further away the softbox is, the closer to the ISL expectations you are.

Are we all actually agreeing and just saying the same things just differently here?

Kind of

Where is am disagreeing is in the significance and relevance of Garry's OP. As I said in my first post on this thread. The ISL applies, and in practical terms it applies pretty closely, to most kinds of light, most of the time.

Sure softboxes are a bit different, because they're not point light sources and that makes a difference at close range. But it's not a massive difference in relative terms and if there is a slight reduction in fall-off over the strict application of the ISL, then regard that as a bit of a bonus (usually ).

TBH, the biggest variance to true ISL parity, in real terms, is with a shoot-through brolly - because they are big, spread light everywhere, and are also often used in highly reflective environments, ie indoors. They spread light around across almost 180 degrees from of the front, then half of of it is bounced straight out of the back (more than half actually - they're very wasteful of light). So in effect, the whole room becomes the light source!

This is both a great asset and a problem. They're very easy to use, not so critical of positioning, give nice soft light with quite a lot of 'auto fill-in' in an average room - rightly popular for that. But if you want control, with light just on the subject and nowhere else, next to useless.

 

[HoppyUK]

Garry, I may be a bit blunt but my comments are absolutely not personal and you should read what I have actually said. If you make contentious posts and then link to irrelevant websites in support of your argument, you should not be surprised if you get criticised. I don't know where I have done anything else here.

As an underlying issue, I have also been critical of your and other's posts re the promotion of Lencarta on this forum, which I feel have overstepped the mark from time to time. I'm not alone in that, and it's not personal either, just a matter of principle. We may agree or disagree on that too which is fine, and the mods have taken a view which I am entirely happy with, but meanwhile, I continue to support and recommend Lencarta, when appropriate, as I always have done.

I am not questioning your motives here at all. It may not bother you but it does trouble me that you have interpreted things in this way

 

[Triggaaar]

I'm getting a few negative comments on this thread (I'm sure that they're not intended to be negative)

Yeah, right - I put my avatar up as a joke, but maybe some people do like to snipe at me - and miss

Anyway, "ignore member" button now clicked, so that's the end of it as far as I'm concerned. I will however respond to any questions/discussion points from other people who raise them.

Wo, what's this all about? You've made some interesting points and sparked a debate about an important law on lighting, and some of us disagree with the odd point you've made.

Other people have either agreed or disagreed, which is fine. Your own responses seem to me to be much more personal

I personaly have nothing against you, I've met you at focus, bought products from you, and I read a lot of your threads (like this one) with interest. As I have already said I disagree with your statement: 'It needs to be suspended because the light needs to be able to travel in all directions, otherwise it cannot quadruple its area of coverage each time it doubles its distance.' because a light doesn't have to be suspended for the light to be able to travel up/down, left/right, which is all that is needed for it to be able to quadruple its area each time its distance doubles (a point light source placed at the end of an absorbing cone would be fine).
so when you say to someone else "it's good to come across someone who understands physics" it's not surprising if others involved in the discussion take offence.

Anyway, to the topic:

Well, whether or not you use a softbox at that distance, a lot of people do - especially for product photography. In fact they are frequently used much closer than that. It may be that a lot of people don't use softboxes at that distance, but that doesn't make my test irrelevant.

We agree that in order for the law to work, the area lit has to quadruple every time the distance doubles. The main things that stop the happening are:
1) reflections - we all agree that these can interfere with the results
2) the shape of the light - if a modifier creates a beam of light comparable in size to the source, the area lit will not increase as much over distance
3) the size of the light source in relation to the distances where the measurements are taken.

So if you have a 4 foot softbox and take measurements at 10cm and 20cm, you can throw the law out of the window for that measurement. The source is too big in relation to measurement distances. That's an easier point to put across to fellow togs that you're trying to help - the ISL has its merits, but don't use it for close distances with large light sources.

 

[Kaouthia]

Well, you've identified the crux of it, the main reason why the inverse square law doesn't apply in the way that people assume it to - if the direction of the light is controlled in any way then it can't disperse freely and so can't quadruple the area covered each time it doubles its distance, leaving only a quarter of the power at any given point.

Funny, when I made a comment to that effect on here last year, you had a completely different response toward me.

 

[Garry Edwards]

Mike,

I have absolutely no problem with people disagreeing with me - the whole point of the thread in the first place was to get people to think and test, rather than blindly accept that the ISL always applies to photographic lighting, and obviously different people will have different views - and as different people use different types of equipment in different environments then they will have different experiences too.

My problem is with Richard, not because he happens to disagree but because he appears to be questioning my motives, and therefore my integrity.

We agree that in order for the law to work, the area lit has to quadruple every time the distance doubles. The main things that stop the happening are:
1) reflections - we all agree that these can interfere with the results
2) the shape of the light - if a modifier creates a beam of light comparable in size to the source, the area lit will not increase as much over distance
3) the size of the light source in relation to the distances where the measurements are taken.

These are exactly the points I've been trying to make. It's all about relative size - even a hotshoe flash is a long, long way from being a point source of light if the distance is very short - and even something as big as the sun can become almost a point source when it's 93,000,000 miles away.

As I have already said I disagree with your statement: 'It needs to be suspended because the light needs to be able to travel in all directions, otherwise it cannot quadruple its area of coverage each time it doubles its distance.' because a light doesn't have to be suspended for the light to be able to travel up/down, left/right, which is all that is needed for it to be able to quadruple its area each time its distance doubles (a point light source placed at the end of an absorbing cone would be fine).

You have a valid point there, but my point is that a large number of the lighting tools we use don't spread light in that way, so the ISL doesn't follow its own rules.

So if you have a 4 foot softbox and take measurements at 10cm and 20cm, you can throw the law out of the window for that measurement. The source is too big in relation to measurement distances. That's an easier point to put across to fellow togs that you're trying to help - the ISL has its merits, but don't use it for close distances with large light sources.

That's a valid point too. I didn't test at those very close distances because, at those distances, the accurate positioning of the measuring device is absolutely crucial. Therefore, my closest measurement was at 50cm. There was a rule on the floor and a vertical spirit level leading up to the flash meter, to make sure that the meter was at the right distance.

Obviously the ISL can follow the forumulae, some of the time. A hotshoe flashgun at a long distance becomes almost a point source of light, so does a snoot, and as long as there are no measurable reflections then the results will be as predicted. The point I was trying to make is simply that most photographic lighting doesn't consist of a point source, and most studios reflect light that interferes with the formulae too.

 

[Garry Edwards]

Funny, when I made a comment to that effect on here last year, you had a completely different response toward me.

Which thread was this?

 

[HoppyUK]

Let's look at this another way, and ignore the ISL for a moment.

Take a torch and shine it on a wall. It will project a pool of light of a certain size. If you then move back to double the distance, the pool will be twice the width and twice the height, which is four times the area of light. Therefore, the brightness of that light is reduced by two stops.

A torch is not a point light source radiating in free space, as the ISL strictly dictates, but the result is the same and that is what drives the reduction in brightness in photography.

The effect becomes reduced when you use a large light source like a softbox, because at close range (which is how we generally use them) the light pool is not absolutely quadrupled, maybe only trebled, but the effect still applies in broadly similar terms and the exposure must be substantially adjusted accordingly.

 

[Triggaaar]

my point is that a large number of the lighting tools we use don't spread light in that way, so the ISL doesn't follow its own rules.

Yes, some modifiers will have more effect than others. I'm going to have to take a light with a standard reflector outside later (I'm not looking forward to it, I will be spotted), as I don't understand the results you posted earlier. I'll post my results later.

 

[Kaouthia]

Which thread was this?

I don't recall specifically at the moment, and I didn't argue the point too hard.

But, it seemed that people believed that how the inverse square law applies to a point light source also applied to modified directional lights, to the letter. While, mathematically it might (because a softbox is essentially just an infinite number of point light sources of a much lower intensity, all combining to converge at different points with varying brightness, which then may or may not conform to the ISL at different distances from the light source) those calculations would be infinitely complex and in a practical sense it often does not stick to the letter of the law because your subject is at multiple distances from many different "points" on the front of a softbox.

 

[Stoptime]

Practical inverse square law demo:

http://www.youtube.com/watch?v=JW3tT0L2gpc&feature=related

enjoy!

 

[Triggaaar]

Right then, lots of testing in the dark and here are my results:

Studio flash head
Standard reflector, 19cm diameter & 90cm octobox
Measurements taken from the edge of the reflector (away from the bulb)
Tape measure held against light meter, flash fired with other hand while end of tape touched flash (except 4m test, where I measured from a vertical pole back to the flash head)

1st test, reflector outside:
0.5m = f22 d0, f22 d1, f22 d0, f22 d0
1.0m = f11 d1, f11 d1, f11 d0, f11 d0
2.0m = f5.6 d1 (*4)
4.0m = f2.8 d1, f2.8 d2, f2.8 d1, f2.8 d2

2nd test, reflector inside studio (fair amount of white walls)
0.5m = f22 d0, f22 d0, f16 d9, f16 d9
1.0m = f11 d1, f11 d0, f11 d2, f11 d2
2.0m = f5.6 d2, f5.6 d2, f5.6 d3, f5.6 d2
4.0m = f2.8 d7, f2.8 d7, f2.8 d8, f2.8 d8

3rd (final) test, softbox outside
0.5m = f22 d0 (*2, got bored/embarrassed)
1.0m = f11 d6, f11 d5, f11 d6, f11 d5
2.0m = f8 d0, f8 d0, f5.6 d9, f8 d0
4.0m = f4 d0 (*4)

Conclusion:
test 1 - follows ISL very closely
test 2 - again follows ISL closely, although reflections affecting the results a lot at 4 metres
test 3 - from 2m to 4m follows ISL perfectly. At closer distances, the size of the softbox had an effect on the ratio of distance vs area of light. This would be more pronounced with a larger softbox and close distances to a model.

 

[Triggaaar]

Having shown that the ISL is alive and well, I'm now back to these odd (IMO) results:

Tests done in exactly the same testing environment at the same time with a reflector showed the following results:
At 0.5 m f/22 (benchmark)
At 1 m f/11 d6, an error of .4 stops
At 2 m f/5.6
At 4 m f/4 d3, an error of 1.3 stops

those measurements seem very odd to me, and I can't explain them without assuming you've got reflections all over the place. Do those measurements not surprise you Garry?

No, they don't surprise me - fairly typical I'd say.

So we're saying that the light intensity falls because the area lit is getting bigger (4 times bigger for 2 times distance according to ISL).

From 0.5m to 1.0m you've lost 1 1/2 stops of light, so the area lit has grown by about 3 times.
From 1.0m to 2.0m you've lost about 1 1/2 stops of light again, so again the area lit has grown by about 3 times.
Then suddenly, from 2.0m to 4.0m you've only lost 2/3rds of a stop, so the area lit has less than doubled.

If all measurements are accurate, and we don't allow for reflections, that makes no sense. If the area lit expands by 3 times, and then expands by 3 times again, it can't suddenly stop expanding so much and taper in.

Do you think I've missed something here Garry that would explain why your 4m measurement is brighter than I suggest it should be (my guess would be reflections but you say that's not likely).

Cheers

 

[HoppyUK]

Good work Mike Though I hesitate to mention that the key test, the softbox indoors, is missing.

However, my own tests are very much in line with yours and I didn't find that indoors with a softbox was any different to outdoors at close range. The envionment/spill etc is overwhelmed by the direct incident light anyway.

I think the critical numbers are the softbox between 1m and 2m, as that is how we generally use them, and a fall-off there of 1.5 stops seems quite typical. It doesn't follow the ISL exactly, but manages three-quarters of it - which is a very substantial reduction whichever way you look at it.

 

[Triggaaar]

Good work Mike Though I hesitate to mention that the key test, the softbox indoors, is missing.

I missed it deliberately, my studio is mostly white, so I figured it would just show I have light bouncing everywhere. Outside I kept the lighting stands high to reduce reflections from the ground, and I tried the reflector inside just to see how much different it was. I figured at 0.5m and 1m the walls wouldn't make a lot of difference since the beam was narrow and I was a few metres from the walls, and understandably the 4m test was most effeted.

However, my own tests are very much in line with yours and I didn't find that indoors with a softbox was any different to outdoors at close range. The envionment/spill etc is overwhelmed by the direct incident light anyway.

Yes, I guess I could test it, but I don't really need to in order to check whether ISL is working.

I think the critical numbers are the softbox between 1m and 2m, as that is how we generally use them, and a fall-off there of 1.5 stops seems quite typical. It doesn't follow the ISL exactly, but manages three-quarters of it.

I'm not sure we need to know it that accurately unless we hate using light meters, as long as we understand the principle. I've been using the principle to play with fill lights to mimmick Venture's setup.

 

[Kaouthia]

If all measurements are accurate, and we don't allow for reflections, that makes no sense. If the area lit expands by 3 times, and then expands by 3 times again, it can't suddenly stop expanding so much and taper in.

It's all due to the fact that a softbox is NOT a point light.

The light meter might be 0.5m away from the CENTER of the softbox, but it might be 0.6m away from the light emitting closer to the edges of the softbox - and your light meter is reading all the light emitting from the entire softbox surface area.

Once you move 1m away, those edges might now be 1.1 meter away. So while you've doubled the distance to the center of the softbox, you've NOT doubled the distance to the light emitting from the edges of the softbox (and all the points on the surface in between).

The further out you get towards the edge of the softbox, the less relative distance to the light meter compared to the center.

That's why the inverse square law isn't accurate to the letter of the law (especially at closer distances as the light source is larger relative to the distance from the center) once you start using softboxes and other light modifiers.

 

[HoppyUK]

It's all due to the fact that a softbox is NOT a point light.

The light meter might be 0.5m away from the CENTER of the softbox, but it might be 0.6m away from the light emitting closer to the edges of the softbox - and your light meter is reading all the light emitting from the entire softbox surface area.

Once you move 1m away, those edges might now be 1.1 meter away. So while you've doubled the distance to the center of the softbox, you've NOT doubled the distance to the light emitting from the edges of the softbox (and all the points on the surface in between).

The further out you get towards the edge of the softbox, the less relative distance to the light meter compared to the center.

That's why the inverse square law isn't accurate to the letter of the law (especially at closer distances as the light source is larger relative to the distance from the center) once you start using softboxes and other light modifiers.

I don't think that answers the question Trig was asking. He was questioning the apparent inconsitency of Garry's figures.

But that's not the point I want to make here, which is, that I don't think it's as simple as a softbox not being a point light source, or being a multitude of point light sources.

The difference in distance from centre to edge, at 1m distance with a 1m octagonal softbox is only about 10cm or 10% (as best you can measure these things) and half that percentage at 2m. But we're seeing a greater deviation from the ISL than that, and while the that difference isn't huge, I don't think that explains it. In fact, the more I ponder this, it seems that the whole 'point light source radiating in free space' thing is a bit of a red herring.

What drives the reduction in brightness with studio lights is the fact that when you double the distance, generally speaking the area of the light pool projected is quadrupled. Or at least is does normally, but that is clearly not quite the case with a softbox

What may be more significant is the internal design of the softbox, which has a degree of parabolic curve about it. In other words, the light is at least partially concentrated and 'projected'.

It would be interesting to test this with identical size softboxes in three different scenarious:
- a normal octagonal softbox
- same softbox with a grid
- and then a third test with a same size 'deep octa' softbox

Deep octa like this Elinchrom http://www.theflashcentre.com/elinchrom-rotalux-100cm-deep-octa-i3608.html which is a fabulous light and well known to provide a more tightly controlled light pool with slightly stronger shadows and far less spill to the sides, due to its parabolic shape.

Anybody got one? This makes more sense to me than simply saying a softbox is bigger and therefore the point light source theory doesn't apply.

 

[arad85]

What may be more significant is the internal design of the softbox, which has a degree of parabolic curve about it. In other words, the light is at least partially concentrated and 'projected'.

Both Garry and I mentioned the word collimated in posts above. This is exactly what collimated means (apologies if you know this, but it seems you didn't pick up on it - see: http://en.wikipedia.org/wiki/Collimated_light). The more modifications you put on it - i.e. adding grids etc to remove the spread of light at source, the less spreading out the radiation does as it travels, so the more consistent the light becomes. You may well have to put a lot of power into the flash to get enough light to light what you want, but the beam is partially collimated (or projected as you put it)

For example, if you put a 10 degree honeycomb over a light, the light cone will be a 10 degree light cone. You can think of that as a point source somewhere behind the flash. For a 7" flash with a 10 degree reflector, that equates to a point source almost a metre behind the honeycomb (from basic trig). Clearly, as you get tighter and tighter honeycombs, that effective point source moves further and further behind the flash head front. It is the distance from the effective point source that will dictate the doubling of area.

If someone wants to try this with a 7" reflector plus 10 deg honeycomb, the maths says the two stop points (i.e. quarter power) should be at 0.5m, 2m, 5m...

 

[Garry Edwards]

Both Garry and I mentioned the word collimated in posts above. This is exactly what collimated means (apologies if you know this, but it seems you didn't pick up on it - see: http://en.wikipedia.org/wiki/Collimated_light). The more modifications you put on it - i.e. adding grids etc to remove the spread of light at source, the less spreading out the radiation does as it travels, so the more consistent the light becomes. You may well have to put a lot of power into the flash to get enough light to light what you want, but the beam is partially collimated (or projected as you put it)

For example, if you put a 10 degree honeycomb over a light, the light cone will be a 10 degree light cone. You can think of that as a point source somewhere behind the flash. For a 7" flash with a 10 degree reflector, that equates to a point source almost a metre behind the honeycomb (from basic trig). Clearly, as you get tighter and tighter honeycombs, that effective point source moves further and further behind the flash head front. It is the distance from the effective point source that will dictate the doubling of area.

If someone wants to try this with a 7" reflector plus 10 deg honeycomb, the maths says the two stop points (i.e. quarter power) should be at 0.5m, 2m, 6m...

Some good points and a thorough test. I'll reply in detail later, must do some work now
Also, I can do a test with a standard (7") reflector with and without a 10 deg honeycomb, come to that I can do the test with a 5 deg honeycomb too, and can also do it with a softbox, with and without a honeycomb - will post the results when I can.

 

[irelanst]

I did some maths on this ,

For a square softbox (900x900) the percentage of light emitted falling onto a 250mm diameter sphere (someones head) vs. the separation distance from the centre of the head to the soft box front diffuser is;

Distance Light falling
(mm) on target % Reduction
250 6.71083% N/A
500 3.58012% 53%
1000 1.30560% 36%
2000 0.37195% 28%
4000 0.09644% 26%
8000 0.02434% 25%
16000 0.00610% 25%

The third column of numbers is the percentage reduction in incident light from the previous step. As can be seen initially the softbox does not follow the ISL, but from 2m and more it is very close.

This doesn’t take into account any reflection effects, that the face of the softbox is a perfect diffuse surface, and also assumes a uniform light intensity over the face of the softbox (admittedly a big assumption).

It also goes to show how very few of those precious watt seconds actually go into illuminating the subject!!!

I think that it shows that a (relatively) large softbox does not obey the ISL exactly when positioned close to the subject, but the light fall off is certainly rapid as you move it back.