Lens Vignetting vs Strobe Coverage

I see many people getting obsessed with strobe field of view and coverage and I thought this post may be useful to relax some of them.

Frame coverage with two strobes with each 90 degrees is generally sufficient for wide angle. 

A fisheye like lens behind a dome will not cover 180 degrees due to the proximity of the focus point. Lens field of view is specified at infinity but when your target is close the field of view can drop significantly.

Even at specification level a diagonal fisheye is at best 92 degrees on the vertical axis in the centre of the frame and 142 on the horizontal.

A good check is to see shooting a pool wall how good is the cover of the vertical line in the middle of the frame

If this test is passed you know that the sides will be covered too. 

When it comes to the edges of the frame most mirrorless lenses have significant vignetting that is potentially corrected in post processing. 

This is a diagram from lenstip

In the corners the lens is 3 stops less than in the center. That means only 12.5% light intensity makes it there

Even after correction you do not get back the full 3 stops because otherwise the image will present severe artefacts and banding.

I generally disable vignetting correction underwater anyway as seldom there is something of interest in the edges.

It follows that no matter your strobe coverage even at short range your frame will not be lit homoegenously it does not matter how perfect is your strobe.

Obviously if your strobe is terrible this will degrade even further however the point is that the strobe is not necessarily to blame for dark corners if you have a mirrorless camera.

Vignetting diagrams for fisheye lenses are generally not available on full frame however the vignetting at the edges is of course 100% as it hits the lens border.

Some older DSLR lenses like the Nikon 28-70 f3.5-4.5 only have 0.5 Ev vignetting so with those and a water contact optic you generally would see brighter corners however the adapters have light fall off themselves and this is not measured as of today.

Strobe power instead is an issue at long range. When you move further away from the target to the point where the light source becomes a point regardless of shape of the tube only power matters. If you are interested in big wide angle scenes you are better off with a strobe that has a lot of power than one that is very smooth and wide but can't cover distance and will dissipate all the light without penetrating in the frame. Of course illumination will not be homogenous across the frame that is indeed impossible due to inverse square law but at least some of the fish in the middle of the frame will have color.

Here an example of a significantly large school of snapper shots at more than 2 meters with a pair of sea and sea YS-D2 at full power you can see the red dropping as you go deeper in the frame but still for a decent effect

Thanks, very helpful!

What about the issue of light temperature? Is it possible to correct this from raw or do i lose color information?

33 minutes ago, fruehaufsteher2 said:

Thanks, very helpful!

What about the issue of light temperature? Is it possible to correct this from raw or do i lose color information?

The light temperature issue is a different one. I believe @Alex_Mustard pioneered this topic and asked Inon at the time to do several filters.

In general terms flashlights normally match daylight color temperature this is true topside and to a good extent remains true underwater.

Now when you take a photo the light only covers part of the frame mostly the central area and at the edges due to inverse square law the effect drops. This is important from the point of view of the color of the water.

The warmer your strobe more color temperature difference there is with the water that is not covered by the strobe.

So with a warmer strobe you can get a deeper blue once the white balance is set correctly. 

Ultimately this difference is around 1000K and using selective masks you can change the color temperature of the water anyway separately from the subject.

In my opinion a strobe close to daylight 5200K is ideal and works as the camera auto white balance expects.

I know many people are adamant that 4600 or 4900 is better but the difference is minimal anyway.

A too cold strobe though say 6000K  or colder is however problematic especially in the shallow water where the water is not deep blue.

I am very happy with my sea and sea YS-D2J they are there at 5250K and I get the colors I want. This is not something I loose my sleep at night and I have tried all the 4600 4900 5200 options on the Inons too.

You also have to consider that as light absorbs warmer colour more than colder ones if you have a light source that is super warm this becomes colder as light travels anyway which is why you rarely see super warm strobes for wide angle.

So in short strobe color temperature affects the color of the water not your main subject so if you are not happy about the blues perphaps this is something you want to address. 300K do not change the world but 1000K are hard to manage.

How does the use of two strobes impact this analysis?  For wide angle people generally use two strobes pointing slightly outwards so the the beams meet in the middle at around the distance of the subject.  This is strobes 101 for dealing with backscatter.  There will be some overlap of beams at this point, but effectively the centre of your composition is covered by the inner edge of the beams.

In this case I can see value in having even coverage across the beams as the edge of the light cone is brighter than it otherwise would be.  What impact does this have on strobe usage?  Perhaps a strobe with more even coverage can be pointed outwards more so than one that that has rapid fall off towards the edges?  A sharp cutoff at the very edge would also help with controlling light.  In any case the strobe may be able to be turned down compared the case where the light drop was more severe.

I do agree that extreme angle coverage is not really required to cover even a fisheye.  At best its field is only 180° in the corners and as mentioned is about 140° on the horizontal axis at most.  In theory this could be covered easily with two 90° beams.  The other consideration with very wide beams is backscatter - light intruding into the triangular area between the camera/strobes and subject and illuminating particles that would not be lit with a narrower or better defined light cone. 

4 hours ago, ChrisRoss said:

How does the use of two strobes impact this analysis?  For wide angle people generally use two strobes pointing slightly outwards so the the beams meet in the middle at around the distance of the subject.  This is strobes 101 for dealing with backscatter.  There will be some overlap of beams at this point, but effectively the centre of your composition is covered by the inner edge of the beams.

In this case I can see value in having even coverage across the beams as the edge of the light cone is brighter than it otherwise would be.  What impact does this have on strobe usage?  Perhaps a strobe with more even coverage can be pointed outwards more so than one that that has rapid fall off towards the edges?  A sharp cutoff at the very edge would also help with controlling light.  In any case the strobe may be able to be turned down compared the case where the light drop was more severe.

I do agree that extreme angle coverage is not really required to cover even a fisheye.  At best its field is only 180° in the corners and as mentioned is about 140° on the horizontal axis at most.  In theory this could be covered easily with two 90° beams.  The other consideration with very wide beams is backscatter - light intruding into the triangular area between the camera/strobes and subject and illuminating particles that would not be lit with a narrower or better defined light cone. 

I do not point the strobe slighlty outwards. I push the strobe outwards using my arms once the subject is further away than the arm lenght the beams blend as shown in the image of the pool wall. In fact in many cases I point the strobes inward.

This is a test I did time ago at 1.5 meters with 8+12 inches segments strobes are point straight ahead

This is a false colour analysis of the same frame. You can do the same using photoshop with a bit more effort

Because light scatters in water you get this type of effect

Pointing strobe outwards is only a recipe when the strobe arms are too short and may result in dark areas in the center of the frame which are ugly.

I shoot predominantly wide angle and I am almost never in a situation where backscatter becomes a predominant issue although I shoot typically in places where the visibility is around 10-15 on a good day and there are plenty of particles.

I see sometimes people obsessed with backscatter trying to find a magic formula to avoid it entirely but there isn't one. If there are particles in the water you will see them what is important is that they are not predominant and don't ruin your subject.

If backscatter is a real issue you need to get closer thereby reducing the amount of water and suspended particles.

Pointing light away from the subject simply means not having the light and the color on the subject and hitting it with the lower quality of the beam which is always the edge and potentially creating big clouds of backscatter anyway on the outer of the frame because you have to pump up the power if you want to use the edges for the subject.

If we are talking about a fish portrait where there is a purpose of subject isolatio then you make your light more directional eliminating diffusers or using devices to focus the beam and with the light source at an angle you reduce the probability of distraction but there is no doubt you will point the light at the subject.

Lots of this backscatter scaremongery comes from ancient times where people would shoot something with a compact camera with on board flash which of course will create significant problems but this is why you have strobe arms (in addition to the fact that flat front lighting is ugly).