Some data on Oly "Focus Stacking"

I was curious to get some feel for how the "differential" step between 1 and 10 worked out in practice using the Oly "Focus Stacking" feature on my E-M1. With the Oly 60 mm macro lens mounted, I set up a Lens Align target at 500 mm from the sensor plane, and snapped away with the 8-frame standard focus stacking bracket using apertures of f/2.8, f/4.0, and f/5.6 and differentials of 1, 2, 3, and 4. The setup is shown below.

Setup so as to read off DOF in the image

The two graphs show the resulting data. It was very interesting to see that the actual size of each of the actual focus steps depends on the aperture. For the 60 mm macro lens set-up, only differentials 1, 2, and 3 provided a focus stack of overlapping frame focus ranges. The step size for differential=4 was larger than the depth of focus, and the in-camera stack merge failed (no disrespect to Oly, it would also have been similarly ugly using any other focus stacking software). The actual step size for differential=3 was just within the depth of focus, and so gives a maximum bracket width from the 8 frames.

Actual focus step size (subject at 500 mm) per differential, size of DOF shown for comparison

Width of the focus bracket after merging the frame stack in-camera, single-shot DOF shown for comparison

My take-away is that, if you use the Oly "focus stacking" feature, you should be quite safe with differentials of 1, 2, or 3. You'd look up your DOF (using an app, eg HyperFocal Pro) for your particular lens and subject distance (remember to measure/estimate from the sensor plane, not the front of the lens!), and then you would set your aperture to accommodate the bracket width you wanted, knowing that this is about 3 times DOF for diff=1, 4.5 times DOF for diff=2, and about 7 times DOF for diff=3.

Thank you very much, Lester. I have been using the Focus Stacking and Focus Bracketing (using Helicon Focus for the final image) and had more or less found results similar to yours, but without the quantitative aspect, which make this work very useful to me.

I will print your graphs and have them handy, in future!

But we really, really should have one of these graphs for each subject distance (which is totally impracticable, or a 3D plot (equally impracticable), or even a calculator, perhaps an App (very clumsy).

I wish that Olympus, in the next version of this excellent process, would add just a bit more practicality:

First, we should set the closest focus point.

Then the more distant point to be still in focus.

Then indicate the desired number of steps.

A further refinement (but I would be satisfied with only those I mentioned above) would be to have the camera calculate the number of steps from the DoF corresponding to the f stop.

Then it would be extremely useful, no more guesswork, or very little, bull's eye each time!

Please listen, Olympus!

It's people like you that make this forum so helpful, practical and interesting! Thanks!

And I haven't even (yet) a camera on which it can be implemented!!

Mike

Pedagydusz wrote:

Thank you very much, Lester. I have been using the Focus Stacking and Focus Bracketing (using Helicon Focus for the final image) and had more or less found results similar to yours, but without the quantitative aspect, which make this work very useful to me.

I will print your graphs and have them handy, in future!

But we really, really should have one of these graphs for each subject distance (which is totally impracticable, or a 3D plot (equally impracticable), or even a calculator, perhaps an App (very clumsy).

I wish that Olympus, in the next version of this excellent process, would add just a bit more practicality:

First, we should set the closest focus point.

Then the more distant point to be still in focus.

Then indicate the desired number of steps.

A further refinement (but I would be satisfied with only those I mentioned above) would be to have the camera calculate the number of steps from the DoF corresponding to the f stop.

Then it would be extremely useful, no more guesswork, or very little, bull's eye each time!

Please listen, Olympus!

I did express that for E-M1 v2. First ponit, last point, and make steps on your own.

Is Helicon much better than Adobe PS in stacking, i used Zerene sometime, but more Adobe.

Petar Veliki wrote:

[...] Is Helicon much better than Adobe PS in stacking, i used Zerene sometime, but more Adobe.

[...]

If you deal with many frames, Helicon Focus or Zerene are way, way faster than PS. Plus they can have different stacking algorithms that are better suited to one or other conditions and subjects.

Another plus for HF is that it can process DNG files, so you are essentially stacking RAW images, but somehow that doesn't seem to work in my computer (a few years old).

Thomas Karlmann wrote:

[...]

Interestingly, old Canon cameras do something quite similarly. With Canon's old system, you clicked on the nearest thing, then clicked on the furthest thing thing you wanted in focus, and the camera would set the focus distance and the aperture to meet the requirements you set. I thought it was fantastic -- way back then.

[...]

I use a set of motorised focussing rails - Stack Shot - that follow that procedure as well (except calculating the aperture.

And there is an App for Panasonic MFT cameras (at least for GX-8) that is said to work like that (for Android, never got mine to work in IOS).

So, I think that it is widely recognised that that is the method to use. I hope Olympus will, as well!

Pedagydusz wrote:

Petar Veliki wrote:

[...] Is Helicon much better than Adobe PS in stacking, i used Zerene sometime, but more Adobe.

[...]

If you deal with many frames, Helicon Focus or Zerene are way, way faster than PS. Plus they can have different stacking algorithms that are better suited to one or other conditions and subjects.

Another plus for HF is that it can process DNG files, so you are essentially stacking RAW images, but somehow that doesn't seem to work in my computer (a few years old).

Is Helicon downloadable ?

I used "DNG" in Adobe stacking, exporting RAW into TIF.

Here's my old thread ont eh subject.

The rule of thumb I use now is:

Aperture ~= differential.

So, f5.6 = differential 5

f2.8 = differential 2

f11 = differential 10.

That seems to work fairly well for macro level stacking.

Petar Veliki wrote:

[...]

Is Helicon downloadable ?

[...]

You can download Helicon Focus. There are more than one product, and they have a paid license to operate, but you can use a free trial version as with most paid applications.

I took another series of snaps using the 12-40 Pro at 30 mm, with the same target distance of 500 mm.  Far focus depth ran off the angled measuring plate at higher apertures and higher differentials, so the graphs are not as complete as earlier, but I think the trends are reasonably clear and pretty similar to what we saw with the Oly 60 macro.

Step size (average for the 8 steps of a focus stack) of low differentials

We can see that differentials of 1 and 2 give overlapping focus frames, a differential of 3 gives steps pretty much the size of the DoF, and I would expect differentials of 4 and larger would leave quite large oof gaps in the stack.

Bracket width of merged focus stack

Roughly, the total DoF in a merged stack is 2x the single-shot DoF for diff=1, 4x for diff=2, and 8x for diff=3.  This relationship seems to apply regardless of aperture -- in other words, the actual step size depends on aperture and maintains its relationship to the single-shot DoF for that aperture.

I don't know how the Oly algorithm works for different subject distances.  When I have a little more time, I'll check out the behaviour at distances of 250 mm and 125 mm.  I would expect that the parameter here would be the magnification, such that step sizes would be proportionately reduced at increasing magnification.

First of all let me add to the appreciation for this very useful insight into the Oly in-camera stacking algorithm.

A previous post also suggested that step 3 was the max to avoid gaps, but was much less elegantly presented and with more limited conclusions. http://www.dpreview.com/forums/post/56869932

I feel much more comfortable using the technique in close focus settings and would be interested to see comparable data for the much longer distances for use in landscape shooting.

I do have a question about the distance parameter you suggest using for DOF calculation: I know that minimum focus distance - as opposed to working distance - is calculated from the focal plane, but I had always thought that it was indeed lens to subject distance that was used to calculate DOF. Can you clarify?

Thanks

woodybrown wrote:

I had always thought that it was indeed lens to subject distance that was used to calculate DOF. Can you clarify?

I use the PhotoPills app to calculate DoF, and give it the subject to sensor distance.  It seems to return a reasonable value...  For example, it suggests the DoF for this shot is 2.1 mm (Oly 60, f/2.8, 33 cm subject distance), and looking at the shot I'd tend to agree!

100% crop ooc (ISO 1600), mm on right, inches on left

fishguy1 wrote:

The author of this article calls the space between shots that are not sharp, "blur banding".

Blur banding occurs when there is insufficient focus overlap between the stack frames, I show an example in this post, http://www.dpreview.com/forums/post/57716893.

Thanks for the feedback!