Brian Caldwell

33 minutes ago, tupp said:

No.  I already addressed this passage previously in this thread.

Let me get this straight - are you denying that the combination of an 80mm lens with a 0.7x focal reducer is a 56mm lens?

4 hours ago, kidzrevil said:

Interesting stuff about the DOF ?, I really like the character of @Mattias Burling shots. Does the DOF equivalence really matter if the rendering looks this good ? Serious question, not trying to troll

Character/rendering are entirely separate issues from photographic basics like DOF and perspective.  Certainly, if you like the character then there's nothing wrong with that.  Heck, if you find that using an 80mm lens with both a 1.4x teleconverter and 0.7x focal reducer (in series!)  to get back to 80mm but with some funky aberrations gives you the results you want then you should certainly do it.  My only real objection is with pseudo-explanations - basically "fake physics" - that cross the line into mysticism.

Mini correction:  techically, aberrations can and do influence DOF.

2 hours ago, tupp said:

It most definitely does not show equivalence -- the foreground (front of lens) stays sharp in both images while the focus on the background changes between images.

 

 

The circled items in the above image do not seem dark or shadow-like:  they appear to be a well lit counter-top edge and a detail on the front of a cabinet door.  Nevertheless, the focus one those to items is most certainly changing.  It is obvious that they go soft and then back to sharper focus.

 

In addition, the bokeh changes significantly -- a telltale sign that the DOF/background-focus has changed:

In the read circle, note that the bokeh changes size dramatically, while the dark rectangular object that it overlaps changes less so.  It is very obvious how the proportions of that bokeh changes relative to the dark corner it overolaps.

 

The bokeh in the yellow circle changes size and its edge changes softness.

 

The two overlapping bokehs in the orange circle are more blurred and blend together in one image, but are sharper and more distinct in the other image.

 

Bokeh is not shadows.

 

In the blue circle, there is some purple item that is softer in one image, yet becomes more defined in the other image.  Is that a purple shadow?

 

In the green circle the leftward edge of the tripod head bracket is sharper in one image but blurrier in the other image.  It is lit brightly enough to see that the focus there has changed, while the focus on the front of the lens stays sharp.

 

The DOF is different -- not equivalent -- in these two images.

You are expecting a level of precision in this comparison that is entirely unreasonable.  Little things like changes in distortion and entrance pupil position during zooming make it impractical to make a blink comparator test completely perfect.  What the comparison does show - with more than sufficient precision - is that you can optically reproduce all aspects of an image shot on a large format with one shot on a smaller format - or vice versa.  

The notion that, say, an 80mm medium format lens has some inherent "80mm-ness" or "medium formatishness" that somehow stays with that lens after you attach a focal reducer is just silliness.  The combination of a 0.7x focal reducer and an 80mm lens is a 56mm lens.  Period.  Put that 56mm lens on a 24x36mm format camera and it will behave just like any other 56mm lens attached to that camera, the only caveats being related to aberrations and other flaws in the lens and focal reducer.

22 hours ago, andrew mcmillan said:

I saw bokeh split in half as well as other tell tale signs, it looked to me as if it was shot with fishing line behind or infront of the lens.  but also oval aperture disks? 

The bokeh that I saw in this movie was definitely Zeiss Master Anamorphic, which is to say it has an aspect ratio of ~1.64:1 rather than 2:1.  The Zeiss anamorphics are famous for requiring add-on junk to create flare - available from Zeiss themselves, but I would be surprised if anyone would knowingly copy the 1.64x bokeh with an aperture disk.

On 2/23/2017 at 5:33 PM, jcs said:

I did those tests- it's not really possible to do perfect equivalence with physical lenses unless all the settings can be exactly matched. In the first example, the only major difference was shadow detail which could be related to ISO... In the second example, maybe I made a mistake or it's still related to optics not really being equivalent. The 'normal' test matches almost perfectly.

Brain Caldwell, the optical engineer and inventor of the Speed Booster says the same thing regarding FF vs. MF. That's why he wasn't interested in making a MF to FF SpeedBooster...

In any case, the differences are minor and most people couldn't tell the difference. Someone posted computer graphics (ray traced?) examples that matched perfectly, as the math predicted.

You should be aware that this is a religious discussion.  Comparisons, and discussions about comparisons, should be avoided and only discussed in private!

On 2/12/2017 at 3:46 PM, elgabogomez said:

Single focus AND useable on 40mm full frame?? How is this not being talked more? As usual Tito Ferradans, superb review.

40mm on a 1.33x gives the same horizontal FOV as 60mm on a 2x.

On 2/8/2017 at 5:47 AM, Ki Rin said:

I see that there was an OM to m43 speedbooster from metabones a couple of years ago that seems to have been discontinued. What happened?

Seems like it would be a nice option. Is it gone forever?

All of the original m43 Speed Boosters were replaced by the Ultra model a couple of years ago.  The OM version of the Ultra is a current product:  http://www.metabones.com/products/details/MB_SPOM-m43-BM3

On 1/28/2017 at 6:03 AM, webrunner5 said:

Well a 57 Chevy was probably only 1200 bucks new either but now...

And a 1925 Model T was only $260 new . . . .

17 hours ago, Nikkor said:

@Andrew Reid 

Have you tried using the bmcc (Not Pocket) speedbooster on the gx80?

I haven't tried that particular combination, but the 0.64x BMCC Speed Booster has a full millimeter *less* clearance on the camera side than the 0.64x XL.  The XL also has a larger image circle that is compatible with the fill size m43 format, and it is also optimized for the standard m43 filter stack rather than the non-standard Blackmagic stack.  Odds are you can probably get the BMCC version to work, but you'll probably be pressing on the flexible shutter cover and you won't get optimal image quality.

22 hours ago, Axel said:

 

Now the whole lenses questions come up again. Which speedboosters work? Will some vignette on the non-crop sensor (or little crop compared to GH4)?

 

The 0.64x Speed Booster XL and 0.71x Ultra are both designed to cover fullframe 4:3 format (21.6mm diagonal).  As long as you use a master lens with a sufficiently large image circle you wont have vignetting.  24x36mm format lenses will always work.  I expect some issues with certain APS-C lenses.

On 1/4/2017 at 11:46 AM, Andrew Reid said:

"6K/24p Anamorphic Video Mode, while fun, is severely hampered by its 4:3 aspect ratio"

UM!

That's what an anamorphic mode is - 4:3

Someone let our dear friends at Cinema5D know.

Well, you would be slightly hampered by 4:3 if shooting with 2x anamorphics.  But with 1.79x anamorphics on the other hand . . . . .

9 hours ago, Laquaglia Punto Ernesto said:

double focusing is not as hard as with other anamorphic adapters.  and the optical quality is great. 

Can you elaborate on how the double-focusing on the Mesmerizer is not as hard as with other adapters?

21 hours ago, Hans Punk said:

@Brian Caldwell That makes perfect sense.

I've had very good results by 'choking' many of my taking lenses by inserting a fixed spherical aperture disc (relative to f4-5.6 to that lens' aperture size). This enables a non-cut oval bokeh, whilst effectively stopping the lens down. f/4 on 2x anamorphic is already a darn shallow depth of field to pull focus with for anything other than trees and sleeping cats.

The aberrations I've noticed on the FM and Core DNA at wide apertures seem to be more apparent to the edges of the image - highlight 'smear' seems to be more pronounced on the off centre area of the optic - where the curvature of the front element begins. Adding a +0.4 doublet Tokina often seems to help clean up these edge artefacts at wide apertures....albeit at the loss of enabling infinity focus. There also seems to be another artefact that I'm not entirely sure about - that there seems to be optic face to optic face reflection that can create highlight 'blobs' when the elements are separated by distance during focus. The FM (and apparently rangefinder) seem to have this 'blob' artefact which makes me think it might be a coating issue (internal face reflection?)....Pure guess mind you.

Hi Hans:

Using a good diopter will definitely improve the closeup image performance of a single-focus attachment, since it allows the attachment itself to be used closer to its "zero aberration" infinity setting.  I'm not entirely sure what you are describing about the highlight blob ghosting effect.  Most single focus attachments use a plano-concave moving front element followed by a convex-plano stationary element.  The concave surface of the front element is normally very close in radius to the convex surface of the stationary element, and this could set up a pair of reflections that might cause the effect you are seeing.  Its also possible that a reflection from the sensor followed by a reflection from either or both of the plano surfaces can cause noticeable ghosting, although in this case I would expect sharply focused mirror-image type ghosts.  Regardless of the root cause of the ghosting, its appearance can be reduced, but not completely eliminated, by using better AR coatings.

4 hours ago, Hans Punk said:

I'm talking about the limitations of a non-cemented doublet (achromatic) diopter (vs the single element as with these vari-diopter designs) - the single element optics of these vari-diopters is a major cause of aberrations of image at wide apertures.

Having owned a FM and Core DNA, I can confirm that f4 is the stop where both focus solutions start to resolve the cleanest image on high contrast edges...especially when pushed on full frame. Sharpness 'performance' depends on your eyesight I guess.

BTW - Tito's FM that he did the test with used to be mine...so I know that lens particularly well

Awhile back I spent many weeks designing variable diopter systems, including simple 2-element designs using singlets, 4-element designs with achromatic doublets, and a few complex 5 and 6 element designs.  What I discovered was that the limiting aberration in every case was spherical aberration at close focus.  Surprisingly, using more complex designs has very little impact on that spherical aberration, and they have numerous drawbacks including excess cost, and larger size and weight.  It turns out that a simple 2-element variable diopter, such as the ones used in all of the Iscoramas is not such a bad solution at all.  They give fantastic results for distant objects, and only gradually reveal weakness as you focus close.  As an aside, the only thing that I found to reliably eliminate close focus spherical was to allow the elements to get very weak, but this is completely impractical because the elements get huge and the front element motion becomes BIG.

BTW, it doesn't make much sense to talk about f/#'s when evaluating single focus units because absolute pupil size is what really matters.  For example, you might get a terrible result when attaching the unit to a 200mm prime at f/4 because the pupil diameter is 50mm, and yet get a really nice result with a 24mm f/1.4 because the pupil diameter is only 17mm.

1 hour ago, Devon said:

So I had a thought today. WARNING: I tend to ramble, so bear with me  

--I have been thinking about buying an anamorphic lens lately. I shoot with a Sony A7s, and therefore am resticted to 16:9. I don't have a large budget, and have settled on buying a 2x anamorphic lens off of Ebay.

One of my concerns is the ultra-wide aspect ratio 2x produces on a 16:9 sensor (3.55:1.)

If I stretch the anamorphic footage in post, rather than squeeze, I double my horizontal resolution.

But that horribly-wide aspect ratio (3.55:1) it produces is just too wide for my taste.

So, after doing a bit (a lot) of math, I realized that if I just horizontally crop the image in post to fit a 2.66:1 or 2.39:1 composition (correcting the distortion by horizontally stretching the footage), that there is essentially NO need to buy a 1.33x or 1.5x lens.---Depending on the horizontal crop, a 2x lens essentially has all 3 types of anamorphic lenses (1.33x, 1.5x, and 2x) "built in" (depending on the horizontal crop I choose.)

Depending on the horizontal crop I choose from a 2x lens, it also produces the same horizontal FOV the other anamorphic options (1.5x, and 1.33x) would produce. 

So after all this rambling, my question is as follows...

For those of you familiar with anamorphic shooting, is this theory essentially correct?

Thank you for sticking with me  I know this is a lot of info to take in. 

The main reason for shooting 1.33x (actually, 1.344x would be best) is to get 2.39:1 output with no waste of a 16:9 sensor.  Shooting 2x on 16:9 and cropping throws away resolution.  Aside from these considerations, there is the aesthetic look that different squeeze ratios provide.  Many people dislike 1.33x because it doesn't look very anamorphic.  However, even 2x anamorphics don't always share the same look.  Rear anamorphics don't look anamorphic at all.  Zeiss 2x Master Anamorphics actually look more like 1.64x true front anamorphics due to their mixed front/rear design.

7 hours ago, jase said:

Thanks Brian! But other than the crushing aspect, what do you think? And do you know by chance, how much space we have between the Sigma 18-35 and the metabones xl ef?

Its a little complicated due to the fact that the outer surface of the Speed Booster is concave.  But you should have at least 1mm of clearance with this particular combination.  Note for anyone else reading this post that the Speed Booster XL for m43 has more clearance than a Speed Booster Ultra for APS-C.

11 hours ago, jase said:

........But how about using a speedbooster and put the thin filter glas on top of the filter and yet behind the lens? Is this a stupid idea because the flaring would still be the same regardless the position of the filter? Maybe @Brian Caldwell could chime in...

It depends a lot on the lens.  For some lenses, such as the 50/1.2 Nikkor, there is almost no space between the rear element of the lens and the front element of most  Speed Boosters.  I would hate for you to crush your filter between two lens elements!

58 minutes ago, tweak said:

Open to proof showing me otherwise. But no 0.9 lens I've ever seen has amazing IQ wide open.

A Zeiss Otus plus Speed Booster XL has pretty amazing IQ wide-open at f/0.90 on m43.

9 hours ago, IronFilm said:

 

http://www.theasc.com/site/wp-content/uploads/2016/06/AC-A-Beautiful-Planet.pdf

Lenses and cameras are like Doritos:  "Crunch all you want, we'll make more"  

22 hours ago, Andy Zou said:

Huh!  Fascinating.  So I'd need to get another camera.  I assumed that the optics between the front element and the iris would've thrown off this measurement.

I think I read somewhere here or otherwise that the distance between the rear element of the anamorphic adapter and the entrance pupil is inversely correlated with the quality of the pairing.

The optics between the front surface and the iris are what determines the location of the entrance pupil.  "Quality of the pairing" is only indirectly related to entrance pupil distance, but the amount of vignetting is directly related to it.  To minimize vignetting you want the entrance pupil as far forward as possible, and the air gap between an anamorphic adapter and the prime lens to be as small as possible.  

11 hours ago, Andy Zou said:

I recall somewhere reading that the distance from the adapter to the entrance pupil affects the quality of the combo.  How does one measure the entrance pupil?  Is this only visible in a lens' cross-section breakdown?

Best done on an optical bench with a low-power microscope.  But a macro lens will work OK as long as you focus by moving the whole camera.  1)  stop lens under test (LUT) all the way down; 2)  Focus on the iris diaphragm; 3) move microscope or camera away from the LUT until the front lens surface is in focus - you may have to apply some dust or some Scotch tape in order to see the front surface.  The distance that the microscope or camera moves is the entrance pupil distance, e.g., the distance from the front surface of the LUT to the image of the iris diaphragm.

I'm thinking about manufacturing some 114mm clamp-on (fast on-and-off) achromatic diopters for the wide-angle anamorphic attachment I'm developing.  Similar to the now-extinct Zeiss Master Diopters.  Also similar to the Leica Macrolux, except larger and less expensive.  Aside from being useful accessories for my anamorphic attachment, would these be of more general interest to readers here?  Thinking about +0.4D, +1.0D and +2.0D .

2 hours ago, Bioskop.Inc said:

So is this the reason that people think that the 54 is less sharp than the 36, especially when shooting with a lens wide open, because in fact you won't be shooting at the aperture that you think you're shooting at? So in fact, when shooting with a 54 you really are shooting at the right aperture & so it will reveal the taking lens to either be sharp wide open or soft - nothing to do with the attachement, but more to do with the taking lens?

Its definitely a factor, and it certainly helps to be aware of what is actually the limiting aperture in the system.  Even the Iscorama-54 isn't large enough to avoid stopping down an 85mm f/1.2.

Although I don't have the Iscorama-54 design prescription and can't accurately evaluate its performpance, I do expect that it has plenty of aberrations on its own, especially at full aperture.  A good test of the 36 vs 54 would be to use a prime lens with great performance that has an entrance pupil larger than 54mm.  The 85mm Zeiss Otus comes to mind . . .

3 hours ago, icarrere said:

Speaking about the Iscorama 36, I´ve put the taking lense at f:2 and IMO it is sharp. All depends on what taking lense it is.  (I'm use to the Nikon 85 F:2)

Bear in mind that in your case with an 85mm lens the rear of the adapter limits the aperture, so you are shooting at f/2.4 and not f/2.  You would need an Iscorama 42 to shoot at f/2 with an 85mm lens.

Has anyone heard anything about whether the automatic anamorphic crop mode of the GH4 has been retained/improved?  Or will it require manual post-processing?

At this point, NX accessories are probably best done by NX enthusiasts like Luca.  Otherwise something called "opportunity cost" rears its ugly head.  What it means is that labor and financial resources are limited - especially in small companies - so you have to be very careful about choosing which products to develop.