Lens Projection and Real Life images

[Michael Bravin]

What we found over and over when first brought the DigiPrimes out was that lens projection, which is essentially necessary for evaluating lenses for film cameras is only a part of the picture (excuse the pun). Matt Duclos mentioned that the reverse light path on a projector is not an issue. Respectfully I disagree only on a really fine point. The coating formulas are defined by the direction of light when using the lens on a camera, so projection tells us less than we should know about issues of flare and veiling glare, Also colorimetry.
The geometric issues we see on a projector are opposite of what we see on a camera, i.e. pincushion http://www.dpreview.com/learn/?/key=pincushion+distortion
on a projector on wide angle lenses is seen as
barrel distortion http://www.dpreview.com/learn/?/Glossary/Optical/Barrel_Distortion_01.htm
and vice versa.
Colorimetry issues,purity, accuracy etc cannot be judged on a projector so on-camera tests are critical. While I have never seen a lens that looks great on a projector not make good images, I have seen many lenses that look average on a projector look great on a camera. Keep in mind that a camera's sensor and lens combo is the optical system not just the lens.
All that being said we depend on lens projection as a strong comparative reference.
My vote is that the SALT test will tell us the relative performance of all the lenses on existing sensor or target size. Now we have a common reference. The issues of how each lens performs on an imager size and technology that isnt here yet is separate and if the RPPs continue their performance out to 5K then that will be additional info for those who need that, but for a while at least a Super 35 or Academy size image will be the majority of work and a good starting point.

Of course Jim's point is that if you're buying you want to have an upgrade path for a 5K sensor, you will have to buy new designs like RPP or still lens designs. Since they werent designed or built to cover 5K the Master Primes and Ultra Primes and Cooke S4 wont work at 5K.

But.. These are the lenses that professionals use for motion imaging everywhere in the world NOW and are known for their look and feelings they communicate and should be the reference not the image circle IMHO.

Once they all are compared for existing formats then the SALT 5K test could be done or SALT 43mm or something like that. I do think the industry needs common benchmark and that would be performance at 4K on a RED ONE, F35 D21 and 35mm Film , for today at least and thats what I would like to see for SALT.

Zeiss and Cooke and Angenieux can decide if and how they will address large target sizes or just concede that market to RED (doubtful )

No doubt what I saw at the RED event at NAB with RPP was QUITE impressive on the Chrosziel projector. Now they must pass muster on cameras, have consistent build quality after mass production, have continued performance after working on projects and being beat up, and hold up their performance over time with continued use. Lens 1 and lens 1000 should have equal or very close performance. This is how all other lenses are evaluated, projection just a strong starting point.

So far so good for the RPPs thats for sure.

Now lets add the SALT for more flavor

 

[Matthew Duclos]

I agree. Testing on a camera should be considered for specific examinations such as color and general use (mounting, fit, interface, etc..)
My ONLY concern with projection and reversed light paths is mechanically reflected light from surfaces inside the lens, a minor concern at that.
I don't think the RPP or other popular lenses' coatings are bearing sensitive.

 

[Mitch Gross]

Ditto what they said. ;-)

 

[Michael Bravin]

I agree. Testing on a camera should be considered for specific examinations such as color and general use (mounting, fit, interface, etc..)
My ONLY concern with projection and reversed light paths is mechanically reflected light from surfaces inside the lens, a minor concern at that.
I don't think the RPP or other popular lenses' coatings are bearing sensitive.

Matt

I have several times seen noticeable contrast change when a lens is on camera vs projected images of the same lens lacking contrast. Other than coatings and baffles what could it be?

BTW it was a pleasure to meet you face to face at RED NAB. I've known your dad for awhile. You, only by stellar reputation.

Guess I should ask "Ditto" the same question

Mitch-
Any ideas about significance of light direction on overall image integrity or quality from your experience? Anyone else?

 

[michael zaletel]

Sorry for butting in but can anyone list two famous features shot recently with different lenses (different quality and cost levels) where I can see how all of these factors matter? I can certainly understand why lighting and exposure matter but I haven't seen a single comparative example of feature footage shot by top talent under ideal conditions that demostrates the value of Master Primes over lesser glass. Perhaps top talent with big budgets and ideal conditions always use Master Primes or equivalent. That sort of makes the comparison a self-fulfilling prophecy doesn't it?

I mean when's the last time anyone said..."wow, that was an incredible movie, well written, well directed, incredibly lit and perfectly photographed. If only the DP had used Master Primes, then it would be a masterpiece."

Please advise,

-michael zaletel
(shooter)

 

[Michael Bravin]

Sorry for butting in but can anyone list two famous features shot recently with different lenses (different quality and cost levels) where I can see how all of these factors matter? I can certainly understand why lighting and exposure matter but I haven't seen a single comparative example of feature footage shot by top talent under ideal conditions that demostrates the value of Master Primes over lesser glass.

Please advise,

-michael zaletel
(shooter)

Are you saying you cannot "see" any difference in the quality of the lenses (not just resolution and technical specs but look and feel) in different films? Have you seen Gangs of NY? Portrait of a Geisha? ANY Speilberg film? Brand and look of the lenses used shows up as an integral part of the cinematographic visual style of a film.

Gangs of New York was shot by Michael Balhaus with Cooke S4 lenses. It has a "look" that is part of the story.

Master Primes allow you to make a crisp clean contrasty image wide open at T1.3 NO OTHER lens can do this. Its part of their clean look.

Lenses are tools chosen by artists just like a brush with different hair or paint that is oil or acrylic gives a different feel and look.

Or did I miss your question all together?

 

[Michael Bravin]

"wow, that was an incredible movie, well written, well directed, incredibly lit and perfectly photographed. If only the DP had used Master Primes, then it would be a masterpiece."

That would be marketing. Red already won that battle

 

[michael zaletel]

Are you saying you cannot "see" any difference in the quality of the lenses (not just resolution and technical specs but look and feel) in different films? Have you seen Gangs of NY? Portrait of a Geisha? ANY Speilberg film? Brand and look of the lenses used shows up as an integral part of the cinematographic visual style of a film. Or did I miss your question all together?

No, I am not saying one cannot "see" the difference. I am suggesting this is all extremely subjective rather than objective. What I am asking for is an example of a film that you think "should have" been shot with more expensive or higher quality glass (as judged by a lens projector test of some sort) because if it had, it would have been better.

People say things like that all the time about the writing, direction, acting, sound, scoring, fx, exposure, focus, lighting, locations and etc. I've just never heard it said about an inferior lens choice is all. That's why I am asking if it has been said and for an example. I am not trying to be smart, I am trying to learn something and fully expect for someone to point out an example or two I can go and rent for myself so that I'll be more careful, picky, etc. when choosing lenses. Sure there are reasons to use different lenses for different artistic looks. What I am asking is are there reasons to use more expensive, higher quality (in test environment) lenses from an objective quality standpoint.

-michael zaletel
(shooter)

 

[Mitch Gross]

Michael Zaletel -- One example off the top of my head would be the Jennifer Lopez thriller "The Cell." If my fuzzy memory serves me, the "reality" sections of the film were shot on Arri cameras using Cooke S4 lenses and the "dreamscape" sequences shot on Panavision cameras using Primos. The goal of the cinematographer was to help shift the imagery in a subtle way, in this case by altering factors like bokeh. Since the film went through a DI process for completion and effects and is now viewable only on some home video format, I doubt one could measurably detect the difference unless you really knew what to look for. I also know some cinematographers who would switch from Kodak to Fuji when moving from a wide shot to a close-up, and some who insisted on using two layers of 1/4 CTO gel instead of a single layer of 1/2 CTO because they could tell the difference.

It is not a question of how apparent the difference is -- it is a question of whether there is a difference or not. These are craftspeople who take great pride in the precision and artistry of their work. It is as important to them as the difference a painter would feel between using a 2" v. a 2.5" bristle brush. It is meaningful because it affects their craft.

This is a technical discussion about the merits of different precision optical instruments and the methodology by which to evaluate them. What's the point of asking whether it matters or not? By very nature of wishing to compare them it matters.

I mean, a Yugo and a Porshe are both cars with engines, four tires and a steering wheel. Any other difference is a mere quibble, right?

 

[Mitch Gross]

"Ditto" was written at 5am after a very long night getting over whatever bug I caught on the flight home. I actually thought it rather eloquent and to the point.

I do not know of any bidirectionality issue with optical coatings. I do know that there can be light scattering issues that can create differences between a lens on a projector and a lens on a camera. I.E., a camera lens supposes a lot of light in front of it and very little behind, so reflective elements are kept from the front but not as much of a problem in the rear, at least until you push light through them the opposite direction. But that's why most lenses now have a black masking cone surrounding the rear occular.

Interestingly, the designer of the IB/E Rebels noted to me that some of the lens designs came from both projection and capture designs, so he knew that they would perform well on a projector. I'll have to talk to him more about that when I see him again.

 

[Jannard]

It seems to me that there is a technical aspect to measuring lenses and a "feel" aspect as well. Interestingly, you can mostly understand why lenses "feel" the way they do by looking at them on a projector. You can see where they fall off in focus, the geometry, the contrast. Lenses are where science meets art and it makes sense.

If a lens image circle is X, it will never be Y... no matter how much you want it to be. If a lens has a focus falloff at 28mm, it will rapidly deteriorate out from there... no matter what you wish it to be. Same for illumination. Color is a more delicate thing to measure... but the easiest to compensate for.

I am all for shooting lenses on a camera and comparing results. But my strong bet is that the results will strongly correspond with what you see on a projector is most cases.

We are developing a new way to measure lenses through the whole system... to see exactly what happens at the end of the line. It will be like measuring horsepower at the wheels instead of the engine. It will tell you what happens after the IRLP filter and microlens of a given camera sensor in a measured way. Hopefully it will be available for the SALT test so there is yet another way to judge lenses in addition to a lens projector and shooting tests.

Jim

 

[jbeale]

You're probably familiar with it, but check out what Imatest has got if you haven't:
http://www.imatest.com/home

Their program enables a large (!) variety of lens, camera & system tests and produces more detailed results than most people would want, probably...

 

[Jannard]

You're probably familiar with it, but check out what Imatest has got if you haven't:
http://www.imatest.com/home

Their program enables a large (!) variety of lens, camera & system tests and produces more detailed results than most people would want, probably...

Graeme referenced this 2 years ago...

Jim

 

[michael zaletel]

I mean, a Yugo and a Porshe are both cars with engines, four tires and a steering wheel. Any other difference is a mere quibble, right?

I'm sure I'm in quite over my head Mitch but there is an important difference between what I am asking and your Yugo/Porsche analogy. I think comparing a G-Shock watch vs. a Rolex would be a more suitable analogy. I do like the artists' brush analogy but not all of the worlds' finest paintings were painted with the best (or even most appropriate) brushes.

Are you saying this is purely an artist/craft/precision/tool discussion and not a matter of perceptible visual difference? I don't think you are saying that. I am just asking for examples of the differences so I can learn to see them and decide for myself whether the extra cost is worth that difference..."to me".

-michael zaletel
(shooter)

 

[Sanjin Jukic]

Michael,

try this comparison:

Cooke S4 vs. Arri/Zeiss Master Primes on the same camera with the same cinematographer>>>

1. A Beautiful Mind (2001)>>Arriflex 535, Cooke S4 Lenses>>

2. In the Valley of Elah (2007)>>Arriflex 535B, Zeiss Master Prime Lenses>>

Cinematographer: [URL="http://www.imdb.com/name/nm0005683/"]Roger Deakins BSC, ASC[/URL]

DVD or BR_Disks are available...

 

[Mitch Gross]

The problem is that there are so many other tools at the artist's disposal plus they may be going for a different effect from one moment to another or one production to another. Lighting and art direction can be enough to disguise the differences between optics. Really the only way to tell is by testing, not by studying finished productions.

 

[Pawel Achtel]

Of course Jim's point is that if you're buying you want to have an upgrade path for a 5K sensor, you will have to buy new designs like RPP or still lens designs. Since they werent designed or built to cover 5K the Master Primes and Ultra Primes and Cooke S4 wont work at 5K.

Michael, you seem to be mixing up resolution and coverage (image circele). :nono: It depends on the sensor.

Which sensor are you talking about? The 5k S35 Mysterium-X or the 6k FF35 Monstro?

All MPs will cover full 5k frame on the Mysterium-X sensor, especially in 2.40:1 aspect ratio. What they will not cover is FF35 (the Monstro sensor), which is actually 6k. Some of the shorter focal length MPs will also not cover 5k area on the 6k Monstro sensor, but most will. :thumbup:

 

[DAVID McNAMARA]

It seems to me that there is a technical aspect to measuring lenses and a "feel" aspect as well. Interestingly, you can mostly understand why lenses "feel" the way they do by looking at them on a projector. You can see where they fall off in focus, the geometry, the contrast. Lenses are where science meets art and it makes sense.

If a lens image circle is X, it will never be Y... no matter how much you want it to be. If a lens has a focus falloff at 28mm, it will rapidly deteriorate out from there... no matter what you wish it to be. Same for illumination. Color is a more delicate thing to measure... but the easiest to compensate for.

I am all for shooting lenses on a camera and comparing results. But my strong bet is that the results will strongly correspond with what you see on a projector is most cases.

We are developing a new way to measure lenses through the whole system... to see exactly what happens at the end of the line. It will be like measuring horsepower at the wheels instead of the engine. It will tell you what happens after the IRLP filter and microlens of a given camera sensor in a measured way. Hopefully it will be available for the SALT test so there is yet another way to judge lenses in addition to a lens projector and shooting tests.

Jim

Jim,
Any chance of seeing some examples of footage shot with RPP's ? This is the first time I have purchased a set of primes without seeing them or shooting with them. I obviously have the utmost confidence in RED. My experience with my camera has been phenominal and I am hoping for more of the same with the RPPs.

 

[Matthew Duclos]

I mean, a Yugo and a Porshe are both cars with engines, four tires and a steering wheel. Any other difference is a mere quibble, right?

This is Awesome! I'm glad someone else is finally finding automotive theory to be relative to the motion picture industry.

 

[Matthew Duclos]

BTW it was a pleasure to meet you face to face at RED NAB. I've known your dad for awhile. You, only by stellar reputation.

It was very nice meeting you too. It's always nice to meet people in person.
I don't get out much... So a few times a year, trade shows are great for me.

As for the issue at hand. I will be totally honest, I can tell you A LOT about a lens from looking at it on a projector.
On the other hand, I couldn't agree more. A test wouldn't be complete without putting the lens on a camera and testing it.
Back to the automotive analogy:
Ferrari can test their cars all they want in a wind tunnel or in a lab...
But you need to put that nice new Scuderia on a track and let it run.
And to continue with Zaletel's inquiry, THe same must be done with a Toyota Carolla.