Epic Lens Test - Canon vs PL Elite

[Jason Sinclair]

As far as i know the autofocus in dslr's is of two types. The second type, contrast detection, probably is used in the canon mount setup. It has to go either side of the contrast to find the sharpest area which "finds" focus. As far as i am aware, on pro dslr's this is second choice to another technique involving sensor in the camera. Contrast detection is not professional enough to use on the move. Great for finding focus.

 

[Lauri Kettunen]

Not in the core of this thread, but guess you mean DSLRs which have a mirror employ the so called phase difference method for autofocus.

Josh, how abot testing also the autofocus properly? Nobody has done that, and likely, many is interested in it.

 

[paulherrin]

My experience seconds what Evin says about shooting wide open. Say take a Canon 24-70mm/f2.8 and the Red 17-85mm and shooting on low light the lens wide open. I think most dedicated people will notice the Canon image is less sharp towards the corners.

I'm pretty sure that's already a known and accepted thing... not saying it can't be demonstrated again, but how many examples do we need to prove this? I'm more interested in real world application, as I think we all would be better served. The higher quality lenses (and more expensive) certainly perform better in terms of full image sharpness on a test chart wide open, as they should. However, in most of my compositions, I hardly ever try to draw the eye to the edge of the frame. In fact, I often find myself intentionally blurring and vignetting my images in the grading process to keep the viewers attention inward. Frankly, the lens does a better job sometimes, it feels more organic to me, which I like. Sometimes it's nice to reserve that to post-production, of course. In that case, I know how to shoot my lens in such a way that it's fully sharp, usually by clicking the aperture a notch or two. If I can't spare that, maybe I need a lens better suited for that style or learn to live with the falloff.

I just don't always understand this pixel peeping. It's so superficial, and really just gives us something really obvious to talk about (aka manufacturing propoganda). Most of the currently available lenses are pretty darn sharp when used properly. Few lenses deliver edge to edge sharpness wide open. The ones that do are really great achievements, but you pay for it. Some professionals need and want that performance... others don't, or simply can't afford it. All I'm saying is, sharpness isn't everything. If you don't know how to get sharp results from your lens - getting a new, more expensive one isn't going to help much in improving the quality of your work.

Scientific testing, in general and in testing of lenses, needs to recognize subjectivity (we need a new scientific method). For one person, a pristine and 'perfect' lens makes the best images ever seen. To another, these are sterile and lacking in character. The same person may like different lenses for different projects, for their 'look' or ergonomics. Neither is right or wrong. Each serves an artist differently. I advocate testing and discussion that, rather than seeking a unanimous winner, gives way to an enlightened perspective and synthesis of information. Everyone's a winner, right?

Any discovery that does not lead to a whole new slew of ideas, questions, and discussion is not much of a discovery at all...

Parameters I'll suggest... (incomplete, as always)

1. Sharpness (yes, after all that
2. Autofocus/Focus Ergonomics
3. Lens Ergonomics
4. Deployment/Versatility (different situations/scenarios*)
5. Color/Contrast
6. Aberrations, Flare, Ghosts, Bokeh, etc.
7. Feel

*Real world scenarios. Studio Set, Tungsten/Daylight/Mixed/Etc, Outdoor, Night, Nature/Wildlife, Documentary/Events - this could be a good basis for the other parameters.

 

[Lauri Kettunen]

I'm more interested in real world application

Paul, I've never tried to test this with charts. My gut feeling experience comes from shooting landscapes.

 

[Zeb B]

Not in the core of this thread, but guess you mean DSLRs which have a mirror employ the so called phase difference method for autofocus.

Josh, how abot testing also the autofocus properly? Nobody has done that, and likely, many is interested in it.

Great idea! On vacation now in HI, will return and be in touch with you all when I have reliable Internet. Looking for mid to late January for the test. Will be fun and insightful. Love this type of stuff!!

 

[paulherrin]

Paul, I've never tried to test this with charts. My gut feeling experience comes from shooting landscapes.

that's a good place to test it. my gut thinks your gut is right. if everybody has undocumented and unrepeatable yet similar gut feelings and experiences, is that a basis for exploration and evidence? maybe it's a mass hallucination that the canons are a little softer around the edges. 'let's get an empirical test to prove ourselves right'

okay, it's late see you guys, don't let me cause too much trouble while i'm away...

 

[Zeb B]

Autofocus, as it exists in still cameras, is not useful in film production no matter how accurate or fast it is. Focusing is a creative act that is constantly changing dynamically as the shot changes and different actors or elements eneter or leave the frame. Even with face detection you need a focus puller (this is not the DP's job) to properly choose and ramp the focus form subject to subject. Autofocus may provide fine tuning focus tool that can handle the small adjustments, but for real production it's quite a long way from being ready.

My understanding is that the redmote wheel will be useable for focus but this is no more useful than grabbing the barrel, you need a true preston/berger style follow focus with mapping, a marking wheel and hard stops. This is what I expect the redmote pro to include.

The DP can create way better personal focus "art" on a wireless Wacom screen/tablet which talks in real-time to an AF lens . . . they just use the stylus to control the focus zone/point and adjust the pull velocity instead of a dude turning a knob to hit marks. Cost will be a driver here as well - potentially eliminate one more guy on set. That is until someone brings Lytro's "Light Field" technology to cine - then this discussion and big ass mechanical lenses with gears will be stored with the buggy whips.

Play with these examples and tell me that this style of focusing (either during acquisition on Wacom or after through light field technology) is not the future:
http://www.lytro.com/living-pictures/2328
Click on the part of the image you want in focus - it pulls to that

http://wacom.com/en/Products/Cintiq.aspx

 

[Jason Sinclair]

The kinds of tests Ken Rockwell does will suit the sharpness tests. Shoot something with trees and lots of details in the exact same tripod position then pick one area and zoom all pictures for a comparison. Do this on the center and the edges. Post pics. People can choose their own price/convenience per quality ratio based on real results rather than speculation. I would also add a contrast/falloff test to the above list.

 

[Jason Sinclair]

I think the future of autofocus in cinema will be to give the focus puller more control and quicker... more accurate pulls, but think everyone is looking in the wrong direction as far as technology is concerned.

 

[Evin Grant]

Josh, I can't even get a good, solid HD signal wirelesly on set and you think that a touch based focus system with all the inherent lag and processing delay is going to be workable? I'm all for the march of technology, but the system you describe will most likely require an additional technician and setup time to be workable, adding another position, not takng one away. Which is good BTW, we want more work, more positions and higher rates in this industry, not less. You have to remember that the majority of people who make a liing in film and TV are not filmakers but crew.

 

[Ryan De Franco]

You have to remember that the majority of people who make a liing in film and TV are not filmakers but crew.

+1. Lightfield lenses might be around for a few decades before filmmaking is entirely imagination capture (digital way, way beyond Pixar). Until then, I'll take the new O'Connor follow focus or Arri FF-5 for still lenses....

 

[Ryan Patrick O'Hara]

The DP can create way better personal focus "art" on a wireless Wacom screen/tablet which talks in real-time to an AF lens . . . they just use the stylus to control the focus zone/point and adjust the pull velocity instead of a dude turning a knob to hit marks. Cost will be a driver here as well - potentially eliminate one more guy on set. That is until someone brings Lytro's "Light Field" technology to cine - then this discussion and big ass mechanical lenses with gears will be stored with the buggy whips.

Play with these examples and tell me that this style of focusing (either during acquisition on Wacom or after through light field technology) is not the future:
http://www.lytro.com/living-pictures/2328
Click on the part of the image you want in focus - it pulls to that

http://wacom.com/en/Products/Cintiq.aspx

You do know that a 1st AC doesn't 'just' pull focus, right?

This won't eliminate any positions. If your idea were to be a possibility, the 1st AC would likely be pulling focus on the tablet. I don't see a big difference.

Your device is far and away from being ready. First off, you'll need a really quick and intuitive way to change factors such as feathering starts and stops, and speed... change them at a moments notice as the scene may change or be different from the time before. Also, I think the risk of error is higher. I accidentally touch the wrong app icon all the time on my smart phone. I can't imagine trying to touch a specific moving object in frame, if it isn't big and prominent.

Your idea is nice if you're doing simpler work, say an interview or a really obvious pre-set rack focus between objects. But in an active scene with actors, where improvisations, and changes occur, having your hand on an actual follow focus is a great way to make sure you can react and adapt as fast as the action changes.

I can, however, see how people who do not have access to a talented focus puller or cannot pull focus very well themselves would be tempted down this road. But once you have a good focus puller, you'll just decide it's easier to let him sit next to the damn camera and do his job.

It might be useful for macroshots, when DoF is very very shallow, and the shot is likely more set with less need for being able to adapt to changes.

I really don't think DP's are going to be pulling focus in the future. They have 1,000,000 other things to be thinking about and looking for. If a DP does do another job, it will more likely be operating.

 

[Jiri Vrozina]

FYI the new Nikon D800 and D4 which will be announced at the beginning of 2012 will both have 1080p uncomressed HDMI out. The D4 will have 102,000 ISO, that's a pretty compelling C cam for video.
The D800 will be 36MP too.

Until the redmote-pro starts shipping there is no elecronic follow focusing with any still lenses.
Once it does I'll be very curious to see how the Nikon and Canon lenses respond to it. Exciting times.

Dont worry about Canon and Nikon lenses and start using Zeiss ZEs.
No more PLASTIC here.

 

[Evin Grant]

Plastic has it's uses and the Zeiss ZE/F are not in the same leauge as the current G Nikkors, and yes, I've tested them side by side.

 

[Jiri Vrozina]

Well 21mm Zeiss and macro 100 are just perfect....should do more tests....
Nikon really should make new 50mm.

 

[conrad gaunt]

I own a set of zeiss zf's and a set of compact primes. When it comes to a moving subject that requires a focus pull, I'll choose the cp's every time. I can't imagine tring to pull focus on a canon lens,there is so much more to cinema glass than sharpness.

We spent £480 having focus gears attached to our zf set, focus pulling isn't an issue over here.. an iris pull would be tricky though.. except hdrx kinda makes them even less common

 

[D Fuller]

Auto Focus ? It's gonna be a little while before it's suitable for narrative filming... Might be a great way to pull off a shot now and then, when you'll let a computer cut your hair it's probably time to let it pull focus. Someday.

Matt Uhry

This might be the best quote of this thread!

As much as I love this technology, there is this really awful layer that can get between you, the artist, and the image you are making. You can spend 20 minutes getting the focus programmed just right, or you can spend 3 minutes setting marks and get the performance just right. It's your choice. have at it!

 

[D Fuller]

Well 21mm Zeiss and macro 100 are just perfect....should do more tests....
Nikon really should make new 50mm.

Yes. the 100mm macro is brilliant. The 50mm macro is pretty nice too. But Zeiss should be ashamed of the 50mm 1.4.

And yes, Nikon should make a new 50, now that they've shown what they can do with the 85, 35, and 24. And they should make a new 18 as well. and a new 105 that updates the Gauss 2.5. (I could keep them busy for a while if they let me.

 

[paulherrin]

Auto Focus ? It's gonna be a little while before it's suitable for narrative filming... Might be a great way to pull off a shot now and then, when you'll let a computer cut your hair it's probably time to let it pull focus. Someday.

haha, now i want to cast a computer for sweeney todd, set in 2012...

 

[Jiri Vrozina]

Yes. the 100mm macro is brilliant. The 50mm macro is pretty nice too. But Zeiss should be ashamed of the 50mm 1.4.

And yes, Nikon should make a new 50, now that they've shown what they can do with the 85, 35, and 24. And they should make a new 18 as well. and a new 105 that updates the Gauss 2.5. (I could keep them busy for a while if they let me.

+1
I wish ZEISS release new wide ZF/E Lenses soon .... new 15,18mm f2.8 would be nice..........