zooms or primes?

[WesG]

We're starting gear up for 3D with the eventual goal of shooting two Scarlets for 3D.

In the meantime we are playing with Canon DSLRs until we get Scarlets and are about to purchase our lenses. We are about to purchase a bunch of Canon lenses that will go on our Canons but then later be used on the Scarlets. Keen to buy smart now.

My question is whether you guys would recommend sticking to primes before zooms?

Zooms have the benefits of needing less lenses (travel), saving time for lens changing (lots of our work is on-the-fly),
however are zooms difficult to get accurate across two cameras? Are primes a safer route for matching framing of the two shots??

[Dan Sturm]

My 2 cents...Zooms are you friend in stereo. Since no 2 lenses are exactly the same, you'll need to align each lens before you can get good results with your 3d rig. Once you've aligned the zooms they will be aligned through the entire zoom range. Primes will require you to realign every time you replace the lens and you'll lose a lot more time. You can prep zooms a day ahead to save time as well. I wouldn't call primes safer. They just appear safer. They all require calibration.

[KETCH ROSSi]

Zooms definitely allow for a set up in which, depending on zoom range used, offer a much broader variety of shooting, were with Primes, the set up process might need to be repeated according to change of framing.

Unfortunately the matching calibration of zooms will in deed require additional work, as no two zooms act the same while zooming, as the elements come in to zoom, they do not do so in a strait fashion, but more of in a Oscillation fashion, making it very difficult to match in a calibration process, some are much better then others for this use.

But the best IQ is always achieved with Primes, even so I must say that some zooms closely match the IQ of Primes, while the zooms offer flexibility and time saving setting up.

[Alexander Ibrahim]

Zooms create all sorts of calibration and alignment issues in post.

Then again there may be some post tool I don't know about for handling the differences between the lenses.

I prefer primes. They have fewer and fewer types of problems.

They take longer on set though, as everyone else has said.

There isn't a perfect solution, if you want to do 3D you will pay in time somewhere.

[WesG]

Excellent advice guys, much appreciated. I had assumed primes would be pretty perfectly aligned and the simplist solution - sounds like this isn't necessarily the case.

Sounds like zooms might after all be a better solution with good calibration. Any tips on where we can learn about the process of calibrating zooms?

I'm not a fan of actually zooming so chances are we'd be using the zooms as variable primes effectively most of the time.

[KETCH ROSSi]

Just to clarify, the reason I prefer Primes is not only for the always superior IQ, again even so now days that gap is quickly shortening, but also because they bring better results, being that it easy to calibrate two primes then two zooms, for more then one specific, the internal alignment of the elements being a primary one.


As far as set up, Primes are actually faster to set up, but you will have to repeat the set up every time you change Focal, were with a Zoom you don't.

[Stephen Strangways]

When lenses are manufactured, they never perfectly match. That applies to zoom, and to focus on zooms or primes.

When Cooke or Zeiss build a cinema lens, they have batches of rings with different focus markings, to account for the variation. Each lens is tested once it is built, and then a focus barrel which has markings that match the actual focus distance as measured. They have batches, A, B, C, D... and when you shoot 3D, you usually want matching lenses, that is, two lenses from group B, to ensure the focus distance between the two lenses is matched as closely as possible.

Of course, with stills lenses, that sort of precision testing never goes into the manufacturing process, because they just can't afford to do it at the prices they sell the lenses for. So, you have to do the testing yourself, and perhaps spend some more time aligning them on set.

You'll want to test the lenses, to ensure that each is about equal in sharpness, zoom, focus, and zoom tracking (how well it stays centered as you zoom in and out.) Canon does turn out the odd lemon of a lens, so you want to ensure you don't get one of those. And, if they track completely differently while zooming, you'll have to realign them whenever you change focal length. You can see that easily: draw crosshairs on some acetate over a monitor or LCD, align it over a target of some sort, and zoom in and out to see if the crosshairs stay locked on the target, or wobble around in a lazy circle as you zoom.

[Mark L. Pederson]

Optimo Rouge lenses are fantastic for 3D work. They can be be tracked - and re-tracked to the specific camera - even in the field. We are getting great results with these lenses - quality is amazing. I would never shoot a feature in 3D with primes on the primary rig. WAY too inefficient IMHO. MX sensor helps BIGTIME with light loss from beamsplitter.

[Alexander Ibrahim]

Optimo Rouge lenses are fantastic for 3D work. They can be be tracked - and re-tracked to the specific camera - even in the field. We are getting great results with these lenses - quality is amazing. I would never shoot a feature in 3D with primes on the primary rig. WAY too inefficient IMHO. MX sensor helps BIGTIME with light loss from beamsplitter.

Well it sounds as though I really need to try 3D with some Cine zooms.

Has anyone tried 3D with RED Zooms?

Has anyone tried the new Fujinon zooms? At all? Maybe I ought to start a new thread about that.

[Pedro Guimaraes]

Once you've aligned the zooms they will be aligned through the entire zoom range.

You couldn't be more wrong.

My take on primes,

Primes are great for sharpness and repeatability. As they change very little every time they get on the camera. I agree they are very inefficient for the main camera to use on a beam splitter. Lens changes take forever with all the motors etc....especially with a AD breathing down your neck asking how long? every 2 mins..... Even Primes with vary in focal length to a certain extent from lens to lens. Only way to fix it is with diopters or much preferable in post. They are fast......but then again not sure you want to shoot 3D @ t1.2.

They are so repeatable that you can pre-establish what the geometrical corrections need to be made for each lens on your rig. Then even without a monitor you can get your rig aligned. For this reasons i like to use primes on our parallel rigs. I built a variable parallel rig for a project last year that when from 6" IO to 6' IO and we used Primes. This way after testing I knew how to shim each camera to achieve a perfect horizontal alignment. This allowed me to switch the focal length and apply the correct shims to achive a good alignment every time.

Zooms,
They are great for matching you magnification on set. Usually less sharpness etc...slower..... mainly really just speeds up production.

Yes the Angeniuex Rouge Zooms are AMAZING for 3D work. Beautiful quality, allows you to track them on camera, T2.8 etc........

highly recommended.

The RED Zooms are TOTAL crap for 3D. trust me. Multiple issues for 3D I discovered thru much testing.

[Riku Naskali]

Care to elaborate on multiple problems? Are you talking about matching the or something that isn't fixable in post? I'm pretty interested in this, since we are shooting stereoscopic with Red zooms in a couple of weeks. No way to afford any Zeiss primes, since we are getting our camera gear for free...