Warning. Read this before you watch so you understand the context and what I was looking to test…
One thing I find with digital imaging cameras is that they can sometimes be a little too perfect.
There’s no room for messing around on the edges with digital cameras. Take a DP like Harris Savides who’s reputed to have a recipe for cooking his film in an oven. This way he would artificially age and de-nature the film to get a “look”. Whilst I’m sure there are those that think they could replicate this in post, there’s nothing like the randomness of working this way where you’re literally baking in a look. At a fundamental level it affects the way the film itself works as a medium. The very randomness is exactly what we’re chasing.
There’s also bleach-bypassing, cross processing the film or fogging and flashing the film. In these cases you’re messing with what Kodak would say is the “right” thing to do but you’re getting a kind of “mistake” that ends up great for a look. Again, when you’re messing with the building blocks you get things that can’t be replicated in post. Anything to do with exposure works in this way.
This process is very hard to do on a digital camera. They tend to work, or they don’t. When they only kind of work the mistakes generally don’t look so good either.
One way to build in some imperfection into the image with digital orgination is through optics. And a really easy way to do this is with vintage or old lenses. Lenses that are more optically imperfect have their own look that works great with modern digital image making.
Older lenses aren’t designed with the same tolerances. They don’t have the same raw materials and have more impurities. They age and discolour. They might only have a single coating rather than be multicoated. They might not have any coatings at all !
But they do have a look that then becomes built into the visual language of the show you’re doing.
I’m just doing some tests right now for a show that we previously shot using the Zeiss MK 3 Superspeeds. Mark 3′s are optically the same as Mark 2′s with better witness markings and numbers on the barrels. From memory, they were introduced in the early 90′s, but they were based on the same optical design of the early 80′s.
We did actually a lot of testing with the original first series and ended up using Zeiss as well as some older Cooke zooms. You can read a very detailed account of what we did in the first series with example clips here.
With the new 4:3 alexa we also tried something I’ve been dying to do on Australian Television….shoot anamorphic.
A more acceptable way to shoot anamorphic is to shoot using a 4×3 sensor camera like the Alexa and then centre cut the 1.78 from the 2.40 image. This way the network get a 1:78 image deliverable and you can still master in 2:40 so that on DVD for example, you might have the option of releasing it that way. This is a lot like what we used to do when shooting 16×9 and protecting for 4×3, creating a 16×9 master and then doing a sub master that was a 4×3 centre cut.
We actually tested this during the first series and we found the drop off in resolution form a 1:78 camera was far too great. I was interested to see what a 4:3 sensor would do. Obviously it would be even more ideal if there were more anamorphic lenses out there that could do a 1.3X squeeze instead of 2X.
Most of the worlds anamorphic lenses are a 2X squeeze but some companies like vantage are doing a 1.3X version of their very nice HAWK anamorphic lenses. I was lucky enough to shoot with these somewhat rare 1.3X lenses recently on a film that’s about to be released, 100 Bloody Acres. You can read about it here.
So we were also able to wrangle from Panavison a set of Cooke XTAL 2x anamorphics lenses, which are also of a mid 70′s design.
What you’re about to look at here if you can be bothered sitting though the 50 odd minutes this test goes for is a comparison between the Pvintage, the Zeiss MK3 Superspeeds and the Cooke XTAL anamorphics, matching as best I could the focal lengths. For example, Panavision do a 75mm where as the Zeiss is an 85mm.
This was shot using an Alexa PLUS 4:3 and was shot using mainly practical lighting. There is a blanket LUT that’s been applied to this which will be the rushes LUT for Puberty Blues. So no real grading. We also shot using the 2K mode of the Alexa in ProRes 4:4:4 though Ive since heard that this can sometimes introduce additional aliasing. I’ve got to test that still.
The anamorphic has been centre cropped to 1:78.
There is a day exterior, but we were late getting to location and the light really changed quickly so it’s probably not the most fair comparison.
In the first setup, I was really interested in the flare characteristics of each lens, so I set something up with the help of the are department and a dedo light to deliberately look at that.
We were also interested in how the out of focus areas would look in the very very far distance so we looked at a night exterior with a deep horizon of lights.
I tried to shoot at the same T stop, to match the slower Anamorpics as well, but you’ll see I also did some wide open shots hand held towards the end.
We actually had two cameras (one was PV mount the other was PL) and something like 17 lenses to test so we didn’t have a lot of time. My excellent focus puller Pim Kulk did a great job considering it was unrehearsed and we shot only single takes of each lens.
In the first flare test, I also found the Fluro at the end to be distracting in anamorphic so I panned it off and shot two setups again. I have included both takes of this to compare.
At the end there are also some lens charts, plus a lot of additional Pvintage focal lengths that don’t have a matching focal length in the Zeiss Superspeeds. You’ll also notice the 40mm Pvintage is very warm and has since been sent of for a visit to the lens doctor…
Let me know what you can see. I’ve made the file downloadable in Vimeo should you wish, but it’s a smidge under 5gb…
So yeah. There’s no sound. Quite frankly it’s a lot of stuff to watch and even I find it a little bit tedious, but I think there will be those of you out there that will be interested enough to take a look. To me the differences are pretty subtle.
So be warned…. it’s not riveting television, but if you want to see the differences, they’re there to look at.
Special thanks to Patrick Rohr for modelling. Next time you’ll be going in the water.
** Just noticed I mislabelled the Zeiss superspeeds as MK2′s when they are in fact MK3′s. As i mentioned,there’s no optical difference between them.