This page has found a new home

DSLR Filmmaking Blog

Blogger 301 Redirect Plugin DSLR Filmmaking Blog

Tuesday, October 22, 2013

Choosing follow focus gear rings

Canon 5D Mark II w/Fotga DP500 II follow focus and generic 60-70mm fixed diameter ring
The cinemazation of DSLR cameras have created a need to add 0.8 pitch follow focus gears to standard photography lenses. Many such gears exist on the market; however, not all are created equal - a fact that is not immediately apparent until you've already experimented with a few.

The prevalent types of follow focus gears available to the prosumer cinematrographer today come in three different variants:

  • Left: Plastic, adjustable diameter gear ring with friction lock and lever
  • Center: Rubberized, adjustable diameter gear ring with tooth lock
  • Right: Fixed outside-diameter ring w/10mm variable inner gear ring diameter (various sizes available), made by Cinematics and others
I've used all three in the field - following are my thoughts on each:

Semi-hard plastic, adjustable gear ring with lever

The adjustable plastic gear ring is my least favorite in that the toothed reverse of the plastic isn't that grippy. I've yet to have it slip, but it does not inspire confidence. Likewise, the lever locking mechanism of the lever is cheap, damages the gear teeth, and the lever gets in the way of follow focus units unless oriented to the opposite side of the rig.

On the other hand, the plastic gear teeth engage well with follow focus units, and it is especially handy when using an older, manual telephoto prime for focus control.

 
Rubberized adjustable gear ring
The rubberized gear rings can are soft enough to grip a lens - and grip well - but this also causes the teeth to flex while meshing. The toothed receiver also works well to keep the ring tight.

Unfortunately, coupled with small-diameter lenses and a follow focus mounted in certain places, these rings will jump teeth - and visibly so on screen. Nevertheless, they are effective on larger-diameter lens barrels.


Cinematics 80-90mm at left, generic 60-70mm at left
These fixed-diameter rings are my personal favorites, having the ability to reach a follow focus unit with ease, regardless of lens barrel diameter. This style of gear ring is presently available as a generic (only one at present, it seems) and by Cinematics with slight differences between each. Both companies produce inner diameters ranging from 60-70mm up to 108-116mm. In my experience, the generics have enough adjustment to fit barrels down to 57mm comfortably; possibly smaller.

The generic's plastic-knob bolt is durable, despite its cheap appearance.

Cinematics' units have a smoother finish than the generics and have an allen head bolt with a proper steel nut, but otherwise show no superiority over the generics in operation or durability. Unlike the generics, they are available in red and blue (I've also seen purple) as well, which is fine if you're a hipster who must color-coordinate your camera rig to your fixed-gear bicycle and Macbook. Personally, I see no point to it - why call more attention to an L-series lens by putting red gear rings on it?

Additionally, a pair of these rings can provide both focus control and a precise zoom control for lenses with fixed zoom rings - such as Canon's L-series. Given the lack of perfect video-quality fluidity in even the L-series lenses - not to mention the fact that most photography lenses are varifocal - controlling zoom through a second gear ring and follow focus unit can be a stress-free method of ensuring a perfect take on the first try.

-Kurt "Man with the 5D"

No monetary compensation was provided for the mention of the any of the products mentioned or shown in this article.

Labels: , , , , , , , , , , , , , , , , ,

Monday, September 16, 2013

CMOS Rolling Shutter Removal - Plugin Shootout!

I'm getting away with whip pans like these.


Or just about, anyway.

That said, I'm not going to bore you with an explanation of what rolling shutter is. Chances are that - before you reached this blog - you already went through the torture of a few hundred posts and videos of everyone else demonstrating rolling shutter - without the slightest suggestion of a good CMOS skew and jello removal tool.

You can relax now, for I'm going to show you a comparison between the five filters that are available for doing just that, and the most professional one of the lot.

Better yet, the best one isn't even the most expensive.

Following is our list of candidates:
  1. The DeShaker plugin for VirtualDub:
    • Free; requires that you convert your footage to AVI before importing. 
  2. Adobe's After Effects' Warp Stabilizer, available since CS5.5:
    • Requires purchase of one of the Adobe suites. Not necessarily cheap.
    • Note: CS6 is equipped with a dedicated rolling shutter corrector, but I did not have time to download CS6 to try it.
  3. ProDAD's Mercalli PRO Version 2
    • $249 from the ProDAD website
    • Sony Vegas - since version 10 - contains a simplified version of Mercalli; we'll cover only the standalone.
  4. New Blue FX's Video Essentials III plugin, which contains a rolling shutter removal tool.
    • Works with Adobe After Effects, AVID Media Composer, and Sony Vegas, amongst others. $130, plus the cost of your editor.
  5. The Foundry's Rollingshutter plugin for Nuke and After Effects
    • A cool $500.
Do note that I'm taking this from the point of view of PC's, though MAC users have a rolling shutter removal tool in Apple's Final Cut 10, and the Foundry's plugin is also MAC compatible.

Seeing as there is no better way to see the results than a video, I prepared three rolling shutter examples and ran them through each of the five filters:

 

Though not covered in the video, program usability presents a concern as well; after all, a timely workflow is essential to smooth editing.

Conclusions:
  • DeShaker is an absolute nightmare to use. Besides having to convert footage to AVI (I used Pazera MOV to AVI to do so - it was quicker than Adobe Media Encoder) before even importing it, setting the plugin up in VirtualDub is a hassle, and it takes ages to analyze the footage. You'll wind up with some pretty big AVI files to clean off your drive too unless you install and compress with the x264vfw plugin.
  • Warp Stabilizer, along with the New Blue FX and Foundry plugins run the same as most other effect plugins for After Effects. New Blue FX had a tendency of crashing AE at times, while The Foundry's plugin can become a long and drawn-out process.
  • Mercalli is ridiculously fast, and arguably the most user intuitive. It almost looks too simplistic and mainstream to be a professional tool, but - as proven by the footage tests - it is superior to everything else as a general-purpose tool. 
    • At best, the only further improvement would be to correct individual extreme skew issues with New Blue FX before importing into Mercalli. With certain footage, this workflow could remove rolling shutter skew entirely.
One last option remains for CMOS skew removal (exclusively - this method will not correct for jello/jitter), if you do not mind a bit of DIY: Layer seven or eight feathered copies of the problematic footage over each other, with each successive clip (from top to bottom) playing progressively slower than the first - thereby correcting for the top-to-bottom scan delay in most CMOS sensors.

on Vimeo has a brilliant video tutorial of this method: http://vimeo.com/10130633

Regardless, a powerful filter does not negate that one can ignore CMOS rolling shutter when shooting, expecting that you can fix everything in post. That's not a reasonable though - you simply cannot walk into a shoot and expect to get away with it.

If absolute perfection is what you need, the Black Magic Production Camera 4K - with global shutter - should be your next purchase, rather than a rolling shutter corrector.

In fact, I fully intend to become the Man with the BMPC 4K someday, but that's the future.

Where did I leave my DeLorean?

No monetary compensation was provided for the mention of any of the products listed. All were tested during product trial periods or were purchased by myself.

Labels: , , , , , , , , , , , , , ,

Friday, August 9, 2013

Compact-Macros for DSLR filmmaking

Of the various Canon EF lenses I've experimented with for video use, the only one that has consistently impressed me (surprise) as an unusually good value - as a used lens on eBay - is the Compact-Macro 50mm f/2.5:


No, Canon did not pay me to say this (sadly).

Granted, we're still referring to a lens that has aperture limitations for cine use, as - like all Canon EF lenses - the iris is controlled electronically from the camera body and cannot be de-clicked.

Yet, the lens has a beautiful bokeh, and a price range of $150-200 on eBay.

I'd say that's a pretty reasonable expenditure for a lens that can give you images such as these - almost consistently through its range to boot:





These were taken as photos (and were the most convenient example images I could locate when writing this post), but they're no less representative of the results the C-M 50mm yields for cine use, with exception to overall resolution.

In short, it is quite difficult to ruin a shot with it, so long as you're not transitioning from a dark to light environment or vice-versa - which wouldn't seem to be the most common shot to use a 50mm macro for to begin with.

Granted, the need for a lens of this kind on a shoot is uncommon, but when you need a close-up lens to emphasize the slightest detail when shooting a detective thriller, you'll thank your lucky stars that you have it.

My only criticism - and it seems to be shared by most users of this lens - is that the focus ring is one of the worst in existence. It either feels like a cheap toy, or it binds like a cheap toy - but not enough to become a problem if you're not pulling focus during a shot. So be it; seems to be the case with most of Canon's lenses, save for the L-series (and even then, the L's do not come close to the feeling of precision that one gets from most mid-range lenses of 30 years past).

-Kurt - Man with the 5D

Labels: , , , , , ,

Saturday, June 22, 2013

I love it when a rig comes together.

The last bits to (satisfactorily) complete my 5D Mark II rig arrived today:





Though I've run it in various configurations previously, the final setup is nothing short of a joy to use.

-Kurt

Labels: , , , , , , , , , , ,

Friday, June 21, 2013

Lens sharpness does matter, Mr. Ken Rockwell.

Warning: This is a controlled rant alert.

When I first started with a DSLR, one of the first websites I found was Ken Rockwell's lens reviews. Chances are, you've seen his site too - seeing as they rank highly on Google - and he's pretty darn good.

Nevertheless, there is one article on his site that I beg to differ with: Lens Sharpness Doesn't Matter.

Well, Ken - sorry to disagree with you, but sharpness does matter. Sure, lens sharpness may be overrated - especially when comparing some of Canon's mid-range zoom lenses to L series offerings with shockingly similar performance - but it is a topic worthy of some finesse.

Thing is, sharpness is not an all-inclusive definition of a lens. Sure, a lens may be sharp in its center, but that may be the only place it can claim that fame.

Two, older EF-mount lenses prove that point:

Canon 22-55mm f/4-5.6 USM
Tamron AF Aspherical 28-200mm f/3.8-5.6, EF mount

Though both of these lenses are no longer on the market, they're generally quite affordable on eBay - making them appear to be quite attractive to the DSLR filmmaker looking for a dirt-cheap starter lens.

Sucker alert.

If lens sharpness did not matter before, these two lenses ought to change the course of things. There's a certain manner that both of these lenses handled areas out of focus and those areas supposed to be in focus that the line became quite blurred (pun not intended) between what was what.

That, and vignetting (darkening of the edges) occasionally proved to be an issue as well.

Cases in point - from the Tamron:



It's not that this photo isn't sharp (though there is enough purple fringing on all the edges to make it appear as if the camera was shaking slightly during the exposure), but the sharp areas and blurry areas all look about the same.

Technically speaking, the difference between these areas are known as the "circle of confusion" (explained very well here: http://www.cambridgeincolour.com/tutorials/depth-of-field.htm), but - for the purposes of this basic explanation - let us just call it for what we see in the photo: It is a circular, muddy mess, with vignetting to boot.

Mind you, this is a photograph taken in RAW, not video or .JPG, so you can imagine what the results would have looked like with the extra compression (or reduced scan lines as the 5DM2 uses for 1080P). Not so good.

The Canon didn't fare much better:


Aside from obvious fringing, the out-of-focus areas in harsh sunlight look as if the camera was mounted to a pogo stick. Granted, I could have put a gradual ND filter on it for these photos, but I doubt if I would have improved anything other than the exposure.


It's neither sharp, nor pretty - it's muddy.

In closing, lens sharpness does matter, though the overall sharpness of an image is more a result of whether the lens does a nice job of separating the out-of-focus areas from those that are in focus.

Be mindful of what you're buying, and remember this when you find a low-end lens for cheap:

"In the field, bad lenses are worth half of what you paid for them."

-Kurt K. - "Man with the 5D"

Labels: , , , , , ,