effects of changing fps for image sharpness

M

Marshallarts

Well-known member
Let me start by saying I've never much liked the "video" look (i.e. 60i). In the 80's and 90's before I knew anything about 24p, pulldown, NTSC, etc., I, much like others here, could always tell when stock footage was inserted in my favorite TV sitcoms. You know, when they cut away to a shot of a helicopter or something and it goes from a 24p look to a 60i look. Also I hate pan-to-scan for this reason and that just dawned on me since it's probably due to sampling/ scan rate not being 24p. That faster frame rate look really jars the viewer out of a 24fps experience. Pan to scan could also potentially create other issues I guess since it would inherently violate motion blur created during 24p capture unless the scan is made with this in mind. This thread isn't about pan to scan however but I want to take my time asking this because this isn't a simple "video look" question. I really want to hear educated opinions here.

I originally began thinking about this when reading about Ansel Adams and the inverse focal length (reciprocal focal length) rule in 35mm still photography. In 35mm photography it is best advised one not shoot any subject (even a still subject) at any slower shutter speed than the inverse of the focal length of the lens you are using. Ex: If you are using a 60mm lens you should use a minimum shutter speed of at least 1/60 second (even if you use a tripod)! Ansel Adams stated through experimenting his minimum shutter speed for a normal 50mm lens is 1/250 second, even more extreme! This is all so to reduce to the point of eliminating blurring due to camera shake and achieving maximum sharpness (since after all that's why we have the expensive lenses).

The problem in motion picture is obviously at a 180 degree shutter shooting 24fps the shutter speed is locked at 1/48th second. Now I know you can change the shutter angel in film or shutter speed in video and I know how that in turn will also change the image (introduce strobing, etc.) so please don't reply with that, let's think deeper.

We all want a great image. But why do we bother with the expensive gear and difficult techniques when true still photographers methods to achieve sharpness far exceed motions pictures capabilities? One answer would be to increase frame rate of capture. If we shoot 50fps instead of 24fps we effectively get a safer shutter speed of 1/100 second allowing us to use up to a 100mm lens (in 35mm cinematography) according to the reciprocal focal length rule.

Then I realize this would bring us closer to the video look of 60i. I'm not talking interlaced or progressive, only frame rate here. I've been trying to find a way to produce an example of frame rate change from say 18fps to 100fps for purposes of being able to "see" this change. As far as I know this hasn't been done, though please tell me if it has.

I hate the 60i look but may come to appreciate it if I can see how 24fps can become 60fps. I would also be interested in experimenting shooting at this faster frame rate for purposes of getting a sharper image. Perhaps in the future the 24p standard will change when people begin to demand sharper images. Truth is 24fps does have awful motion blur issues (just hit pause on any DVD). In achieving a more life-like look as one sees with their eyes maybe a faster frame rate is needed.
 
You could always just change the shutter speed. 24fps @ 1/250, for example. The result would be noticeably crisper pictures -- but less "film-like" because it would be more stuttery/strobe-y.

Increasing the frame rate would indeed make it look more like "video".
 
Ansel Adams had to worry about losing sharpness to camera shake in part because he was using a mechanical camera. Whenever the shutter flips up on a still camera it causes the entire body to shake. Since Adams was using view cameras with huge shutters, he obviously had problems with the camera shaking even when it was tripod mounted. This caused issues with critical focus and is why most cameras today have a mirror-lock up feature to get over this.

If you're using a digital camcorder you don't have this problem since they don't use mechanical shutters. If you're doing handheld or panning and want super sharp footage then yeah, go for a high shutter speed. But if you're on a tripod placed on a steady surface I think you should be just fine at 24fps
 
Marshallarts said:
In achieving a more life-like look as one sees with their eyes maybe a faster frame rate is needed.
Click to expand...


If you wag a pencil quickly back and forth in front of you, you will see that our eyes have a huge amount of motion blur, way more than what you capture with 1/48s shutter and 24fps.

Which is very interesting.
 
thanks for the feedback everyone. Barry, yeah I was thinking about the increasing shutter speed method but my concern was about the strobing effects it creates. And Dan, I've been reading some interesting studies (sorry, can't find the URL for my life) about human eye and visual perception specifically concerning motion blurring and theoretical fps the eye can perceive. The old notion of persistence of vision and 24fps isn't true as a maximum fps of the eye. Recent studies suggest the human eye is capable of noticing differences up to 200+fps (I think I read this when discussing gaming but honestly I'm not sure, in fact I've read other articles saying up around 90). But truthfully this isn't true since (and i'm really upset I can't find this article, I think it was on wikipedia but can't find it now) newest research show vision isn't perceived in frames but almost more like mpeg-4 encoding where static objects are not refreshed but moving ones are. It's a much more complicated study but it also discussed the purpose of motion blur when, and using your example, you wave a pencil in front of your face.

davide, thanks for your input as well. I didn't take into account the shutter issue and I've heard of the benefits of mirror lock-up to correct this. Still I think his experiments were on 35mm cameras but as you say the shutter wouldn't be as damaging on a digital camcorder. Wouldn't panning at a faster shutter speed only increase an unwanted strobbing effect?

With how big the hand held aesthetic has become, I wouldn't say cinema verite because IMO it is entirely overused and used incorrectly at that, it's amazing how much motion blurring is introduced into our films.

It is interesting that higher frame rates in fact give you a sharper image and this is what I want to experiment with. If you could create a very sharp motion image with a fast refresh rate perhaps your brain would still perceive motion blur when you wave a pencil in front of the lens. Maybe we'll need to rid the black side of the 180 degree shutter for this to become most lifelike.

Mostly I would LOVE to see a demonstration of 18fps becoming 100fps so we can view this change. This is something that would only be viewable on a monitor but it would be a VERY interesting demonstration to watch!!
 
I dont' have any examples that fit that request. But you could try finding some Showscan footage. That's 70mm film shot at 60fps. I've never seen it but it's supposed to look extremely realistic and immersive.
 
Yes, at 60fps it would feel "realistic" but that's also why people like 60fps for sports shows on HDTV - the thing is, that's what a video camera is for!

Remember a couple of things - one, 24fps turned out to be a great choice by film-makers because it's *NOT* realistic. That dreamy, caught-up-in-fantasy effect that we get into when watching a film comes in part from our brain recognizing that we're seeing something we can't quite put our finger on, something decidedly not real. I was in a store the other day and a really big, expensive HDTV panel with 120Hz refresh, had a feature that used adaptive interpolation to show 24p movies interpolated up to 60Hz (or even right to 120 but I suspect it was 60) It made multi-million dollar hollywood look just like it was shot on a video camera!! There's no way film makers want the audience to see their film like that - if they did they would gladly shoot film at 60fps.

Second, motion blur might be more natural than we think. What I mean by this is, that our brain can interpret as sharp, what our eyes see even with motion blur, especially since we're so used to that blur. This makes the "rules" quite different between static photography, and video/film. I notice that most photographs can be quite pleasing with som "Unsharp Mask" commonly used to sharpen in software. But you may not like the effect of that same filter when used on video. Try sharpening each individual frame with USM and it can do un-natural things to the motion blur, not adding any apparent sharpness when viewed back as video.

Sorry to blather on, but I think the glass & the focus to get sharp footage, and clean editing pipelinel to preserve it, will allow our brains to see the footage as sharp. Changing the shutter speed will add un-wanted side effects as others have mentioned - strobing at one end and the "video look" at the other (most video cameras use 360-degree shutter ie full 1/60 second for each field - you can sometimes get interlaced footage to look more film-like by dropping the odd fields, and scan-doubling the even ones to get the 180-degree "look" although of course you sacrifice resolution.)

Anyway that was a little long-winded, but it's a topic I find interesting, so you got to hear my ravings... happy shooting.

Dip
 
dippytwitty said:
Anyway that was a little long-winded, but it's a topic I find interesting, so you got to hear my ravings... happy shooting.
Click to expand...

Dip, your ravings were much appreciated. Thanks to everyone who contributed to this because it was something I've been very curious about though all along I knew in the end I would be wrong. As with Dip, I find these topics interesting and love discussing it. I would still love to see an example of fps shifting from 18-100 though:evil:. I think it would be a pretty cool experiment. I'll get to it as soon as I get to making my movie.

Also, it was interesting to read about ShowScan as Davide discussed. I would love to see that! I did some researching about it and found a very interesting thread very much debate. It discusses ShowScan, 60fps, and addresses my original concern. The author of this thread referrers to a higher frame rate as "temporal high definition" which I found very interesting. Unfortunately the thread is mostly people yelling at him because of the way he argues that a high definition temporarily at 60fps is as important as the common definition of high definition (which he calls spatial HD). Still I learned a lot reading it.
http://www.cinematography.com/forum2004/lofiversion/index.php?t15157.html
 
C'mon guys -- Adams was a STILL SHOOTER and his "rule" is a good one for creating still images of the kind Adams was known for.

Blur and varying degrees of image sharpness are all part of the tool box to create the look & feel a director (you) wants. Look at the (now infamous) "saving private ryan" effect - fast shutter speed, reduced motion blur, & perfect for the feel Speilberg was trying to achieve. Would you shoot a romance scene like that? Maybe not - your choice. Heck, I don't know that I'd shoot an action scene like that. It's all about taking advantage of the ability to join motion and sound for a media that is (so far) the most involving for an audience - make 'em feel they're "there" while they sit in a theater or in front of a TV. "Rules" are useful, but sometimes they're the first thing you should toss out.

Case in point - two of my favorite music CDs this year are Iron & Wine "The Shepherd's Dog" and the collaboration between Robert Plant (led zep singer) and Allison Krause (noted bluegrass queen). Both are rooted in very traditional folk sounds - Iron & Wine having a real CSN vibe, and the Plant/Krause release being very "Americana". In both cases, there's a lot of modern production quirks & tricks that could seem questionable for the material, but in the end really amp up the emotion & drama. They'd both be fine CDs if the producers had followed "rules" about recording "folk-ish" music; but in the end, quirky, creative production enhances the experience to the point they become very affecting works, the kind that sort of lodge into your mind in a cool, hypnotic way.

Sorry for the ramble, and tech-geek stuff does indeed blow my dress way up and I'll analyze your FPS tests til my wife rolls her eyes... but to me, blur good! Grain good!
 
Marshallarts said:
Dip, your ravings were much appreciated. Thanks to everyone who contributed to this because it was something I've been very curious about though all along I knew in the end I would be wrong. As with Dip, I find these topics interesting and love discussing it. I would still love to see an example of fps shifting from 18-100 though:evil:. I think it would be a pretty cool experiment. I'll get to it as soon as I get to making my movie.

Also, it was interesting to read about ShowScan as Davide discussed. I would love to see that! I did some researching about it and found a very interesting thread very much debate. It discusses ShowScan, 60fps, and addresses my original concern. The author of this thread referrers to a higher frame rate as "temporal high definition" which I found very interesting. Unfortunately the thread is mostly people yelling at him because of the way he argues that a high definition temporarily at 60fps is as important as the common definition of high definition (which he calls spatial HD). Still I learned a lot reading it.
http://www.cinematography.com/forum2004/lofiversion/index.php?t15157.html
Click to expand...


Glad you found it interesting. In all fairness I think the people 'yelling' at the poster were right. David Mullen is one of the most talented and knowledgable DPs to post on boards like this and I agree with his argument in that thread. The number of FPS has nothing to do with HD, SD, 4K or whatever. I think using a term like temporal high definition to describe high shutter speeds just adds confusion (as if broadcast standards weren't nebulous enough). If we do that then why not say dynamic high definition for footage that has a high dynamic range?
 
You must log in or register to reply here.
Back
Top