More RED DRAGON footage - skin tones

I don't see what's so unlikely abut different cameras rendering motion differently, they render EVERYTHING else slightly different. DSLRs for instance can't be set to exactly 1/48, they also have rolling shutters with different read out speeds from each model. The Alexa has a very fast readout, the BMC4K has a global CMOS... The DVX100B had CCDs and was very natural looking with 24P.
I bet all of these things contribute to how a camera depicts subjects in motion. Doesn't seem all that difficult a concept to grasp.
 
maarek said:
The only difference is rolling shutter. If you use two cams that use a global shutter, there will be no difference in motion.



There's no such thing as "film-motion". Motion pictures are usually shot quite deliberately with nice dolly shots and planned handheld footage. A lot of the "film-motion" comes from great dolly shots and the way camera is moved. You give that same film-camera to some guy just shooting in his backyard a bunch of flowers and it will look just as bad as any other homemade video.

Except nowadays CMOS-cams have the rolling shutter issue that does make handheld shots look quite insane if your not used to them.
Click to expand...
I completely and utterly disagree with what you've said. Film has a distinct motion that only very expensive/or new and still expensive digital cams can replicate.
 
pulpfiction007 said:
I don't see what's so unlikely abut different cameras rendering motion differently, they render EVERYTHING else slightly different. DSLRs for instance can't be set to exactly 1/48, they also have rolling shutters with different read out speeds from each model. The Alexa has a very fast readout, the BMC4K has a global CMOS... The DVX100B had CCDs and was very natural looking with 24P.
I bet all of these things contribute to how a camera depicts subjects in motion. Doesn't seem all that difficult a concept to grasp.
Click to expand...

It boils down to someone just being incapable of perceiving it. If they can't see it, it doesn't exist.
 
pulpfiction007 said:
I don't see what's so unlikely abut different cameras rendering motion differently, they render EVERYTHING else slightly different. DSLRs for instance can't be set to exactly 1/48, they also have rolling shutters with different read out speeds from each model. The Alexa has a very fast readout, the BMC4K has a global CMOS... The DVX100B had CCDs and was very natural looking with 24P.
I bet all of these things contribute to how a camera depicts subjects in motion. Doesn't seem all that difficult a concept to grasp.
Click to expand...

So what you're saying is that it is the rolling shutter that makes it non-filmic. If it is a very fast rolling shutter, or a global shutter, then it is cinematic? I would get that.
 
Last edited:
Samuel H said:
So what you're saying is that it is the rolling shutter that makes it non-filmic. If it is a very fast rolling shutter, or a global shutter, then it is cinematic? I would get that.
Click to expand...

I think it comes down to the entire system, but yes definitely including the shutter design/readout. Every camera seems to see colors a bit different just on a static shot. Motion is an extremely complex thing (not even talking about the camera being moved where it becomes exponentially more complex)
A digital system has to "build" an entire picture and continue to rebuild it 24 times a second as a subject/subjects move through and around the frame, so every little thing the camera has in it's design will contribute to how it does this. Questions for motion rendering could be: how fast is the readout, how complex is the design of the camera for motion images, does it have a strong enough codec to make motion crisp rather than mushy. Seems to me there should be differences given the complexity of what has to be accomplished by any given camera.
 
Are there any papers on film motion? Would love to read up on the science of it, if it exists.
 
I've been looking for some more info on mechanical movie camera shutters from the film era, and it seems that they are indeed a rotary shutter. Unvelievable.
(the best explanation I could find is, of course, in this forum: http://www.dvxuser.com/V6/showthrea...-on-5D&p=1986320212&viewfull=1#post1986320212 )

So: why hasn't anybody developed a NEX-to-whatever lens adapter with a rotary shutter built in? Even m43-to-whatever, for the BMC.
For a very fast 5ms rolling shutter on a 16:9 image, the diameter of the disc has to be around 3.5 times sensor width: 87mm for a S35 sensor.
The only challenge I see is how to make sure it goes in-sync with the camera's frame capture.

(it gets a lot more complex if you want it to have an optical viewfinder, as in film cameras, but I think we can live without that)

edit: I see where problems start to arise: if the camera's electronic rolling shutter is not very fast to begin with, that challenge of getting the rotary shutter in sync with the frame capture of the camera becomes a lot bigger. The BMC, for example, as a sensor read-out time of 25ms (link); that's more than (1000/24)/2=20.8ms, so there's no way to put a 180º rotary shutter in front of a BMC. Same for basically every DSLR except the GH3. The FS100, RED cameras, etc, could get one, but, even for them, sync error has to be <10ms for this to work
 
Last edited:
tmnt said:
Are there any papers on film motion? Would love to read up on the science of it, if it exists.
Click to expand...

A paper by Alan Roberts here:

http://www.bbc.co.uk/rd/publications/whitepaper053

One downloaded you have to scroll down to the parts about motion...it gets very scientific, lol.

From another forum I read some of Graeme Nattress thoughts about how edge enhancement/sharpness alone can change how we perceive video motion differently than film.
 
Last edited:
Samuel H said:
That can't be. If it is a rolling shutter, it has rolling shutter issues. Don't modern cameras use anything more advanced than that? Or maybe they just use a bigger disc so that the transition is faster in relation to exposure time...

Also, regarding motion blur in film and digital, I would tend to aggree with what maarek just said: if you can't point to a controlled test where I can see the difference, I doubt it exists.
Click to expand...

Found this picture regarding the F65 variable shutter disc :

F65QLIMG_6424.jpg
 
Yes, I ran the math and it's not as bad as I thought it would be: with a disk around 4 times as wide as the sensor, like the one in that pic, rolling shutter time is around 4.3ms. Getting it in sync with frame capture is the big issue, but if you're the camera manufacturer then it should be a piece of cake.
 
maarek said:
Rubbish. If you can't point it out in controlled testing, it doesn't exist.
Click to expand...

So here you say I'm wrong...

maarek said:
Now rolling shutter does make motion look quite odd with the F3 and DSLR's. A lot of these huge "motion differences" are attributed to that one factor. Alexa and RED Epic are quite speedy but still have it slightly.
Click to expand...

Yet here you say I'm right...

maarek said:
If your 5d is "incapable" of filmmotion, you are clearly moving it wrong.
Click to expand...

I stay extremely busy, so I'm obviously doing something right. Probably partially to do with the fact that I can pick up on subtleties.
 
I think what he's saying is that slow rolling shutter is a difference, but there are no more differences beyond that, even though some people would argue there's something deeper that a digital sensor with very fast rolling shutter or even global shutter won't be able to replicate. Those would be my thoughts too right now.
 
I can't believe no one has mentioned Tessive.

For years I pined for global shutter, because I couldn't stand the thought of my video having any skew at all. But after reading Tessive's site and watching their videos, I'm convinced that even global shutter is not enough. I know, I sound picky, but just go there.

Then the ones who claim some cameras do render motion differently can talk more intelligently.

Explanatory video: http://tessive.com/the-time-filter/
Demo test footage: http://tessive.com/gallery/
 
Samuel H said:
I think what he's saying is that slow rolling shutter is a difference, but there are no more differences beyond that, even though some people would argue there's something deeper that a digital sensor with very fast rolling shutter or even global shutter won't be able to replicate. Those would be my thoughts too right now.
Click to expand...

Pretty much. I actually really hate rolling shutter. With a passion. That Tessive shutter just basically looks like a 270/360 degree shutter. It's pretty telling that none of those example shots are at 360 degrees.
 
roxics said:
I honestly can't tell the difference between any of this footage anymore. Be it a GH2/3, BMC, Canon, Nikon, GoPro, whatever. It all looks the same to me these days. The only camera that looks drastically different to me is the Alexa. Sure I can see the dynamic range difference in the raw footage, but beyond that, the effect is the same.

Now the motion is different. Maybe that's why I like the Alexa, it moves like film. Granted the Red does too, but not as much.
Click to expand...

Most quotable PullQuote of The year! I will incorporate it with attribution into my signature if I may !

Or as I like to put it... sooner or later, everything will end up on a 5"-10" tablet at 1080p or yikes- Google Glass HUD !
 
You must log in or register to reply here.
Back
Top