Will CMOS work for me?

[Aric Mannion]

I mostly shoot fast action and green screen work. I've never shot with a CMOS camera before, and just started researching it. I'm now concerned that the fixed Scarlet (or any other CMOS camera) won't work for me at all.
If I have footage of a backflip should I expect a slanted/distorted figure? If I stabilize this footage, will it truly become a wiggling image? Or has CMOS come a long way since these articles?

http://www.dvxuser.com/jason/CMOS-CCD/
http://www.ptgrey.com/support/kb/index.asp?a=4&q=115

-I know we can't judge the Scarlet yet, but I am wondering if CMOS is not recommended for this imagery as a general rule.

 

[Graeme Nattress]

You'll probably be fine because the shutter in Scarlet is so very fast you'll probably never see it's effects.

Graeme

 

[J. Eric Camp]

Watch Gamer or Book of Eli. Tell me if in the action scene's you see any slanted figures: red CMOS in action.

Also most commercials you see these days that have slow-mo.... a different CMOS in action.

You seem to be combining attributes too. CMOS and rolling shutter are not synonymous. Additionally different sensors will have different read reset times.

 

[PatC]

Not all CMOS are created equal.

 

[Jeremy Kromberg]

I can't really imagine the MX not working for anyone but I am a Red fanboy.

 

[Noah Kadner]

Sort of like asking can I drive with unleaded gas. Get used to it- CMOS is in more camera models than it's not these days...

Noah

 

[Gavin Greenwalt]

With the much less noisy blue channel, decreased compression artifacts and faster reset time the Scarlet should be more than adequate for Greenscreen/Tracking. Certainly better than anything in its price range.

 

[brian hanson]

i've had the same thoughts, i shoot a majority of action sports. there is a skateboarding shot in the mx reel that looked great.

 

[Graeme Nattress]

So to summarize:

short answer - Yes.
long answer - most probably yes.

Graeme

 

[george manzanilla]

So to summarize:

short answer - Yes.
long answer - most probably yes.

Graeme

I'm new to the red product line, mostly because we have still been using 16mm and ccd based cameras, and quite frankly this is the only issue i was concerned about.

When you say the "shutter speed in Scarlett is so very fast", does this mean that if i shoot an image with a 1/120 speed as opposed to a 1/500 shutter speed we may see a varying amount of rolling shutter artifacts? if one shoots at a slower shutter speed, will there be more of the jello artifiacts that i've seen plague footage shot on Canon 7ds etc.

One of my concerns is stabilizing footage. I shoot a lot of handheld, i know there are filters to correct this type of cmos rolling shutter effect, but having to do that all the time may not be time efficient all the time. thanks for any knowledge you can share, much appreciated.

 

[Sean Finnegan]

Watch Gamer or Book of Eli. Tell me if in the action scene's you see any slanted figures: red CMOS in action.

I saw Gamer at a pre-screening at my school because Brian Taylor (the director) is a graduate. I'm not really a fan of the look he likes for his movies (which is pretty consistent from Crank, to Crank 2, to Gamer). Nice guy though. At any rate, those films are shot at extremely high shutter speeds. For those of us who prefer a more classical "film look" with a shutter speed of 1/48, will we see rolling shutter effects with camera movements or moving subjects?

 

[Gavin Greenwalt]

When you say the "shutter speed in Scarlett is so very fast", does this mean that if i shoot an image with a 1/120 speed as opposed to a 1/500 shutter speed we may see a varying amount of rolling shutter artifacts? if one shoots at a slower shutter speed, will there be more of the jello artifiacts that i've seen plague footage shot on Canon 7ds etc.

No. The best way to think about it is a song sung in a round. The first row of pixels starts singing "Frere Jacques..." and then the second row and the the third row and so on and so forth. Eventually they're all singing at the same time. And then the first row finishes the song (its exposure length) and then the second row etc, etc... Each singer sings the same length of song regardless, each row exposes for the same amount of time. The delay between rows/singers starting is how much 'rolling' shutter there will be which is what Graeme is referring to by "fast shutter". The rate at which each pixel row starts its exposure is much faster.

The Mysterium-X's delay between each row starting its exposure period is much shorter than Mysterium's so there are less rolling shutter artifacts.

This is in contrast to a global shutter which is like a choir all starting at the exact same time.

 

[george manzanilla]

No. The best way to think about it is a song sung in a round. The first row of pixels starts singing "Frere Jacques..." and then the second row and the the third row and so on and so forth. Eventually they're all singing at the same time. And then the first row finishes the song (its exposure length) and then the second row etc, etc... Each singer sings the same length of song regardless, each row exposes for the same amount of time. The delay between rows/singers starting is how much 'rolling' shutter there will be which is what Graeme is referring to by "fast shutter". The rate at which each pixel row starts its exposure is much faster.

The Mysterium-X's delay between each row starting its exposure period is much shorter than Mysterium's so there are less rolling shutter artifacts.

This is in contrast to a global shutter which is like a choir all starting at the exact same time.

thank you very much Gavin.

 

[Doug Horton]

So why did RED choose a rolling shutter over global then?

 

[Elsie N]

So why did RED choose a rolling shutter over global then?

Possibly to keep the cost of the camera under 6 figures... but I'm just guessing.

 

[David Rasberry]

So why did RED choose a rolling shutter over global then?

It reduces bandwidth requirements when reading the sensor for one. Also the rolling shutter is much closer to emulating the cadence and look of film that everyone seems to love. Given the quality of images a Red produces I think it is pretty much a non issue except for a few situations like unsynchronized strobe lights.
If rolling shutter is really that serious an issue for you, you should buy a 3-chip video camera instead. That is why all these different options exist in the market.

 

[Shane Nassiri]

In my experience in shooting RED, it's so completely negligible that is is a non-factor. Quite the opposite when comparing that to 5D/7D, which is atrocious. We typically shoot EX1 unless we rent RED, and the jello in the EX1 is 10x worse than RED, and my 5D is 10x worse than the EX1.

There are a lot of RED detractors for whatever reason, and this is one of the things they point to. Case in point, I am working with a guy doing some graphics on a little film project that he shot on an HVX (or HPX, or something). It was the first time I met him, and thought I'd ask how he liked shooting with the HVX. He touted the cameras ability and took an unprompted swipe at RED for rolling shutter artifacts. It was pretty obvious to me that he had never actually used the camera and only read on a message board somewhere that RED and CMOS have rolling shutter issues.

All that to say: use it and I think you'll see it's a non-issue. And from the reports, the MX sensor is even less of an issue.

 

[Doug Horton]

It reduces bandwidth requirements when reading the sensor for one. Also the rolling shutter is much closer to emulating the cadence and look of film that everyone seems to love. Given the quality of images a Red produces I think it is pretty much a non issue except for a few situations like unsynchronized strobe lights.
If rolling shutter is really that serious an issue for you, you should buy a 3-chip video camera instead. That is why all these different options exist in the market.

not complaining, simply trying to understand the technology better, but thank you. I currently own an HVX but would love to own a Scarlet if the funds allow me to do so.

 

[george manzanilla]

not complaining, simply trying to understand the technology better, but thank you. I currently own an HVX but would love to own a Scarlet if the funds allow me to do so.

i'm on that same boat. hvx has been a great camera, although it obviously has its drawbacks, i think most people that come from that camp are pretty comfortable working around some of its issues inherent in ccd's. that camera has had quite a return on its investment.

that being said, cmos technology is new to people that come from the ccd based camera systems (ive been shooting panasonic cams for 9 years now) and some of the talked about issues SEEM like deal breakers, although from hearing the users speak about it, i am a bit more convinced. will have to try it out!

 

[Michael McVicar]

No. The best way to think about it is a song sung in a round. The first row of pixels starts singing "Frere Jacques..." and then the second row and the the third row and so on and so forth. Eventually they're all singing at the same time. And then the first row finishes the song (its exposure length) and then the second row etc, etc... Each singer sings the same length of song regardless, each row exposes for the same amount of time. The delay between rows/singers starting is how much 'rolling' shutter there will be which is what Graeme is referring to by "fast shutter". The rate at which each pixel row starts its exposure is much faster.

The Mysterium-X's delay between each row starting its exposure period is much shorter than Mysterium's so there are less rolling shutter artifacts.

This is in contrast to a global shutter which is like a choir all starting at the exact same time.

Just a thought, with fewer pixels and rows in the 2/3 sensor would we expect to see faster read/reset times? I know it is a different chip and we'll see what we see when we see it, but just wondering. Any thoughts?