Why different RED One cameras' Color Tempetures are totally different?

[Yang Liu]

We run a rental house in Beijing , in this half year, a few crew were shot on A/B RED one cameras on set, everything is fine excpet the CT. Always, A/B camera look different. At very beginning I thought it was the problem of Monitor(we use TVLOGIC 17''), but I realized the RED LCDs look different as well.

I'm very very sure all the SETTINGs are the same, even I shot on the same lens,at the same location, under the same lighting... When DP asked me can you calibrite the cameras, my only answer is we can do it in post.

But can anybody explain why they're different?

 

[Mitch Gross]

It may be a Digital Cinema camera, but camera sensors are actually analog devices. And they have chemical dyes in their Bayer masks, both of which are manufactured in production batches. So there are variations that can occur in these cameras just as they can in every other camera. In designs that are not RAW based, there color matrices applied to balance this out. What you are likely seeing is the true nature of the camera sensor.

 

[Frank Cueto]

It may be a Digital Cinema camera, but camera sensors are actually analog devices. And they have chemical dyes in their Bayer masks, both of which are manufactured in production batches. So there are variations that can occur in these cameras just as they can in every other camera. In designs that are not RAW based, there color matrices applied to balance this out. What you are likely seeing is the true nature of the camera sensor.

Where is the LIKE button for this post???

J/K!!!

 

[Shawn Booth]

Did you white balance both cameras with a grey card set in your lighting?

 

[Jordan Livingston]

Try performing a black balance calibration on both cameras in the same location / lighting conditions?

 

[Yang Liu]

It may be a Digital Cinema camera, but camera sensors are actually analog devices. And they have chemical dyes in their Bayer masks, both of which are manufactured in production batches. So there are variations that can occur in these cameras just as they can in every other camera. In designs that are not RAW based, there color matrices applied to balance this out. What you are likely seeing is the true nature of the camera sensor.

Thanks for your explaination, Mitch. I know Abel also have a few RED Ones, how do you guys deal with this problem if DP and director ask for using the same look cameras? I can't explain why they look different to the Director, they only want the results.

 

[Yang Liu]

Did you white balance both cameras with a grey card set in your lighting?

We keep the white balance on 5600/3200 for both cameras.

 

[Yang Liu]

Try performing a black balance calibration on both cameras in the same location / lighting conditions?

If u mean black shading, yes we tried many times. But it seems like the sensors are different.

 

[Mitch Gross]

I tell them exactly what I told you.

We have a few sequentially numbered cameras that are very close matches if it really becomes an issue.

 

[Jeff Kilgroe]

Mitch speaks the truth. All sensors are different, this is not just an issue with the RED, but with any digital sensor or camera system. I see RAW variations all the time between sensors in medium format digital backs, even when they are supposed to be "the same". Every sensor must be profiled and each sensor has its own character, almost like a digital finger print. You can create individual LUTs for each camera to balance these differences out. But when looking at the RAW, you will see the subtle nuances between each sensor when shot under the same conditions. In most cases, it should not be so different that it's truly noticeable unless both cameras are shot with the same glass, from the same perspective at the same subject, same lighting, etc.. If you are seeing a huge difference, like off by 500°K, then you should contact RED.

I brought up glass because often there can be differences in color-casts from one lens to another, even of the same make and model. Increasingly more so on older lenses. Other objects in the optical path definitely play a role too, like filters.

 

[Mitch Gross]

(I can make fun because this stuff doesn't exist anymore.)

One of the fun things about the dear departed Dalsa Origin & Evolution s that each sensor had a different look file that traveled with it. They went through a lot of improvements during their foray into the Digital Cinema world, and ith each sensor upgrade came a new look file for anything that was shot with that sensor. Not such a radical thing -- many cameras do this. But usually it is embedded within the header as color offsets, sort of a shading control. On the Dalsa it travelled as a separate reference file and one had to be sure to keep them locked together in post otherwise a small nightmare would ensue as you tried to match everything back ito place.

Those sensors were gorgeous, but this was just one of the many little headaches that kept this system from seeing widespread use.

 

[Yang Liu]

I tell them exactly what I told you.

We have a few sequentially numbered cameras that are very close matches if it really becomes an issue.

Thanks Mitch, next time I will tell them what you told me.:001_smile:

 

[Yang Liu]

Mitch speaks the truth. All sensors are different, this is not just an issue with the RED, but with any digital sensor or camera system. I see RAW variations all the time between sensors in medium format digital backs, even when they are supposed to be "the same". Every sensor must be profiled and each sensor has its own character, almost like a digital finger print. You can create individual LUTs for each camera to balance these differences out. But when looking at the RAW, you will see the subtle nuances between each sensor when shot under the same conditions. In most cases, it should not be so different that it's truly noticeable unless both cameras are shot with the same glass, from the same perspective at the same subject, same lighting, etc.. If you are seeing a huge difference, like off by 500°K, then you should contact RED.

I brought up glass because often there can be differences in color-casts from one lens to another, even of the same make and model. Increasingly more so on older lenses. Other objects in the optical path definitely play a role too, like filters.

Thanks Jeff. And I think you are right about the glass issue. Thanks for your response.

 

[Peter Hodgins]

Try performing a black balance calibration on both cameras in the same location / lighting conditions?

same location / lighting ? for a black balance ? Should not be too hard.:hurray:

 

[John Marchant]

It makes perfect sense for there to be variations, but I remember being surprised a while ago at having a seriously hard time getting two XDCAM-HD cameras to balance nicely for a 3D shoot. Shooting closely matched images for 3d meant I was being much more critical than normal but the inherent sensor differences were there.

 

[Rick Burnett]

When I design CMOS chips, we simulate 5% to 15% variation in circuit performance possible across and entire design. This is on top of 100% to 200% equal variation across the entire chip. There is actually a significant amount of variation that can occur in the processing of chips. And this is not including the filter layer which also probably has variance I am sure. When working at these small sizes, it can be challenging. This is why bigger sensors cost A LOT more to manufacturer because you end up making more of them and picking out the ones with the least variation. All chip design tends to work like this on high speed or analog designs.

 

[Lonny Danler]

Please correct me if you all have a different opinion, but though accuracy is always the goal, RAW allows for some shift and as everyone is saying is inherent in the system. Even going out to monitor (what you see on set) doesn't match what you get in post off the r3ds. Michael Cioni has a demo of this at digitalcinemalessons.com and made sense to me.

 

[Ketch Rossi]

It may be a Digital Cinema camera, but camera sensors are actually analog devices. And they have chemical dyes in their Bayer masks, both of which are manufactured in production batches. So there are variations that can occur in these cameras just as they can in every other camera. In designs that are not RAW based, there color matrices applied to balance this out. What you are likely seeing is the true nature of the camera sensor.

This is why I keep stressing the fact that when thinking on spending some serious money in a S3D rig kit complete with Cameras, it is a MUST to talk to the Manufacturer and work together in getting not only two cameras with Sequential Serial Numbers, but more so make sure that the sensors in them are from the same BATCH, this will further improve a good image match coming from this cameras when shooting in S3D and leave much less work in post.

 

[Yang Liu]

It makes perfect sense for there to be variations, but I remember being surprised a while ago at having a seriously hard time getting two XDCAM-HD cameras to balance nicely for a 3D shoot. Shooting closely matched images for 3d meant I was being much more critical than normal but the inherent sensor differences were there.

It's stressful to balance two cameras, especially there was an asshole 1st AD and 'super stars' on set, the entire crew was waiting for you, it's really fun.

 

[Yang Liu]

When I design CMOS chips, we simulate 5% to 15% variation in circuit performance possible across and entire design. This is on top of 100% to 200% equal variation across the entire chip. There is actually a significant amount of variation that can occur in the processing of chips. And this is not including the filter layer which also probably has variance I am sure. When working at these small sizes, it can be challenging. This is why bigger sensors cost A LOT more to manufacturer because you end up making more of them and picking out the ones with the least variation. All chip design tends to work like this on high speed or analog designs.

Now I understand why it's expensive to replace a CMOS at RED.