Native chip colour balance

[JD Holloway]

I'm hoping Epic's sensor will be Tungsten balanced to reduce the number of "noise" threads we have to pour through!

 

[Charles Angus]

That is the equivalent of putting an 80A filter in front of the camera, with the commensurate 2 stops light loss.

Would you rather have a 320ASA 5000K camera, or a 75ASA 3200K camera?

This solution has been proposed and rejected as not feasible/desirable/practical a thousand times.

 

[Michael Lindsay]

I'm hoping Epic's sensor will be Tungsten balanced to reduce the number of "noise" threads we have to pour through!

I argued on these forums about a year ago for a Tungsten sensor but I am no longer convinced.

1st of all I'd rather not lose a stop or 2 on the other 2 channels (with no option).

And secondly I'm finding it 'easy'ish' to deal (post B16) with the 5000k sensor under tungsten light using a either a 80D/80C or contrast control.

In general I am as often nearly clipping the blue channel as I am underexposing it. So we're just getting on with it and learning as we go along.

I'd rather have a camera that (via a 2 position OLPF/ND combo) could be rated cleanly at either 50asa or 500asa. Throw in 2 more stops and I'd be fall over with dizzyness.

Sorry to disagree

Michael

PS I don't really expect the OLPF/ND sliding filter to work. the tolerances within the world of back focus errors would make it too tricky

 

[Jeff Kilgroe]

Daylight balance is the way to go, IMO. :tongue:

Not everyone shoots under tungsten light. Many of us do most of our shooting in natural / available light. Going for a 5000K or wider color balance is better as it increases the sensitivity across the entire spectrum.

Many of the noise issue threads we see are more of a product of reduced blue-channel sensitivity. Which is a common property of most digital sensors and (AFAIK) a problem with all current CMOS designs. As RED evolves their sensor design, increases DR and hopefully blue channel sensitivity in relation to red and green, many of these concerns should be put to rest.

 

[JD Holloway]

Many of the noise issue threads we see are more of a product of reduced blue-channel sensitivity. Which is a common property of most digital sensors and (AFAIK) a problem with all current CMOS designs. As RED evolves their sensor design, increases DR and hopefully blue channel sensitivity in relation to red and green, many of these concerns should be put to rest.

This is what I'm getting at. While Red's sensor development program is in its infancy, I'd like to suggest that this be considered. CMOS cine-design needs to be re-envisioned to the realities of on set-lighting and lighting distribution. Until high output low cost LEDs/Kinos become the norm for set lighting (not holding my breath) tungsten will rule the sound-stages. God bless HMI's but they're heavy, complicated and unreliable at times for the thousands of reasons that any Gaffer will gladly point out.

Available-light cinematography typically has an abundance of light; less so when we go inside. Lets all be ready for interiors.

 

[Graeme Nattress]

I've said it before and I'll say it again:

silicon sensors are insensitive to blue light. Unless you know how to change physics, we're stuck with that.

Graeme

 

[Evin Grant]

How bout using the new "Black" silicon?
100-500X more light sensitive!
http://gizmodo.com/5062412/black-si...ange-digital-photography-night-vision-forever

 

[JD Holloway]

I've said it before and I'll say it again:

silicon sensors are insensitive to blue light. Unless you know how to change physics, we're stuck with that.

Graeme

understood...

gotta change the game...

 

[Graeme Nattress]

Yeah, and what does that do to your IR issues. And "silicon absorbing light" - what on earth does that mean....

 

[Pawel Achtel]

Available-light cinematography typically has an abundance of light; less so when we go inside. Lets all be ready for interiors.

How about 9000k balance for underwater use in available light? You see, you have your application, others have other applications. I personally think 5000k is a nice "middle" spot and I'd much rater have a sensor balanced for 5000k than 3200k.

Just put a piece of blue gel over your tungsten light and solve your individual problem with a workable $0.50 solution. Either that or use LED or fluoro lighting.

And, as Graeme mentioned, unless you can change physics, all sensors, not just Red, are native daylight or more sensitive to blue than red spectrum. You can compromise this by placing a filter and reduce the overall sensitivity by a stop or two, if this is what you want. But, I find it somewhat funny people requesting a feature requiering a change in physics...

 

[Deanan]

At some point, tungsten lights will be outlawed like CRT monitors because they waste 90% of the energy put into them to make the 10% light output.
Then, only outlaws will have tungsten lights.

btw, this post is a joke

 

[Stephen Williams]

And, as Graeme mentioned, unless you can change physics, all sensors, not just Red, are native daylight or more sensitive to blue than red spectrum. .

Hi,

The one that does the impossible wins the game!

Stephen

 

[Stephen Williams]

At some point, tungsten lights will be outlawed like CRT monitors because they waste 90% of the energy put into them to make the 10% light output.
Then, only outlaws will have tungsten lights.

btw, this post is a joke

Hi,

In general tungsten lights will be outlawed, however there will be exceptions.

Stephen

 

[JD Holloway]

How about 9000k balance for underwater use in available light? You see, you have your application, others have other applications. I personally think 5000k is a nice "middle" spot and I'd much rater have a sensor balanced for 5000k than 3200k.

Just put a piece of blue gel over your tungsten light and solve your individual problem with a workable $0.50 solution. Either that or use LED or fluoro lighting.

And, as Graeme mentioned, unless you can change physics, all sensors, not just Red, are native daylight or more sensitive to blue than red spectrum. You can compromise this by placing a filter and reduce the overall sensitivity by a stop or two, if this is what you want. But, I find it somewhat funny people requesting a feature requiring a change in physics...

Certainly underwater filming is always hurting for light. As I'm sure you know it can be hard work getting accurate colour renditions. No doubt my interests conflict with yours in this department as I do little to no underwater work.

Putting CTB on all your tungsten lights is not always a solution. There is simply not always a budget for enough "big guns" (10Ks/20Ks) to light large spaces and/or shoot off-speed and/or with <180 shutter. Lens options also rear their ugly heads here. Then there is the budgetary constraints of power distribution, generators and cabling. Effectively throwing away half your light or more with gel (more inefficiency), generating heat, is simply not a solution. A hot stage is expensive. Actors are tired/pissed, producers have to pay for air con and more equipment blah blah blah.

LED or Flos? You're kidding right? These lights, while very useful tools, do not (yet) effectively replace tungsten in output or cost/lumen in large interior stage set ups. Furthermore shooting in rooms that have any "practical" lamps/pot lights/candles/firelight etc bring into the forefront issues of colour balance. Lighting to 3200K and dropping an 80a filter on the lens in an already low light situation is not an answer (though if the camera can handle iso3200 noiselessly maybe it might be!).

I understand (to the extent my brain works this morning) the issues with silicon CMOS technology and its insensitivities. Certainly SiOnyx has a breakthrough. I like the linear sensitivity curve in the 400-600 nm range. Less Blue channel gain would be a happy day for all reguardless of our shooting environs. I would think their unit would need serious IR filtering but this could be the stuff of the future.

Happy (Canadian) Thanksgiving!

 

[J. Bernard Vallon]

the solution is to reduce sensor noise in general. Modern Still cameras use CMOS sensors @ high ISOs under tungston lights no problem. Sure you have to boost the blue channel, but there is such low noise there anyway it still looks great.

 

[Dj Joofa]

unless you can change physics, all sensors, not just Red, ...

One may not have to "change physics." Just look at how human vision does illuminant change compensation and you might get a clue on how to accomplish some of this goal in digital cameras, which apparently many people are not using.

 

[JD Holloway]

Interesting Joofa?
Would you use some level of colour AI?

 

[Pawel Achtel]

the solution is to reduce sensor noise in general. Modern Still cameras use CMOS sensors @ high ISOs under tungston lights no problem. Sure you have to boost the blue channel, but there is such low noise there anyway it still looks great.

Agreed, I think this is solution that is going to attract to all, rather than changing the native 5000k colour balance.

 

[Dj Joofa]

Interesting Joofa?
Would you use some level of colour AI?

There are some differences in how the white balance is done in digital cameras and how the human vision adjusts its sensitivity for chromatic adaptation. The corresponding "change in sensitivity" in digital cameras is done by adjusting the gains between color channels. However, that is a fixed procedure that most manufacturers follow. I have personally gone through at least more than 20 books purely on "Colorimetry", and countless research publications, however, all most all of them miss on this point or that point on how to accomplish things in practise. White balance, and especially, auto white balance, is not an easy task, and there are several issues relating to color fidelity.

For a camera such as Red, some of them don't apply, for e.g., Red doesn't do auto white balance, and the resulting footage is color corrected by digital artists manually for a look that is pleasing, not necessarily, that is correct.

I have talked with many people on these issues, and I found that a good understanding of issues of colorimetry as applied to the development of a modern digital camera is not very clear among practitioners and theorists alike. There is a need for explaining such complex issues.