noisy blue channel

D

DanHD

Well-known member
Has anyone noticed any additional noise in the blues when tungsten balanced on this camera?

Is the 170 inherently daylight balanced, in that the blue channel is amplified when shooting tungsten by adding 6db gain?

I've seen this problem arise first ACing on a Red shoot, where the DP would light everything 5600 or 5000k to "feed the blue" and avoid the noise, and wanted to know how pronounced it is on this cam

Should I bother balancing all my lights daylight for indoor shoots?
 
I read this in a thread on RedUser

"Due to the physical characteristics of silicon sensor manufacture and other issues beyond the scope of this discussion CMOS sensors like the Mysterium are more sensitive to particular parts of the spectrum. To compensate for this non-linear spectral response gain is added, most notably to the blue channel. It should be noted that the debayering algorithms that Graeme has developed are assumed to be using the full RAW data set to intelligently interpolate blue channel information even from pixel sites that do not have a blue mask. If that is incorrect then perhaps Graeme will see this post and clarify. In any case, the key point is the blue channel gets pumped up and that process can create noise'

Does this mean CCD's are more immune to this occurence?
 
The CMOS is a totally different character than a CCD. As the world makes a transition to CMOS you need to learn new stuff in order to to optimize its performance and not have it hit you over the head with surprises.

To directly answer the original posters question, No, I have never seen any additional noise in the Blue channel under tungsten light. This is BTW the light I most frequently use to look at this camera, its the light source we have in our lab.


Hope this helps,

Jan
 
When shooting under tungsten (warmer lighting), the camera has to increase its blue channel gain to achieve a neutral color balance.

This has been a topic and a known issue with video cameras for a long time. A quick google search will confirm this. To make things worse for these cameras, sensors are the least sensitive to the blue channel bandwidth. These two conditions add up to more noise under warm lighting.

Dan, you're right, it's really a DoP call. Depending on your set, it can be a pain to go CTB.
It's certainly, if not almost impossible to do this at all locations.

You can always balance to the warm side. A good solution is using warm cards (blue cards) to force the blue channel down.
 
Last edited:
Wasn´t there a thread somewhere to shoot with a magenta filter to get rid of noise?

If someone remembers, please let me know.

Thank´s Frank
 
Hey guys, I do not see this in a CCD imager, haven't for 30 years. Nonetheless the blue channel in HD is only 7% of the signal.

Best,

Jan
 
Thank you all, and Jan

From what I've seen so far on the HVX200 and hpx170, it's more of a nonissue. I've only actually visually noticed noisy blue shadows and such on the EX1 and, suprisingly, the Red camera.

I think I'll just balance for tungsten and if the results are truly worse, I'll notice it.
 
I have never seen the issue on a CCD, the make-up of the CCD is really quite different than that of a CMOS. The two cameras you mention are both CMOS. White balance with the HPX170 and it will be fine.

Best,

Jan
 
Perhaps it is something with the striping of those large single sensors, the Bayer filter syste or something of that sort. But in the thirty years I have been in this business, and granted CCDs have not been a part of our reality that long, more like 22 years, I have never seen this as an issue in a three ccd prism split camera. Neither have the two engineers I passed this discusion past.
 
That's great to know Jan. And it may well be that prism split three CCD cameras are immune, or not nearly as sensitive to this as single chip cameras. I don't know the physics well enough to speak on it.

I will say, that at least perceptually, I have been more happy with the results of my EX1 with daylight balanced light.

Perhaps if I get time next week, I'll try a test with my DVX and post results, and see if people can tell a difference. Might make for some interesting results!
 
I talked to Vincent Roth who is the techinical director of Fofokem and he says that CMOS Bayer patterns are far more susceptable to this occurence, and not to worry if I have a three chip prism split.

So i've had both Jan and himself tell me the same thing...so I think i'll just end up shooting as I always have with Tungsten balance


I will say though, the EX1 looks far noisier under tungsten balance. I did find that when I rented the camera for wedding work.
 
But the EX1 uses a CMOS chip with a bayer patter just as the Red One does do that may be the case

In which case I'm fairly dissapointed.
Do we have to trade between CMOS and less blue channel noise? Everyone over on the Redboards use Kinos and HMIS when they shoot, no tungsten

i prefer to use my tungsten lights, I have some pretty good Mole Rishardsons as well as Lowels
 
DanHD said:
But the EX1 uses a CMOS chip with a bayer patter just as the Red One does do that may be the case
Click to expand...

But the EX1 is not a single chip camera. It is a 3-chip unlike the RED.

DanHD said:
i prefer to use my tungsten lights, I have some pretty good Mole Rishardsons as well as Lowels
Click to expand...

So do I. So gel the lights or filter the camera. Problem solved. And quite cheaply I might add. 77mm 80A is what, ~100?
 
Problem is solved, yes, at the expense of losing two stops of light. Either CTB gels on the lights, or an 80A filter, is going to result in a net reduction of transmissible light to about 25%. Not saying it won't work, just pointing out that, as always, there's a tradeoff somewhere.
 
Barry_Green said:
Problem is solved, yes, at the expense of losing two stops of light. Either CTB gels on the lights, or an 80A filter, is going to result in a net reduction of transmissible light to about 25%. Not saying it won't work, just pointing out that, as always, there's a tradeoff somewhere.
Click to expand...

Nature of the business. No way around it. So get in bed and get comfy with it.
 
Haven't seen the extra blue noise at 3200K with the 170, blue channel is always noisiest with
any CCD.

Haven't found th EX1 to be noisy under tungsten at all. EX1 is clean but slow at 1080.
 
If I may ask,


Why are cameras made daylight balanced?

Is it because camera manufacturers assume that most shooting will be done in daylight?

with the amount of indoor shooting, and the abundance of tungsten, it seems as if cameras should be MORE sensitive to blue not less

and another question.... why cant cameras be made to be equally sensitive to all three colors?
 
DanHD said:
If I may ask,


1. Why are cameras made daylight balanced?

...with the amount of indoor shooting, and the abundance of tungsten, it seems as if cameras should be MORE sensitive to blue not less

2. and another question.... why cant cameras be made to be equally sensitive to all three colors?
Click to expand...


Let's take #2 first. Answer, they are. But since they are equally sensitive, and our tungsten light offers so little blue, the problem becomes obvious. Whe we learn how to make sensors more sensitive we don't just apply it to the blue, we apply it everywhere.

Which leads to #1. Filmmakers and parents shoot indoors. The world shoots outdoors. Sports, nature, pets, vacation. Let's face it. Consumers buy cameras to shoot outside. You need only look at the film industry for examples. Tungsten based camera film nearly died off becase EVERYONE bought daylight film. In fact, in ye olden days, it said right on the film box if you were going to shoot daylight film indoors, to use an 80A filter. No one did. Everyone had red pictures and didn't know why. Eventually, the labs learned to just fix it in printing. A simple color timing change.

Actually, you could very well do the same thing. Don't color balance anything and simply make the color timing change in post. That way, you'd not lose 2 stops of light with a filter or gelling your lights. I've done it for others when the color balance was wrong, or when they left auto-white-balance engaged and I had to re-time the footage every 3 seconds. Royal PITA.
 
DanHD said:
If I may ask,

Why are cameras made daylight balanced?

Is it because camera manufacturers assume that most shooting will be done in daylight?

and another question.... why cant cameras be made to be equally sensitive to all three colors?
Click to expand...

Every video camera I've ever owned, primarily broadcast Plumbicon and CCD cameras, were preset for 3200K. The only video camera I've heard of that is
designed for 5600K is the RED ONE.

Cameras cannot be made to be equally sensitive to all three colors because the
three colors represent different luminance values, with green making up the largest percentage of the luminance by far, then red, lastly blue.
 
You must log in or register to reply here.
Back
Top