Sorry I guess I fell down the same rabbit hole. I did test and indeed the exposure at 5600 with an 85 filter did take 2/3 stop to match no filter but I guess that's meaningless anyway - oh well never mind. The basic question still remains - what color temp is native to the FS7/F5 and probably all Son'y 's for that matter and should this affect exposure and noise?
FS7: Overexposing on the FS7 for cleaner shadows
- Thread starter RASL
- Start date
broughtonfilm
Well-known member
It would be good to know and I'm curious to find out myself but personally, I don't worry about it too much. I guess the difference would have to be really jarring for me to lose sleep over this. I'm not worried because I feel that in situations where you'd really push the camera's ISO rating to the limits of risking something are in the scenes where a slightly increased noise is somewhat acceptable (I know I will get eaten alive for this statement). When shooting on film it was a common practice to use 50 - 250 ISO D daylight film stock for daytime exteriors and 320 - 500 ISO T film stock for tungsten lit interiors. As a result the interior shots always had a slightly increased grain as compared with exteriors. Was it done by design? No, I agree that in most situation it was a matter of necessity (and also convenience, no need to filter the T stock with 85 and ND filters, etc) but sometimes it was even intentional, as a part of dynamic look of the film. And yes, noise is not exactly the same as a film grain. I guess I don't worry also because we get to see what we are doing on a monitor, in other words if you do look carefully at what you are doing you run all those test all the time anyway without even making a specific test but rather as a test of all the factors at the same time - hey, that's exactly where digital cinema has an edge over shooting film, let's use it to our advantage! I guess the only true significance of knowing which setting has an edge would be when changing the color temperature of lights is as easy as a turn of the knob on the LED panels or replacement of Kino bulbs. Then yes, of course use the setting which is the cleanest. But this is a studio situation where usually we don't even have to push the limits of the camera and can usually afford to rate the camera lower anyway.
Ben Scott
Veteran
Was pretty certain that all modern CMOS sensors perform 'best' when given enough blue channel information to do their job, which denotes a daylight balance being preferable. However, what is 'best' from a technical standpoint may not be what is best for what you want from your image.
Otherwise we wouldn't still be employing diffusion filters and adding grain to images.
Otherwise we wouldn't still be employing diffusion filters and adding grain to images.
Doug Jensen
Carbonite Member
I've never bothered to find an answer to this question because it makes no difference to me whatsoever: 1) You can't change it. And 2) I have no interest in using tungsten lighting whether that is the native color balance or not.
But with that said, I'd wager that the FS7, F5, and F55 are native 3200K. That's what the PDW-F800/700 and all the Betacams were, so it's logical to assume that other Sony cameras in the same family would be the same, unless of course, someone finds conclusive evidence from a Sony source that they are not.
morgan_moore
Major Contributor
If the native is 3200 this means the sensor is extremely sensitive to blue
55k would be adding -Ve gain to te blue channel (or boosting g and r)
Now.. Speculation
Would use of a physical yellow filter be of interest instead of adding that -Ve gain???
Well I have a feeling that it might (the same way as I think a physical blue filter is good on Red under tungsten)
The results might be cleaner although light levels would need upping
Also as I mentioned above it might mellow off some of the horrific blue led clips that we see on occasion
In general I have been wondering about camera filtration under stage and conference led hell.
55k would be adding -Ve gain to te blue channel (or boosting g and r)
Now.. Speculation
Would use of a physical yellow filter be of interest instead of adding that -Ve gain???
Well I have a feeling that it might (the same way as I think a physical blue filter is good on Red under tungsten)
The results might be cleaner although light levels would need upping
Also as I mentioned above it might mellow off some of the horrific blue led clips that we see on occasion
In general I have been wondering about camera filtration under stage and conference led hell.
broughtonfilm
Well-known member
This makes zero sense as the negative gain is not applied to color channels to color balance the camera because that would affect the resulting exposure and the manufacturer's goal is to keep it constant regardless of the white balance setting. It is always boosting of the gain of the channel whose color is lacking in the light source to achieve it, in example, boosting of a again of the blue channel under tungsten lighting for a camera that is 5500K natively and not decreasing of a gain of red and green channel to achieve the balance. That's the whole point.
On the other hand, yes, ENG cameras which were balanced to 3200K used an optical 85 filter to balance the camera to daylight. So maybe I should not say "zero sense".
morgan_moore
Major Contributor
And my point is that rather than boosting the weak channels would it create a more pleasant result to filter pjyoscally out the hot colour?
broughtonfilm
Well-known member
That's how it was done in cameras balanced to 3200K. But filtering with 85 filter is different than with 80A when going other direction in how much light lost each costs you (2/3 stop light loss vs 2 stops light loss respectively ), not to mention that if you filter from 3200K to 5500K with 85 you do that to shoot in daylight where light levels are higher so you can afford it. If you go other way when your camera is balanced to 5500K and you have to shoot under tungsten light, you'd lose 2 stops with an 80A filter and that's where light levels are usually much lower. Your 1000 ISO camera would be a 250 ISO one effectively all in a sudden. http://www.xdcam-user.com/2009/07/pdw-700-native-white-balance/
Thanks - I forgot that most cameras were originally balanced to 3200 and used a daylight filter for exteriors , so its likely that Sony kept that arrangement. That would imply perhaps that under daylight a slightly lower ISO would be preferable. Thankfully in very low light we are more often in tungsten.
broughtonfilm
Well-known member
I don't think that's the conclusion. I think since the camera is not using optical filters for color balance, it is opposite - it is likely balanced to 5500K. And what if it is neither?
Aharon Rothschild
Member
I would say underexposure is the cause of noise and underexposed tungsten in particular. If we use a meter we can get away from this and not really need to worry about the filtration either as there will be enough going on across the rgb channels etc. you are going to see gain across blue and red for 5600 as well as 3200. You just see more in blue at 3200 and when underexposed and possibly pushed even further in CC that right there can make things very noisy. Shooting RAW is a big issue here as underexposed tungsten RAW sees the same gain to blue, it's just not locked in, in camera and many shoot RAW in some kind of log mode not seeing how underexposed they are and then crop in on those high res images in post and really don't get that even though it's RAW there is still a native color temp.
But this is all relative to exposure, and this was the last of my suggestion to the OP in order of importance relative to noise. I think matching exposure on set to exposure in post and on screen though all the shots in a scene is probably the best way to avoid noise. Ie. push the image the least amount possible by getting exposure as close as possible and not seeing much of a spread through all the other shots in the scene leading to a minimum of pushing the image up or down.
This is best achieved by running a camera test to find the best noise vs range ratio for your needs and then rating the camera and exposing to that.
Filtration for 3200 will help as well but lets get a meter first and work from there.
But this is all relative to exposure, and this was the last of my suggestion to the OP in order of importance relative to noise. I think matching exposure on set to exposure in post and on screen though all the shots in a scene is probably the best way to avoid noise. Ie. push the image the least amount possible by getting exposure as close as possible and not seeing much of a spread through all the other shots in the scene leading to a minimum of pushing the image up or down.
This is best achieved by running a camera test to find the best noise vs range ratio for your needs and then rating the camera and exposing to that.
Filtration for 3200 will help as well but lets get a meter first and work from there.
morgan_moore
Major Contributor
Yep and it might be a really clean one.
Consider we (or I) often shoot with ND engaged 250 might be ok
S
Mitch Gross
Veteran
All camera sensors are blue-starved. It is the nature of silicon. They are always daylight-balanced at the sensor level, but then color processing later on can shift this to tungste-balance. Please note that not all cameras were or are natively set to tungsten base. Most professional Panasonic cameras allowed one to set the base balance at daylight or tungsten (and oh wasn't it fun when someone forgot and then rotated in the 85A filter wheel -- thanks B&W viewfinders). Noise is always seen first in the blue channel, basically because blue is usually getting pumped up in the processing. So the worst thing you can do when shooting in daylight is have the camera set to tungsten balance (pumping up the blues) and then throw an 85A filter on there (pushing them back down again). I can't recall on the FS7 as I don't have it in front of me right now, but most Sony cameras will let you select a base preset white balance of daylight or tungsten. In the Cine EI mode you can also set the base halfway in between. It's almost always better to set the colors electronically. This will adjust the gains applied to the individual color channels first, and you won't be needlessly adding gain to a color channel only to push it back down again with a filter. If you want to experiment with adding filtration to improve noise issues as well as color channel clipping issues, I would suggest starting with a blue filter such as an 80C. That plays to the native proclivity of the sensor.
Sam, I don't see how you can say that if a camera is high ISO then it is therefore blue-sensitive. Once you compensate for overall exposure, the color balance of the sensor is not changed. CMOS or CCD, pretty much every sensor I've ever tested is blue-starved. It's the nature of the beast.
It's really easy to shoot some quick tests to figure this out and experiment with the results. Shoot whatever subjects you like and bring the shots into Photoshop. There you can look at the individual red, green & blue color channels as monochrome images. You'll be amazed at how noisy the blue channel is. Not a dig on Sony -- this is true for every camera.
Sam, I don't see how you can say that if a camera is high ISO then it is therefore blue-sensitive. Once you compensate for overall exposure, the color balance of the sensor is not changed. CMOS or CCD, pretty much every sensor I've ever tested is blue-starved. It's the nature of the beast.
It's really easy to shoot some quick tests to figure this out and experiment with the results. Shoot whatever subjects you like and bring the shots into Photoshop. There you can look at the individual red, green & blue color channels as monochrome images. You'll be amazed at how noisy the blue channel is. Not a dig on Sony -- this is true for every camera.
morgan_moore
Major Contributor
Mitch early in the thread I said I was only theorising. I may have it completely ar_se about face.
You say sensors are starved of Blue.
Would it not make sense, therefore, not to apply an ND filter, which AFAIK dim all three colours, but to apply a blue filter (a filter that removes green and red) when the opportunity arose?
S
You say sensors are starved of Blue.
Would it not make sense, therefore, not to apply an ND filter, which AFAIK dim all three colours, but to apply a blue filter (a filter that removes green and red) when the opportunity arose?
S
Mitch Gross
Veteran
A blue filter doesn't particularly replace an ND filter in this case. It shifts the color balance as opposed to limiting the light level. Now, a full index compensating blue filter would eat about 1.5 stops of light, so it does have the secondary effect of darkening the overall image. But I would question the use of that filter, as the blue channel is not THAT desperate for blue that one needs to so dramatically choke off the other color frequencies. That's why I suggested experimenting with an 80C filter, which cuts about 1/2 a stop of light.
BTW, I have not tested this on an FS7 and I don't know the degree to which this will improve or particularly alter the image. Test it out yourselves. I suggest buying a bit of gel and taping it in front of the lens to do some experimenting. A sheet of gel is only a few bucks and swatch books are free. That's plenty good enough to do some testing.
BTW, I have not tested this on an FS7 and I don't know the degree to which this will improve or particularly alter the image. Test it out yourselves. I suggest buying a bit of gel and taping it in front of the lens to do some experimenting. A sheet of gel is only a few bucks and swatch books are free. That's plenty good enough to do some testing.
Mitch , Thanks for much for your always valuable insight.
That's why I raised this question originally and verifies suspicions I've had for a long time , ever since I learned about how people were exposing RED - which is to expose Tungsten more than daylight to get less blue noise and that some people shoot with an 80C under tungsten. The curious thing is that I've never heard the same advice applied to Sony's, and I guess I would still like to hear some confirmation from Sony on that.
One reason this is easier with a RED however is that the camera allows you to color control your look or what's basically a LUT so you can see what your shooting without a blue mask and it will go to post with that look already applied as metadata. Looking at the wratten conversion number for an 80C though (-81) it looks like you could shoot as FS7/ F5 under 3200 tungsten with an 80C in Cine EI and use the "Mid' Color balance at 4200K. Should balance out. This would preserve your highlights somewhat as it would save the Red channel from overexposing a bit. But aside from that 1/2 stop saving the red channel , is it really any different from just overexposing - and thus worth the hassle of adding an extra filter? I'm not convinced but its an interesting idea.
It especially raises the question of whether the noise issue that makes people rate their cameras at 1000 and 8000 ISO is as necessary under daylight as tungsten.
Another question that comes to my mind however is if these sensors have always been blue sensitive, why in the world was it standard practice to make Betacam cameras with built in 85 filters for daylight? It would have just added noise. Was this simply because they were mimicking film? Seems odd for Sony engineering to go that route.
That's why I raised this question originally and verifies suspicions I've had for a long time , ever since I learned about how people were exposing RED - which is to expose Tungsten more than daylight to get less blue noise and that some people shoot with an 80C under tungsten. The curious thing is that I've never heard the same advice applied to Sony's, and I guess I would still like to hear some confirmation from Sony on that.
One reason this is easier with a RED however is that the camera allows you to color control your look or what's basically a LUT so you can see what your shooting without a blue mask and it will go to post with that look already applied as metadata. Looking at the wratten conversion number for an 80C though (-81) it looks like you could shoot as FS7/ F5 under 3200 tungsten with an 80C in Cine EI and use the "Mid' Color balance at 4200K. Should balance out. This would preserve your highlights somewhat as it would save the Red channel from overexposing a bit. But aside from that 1/2 stop saving the red channel , is it really any different from just overexposing - and thus worth the hassle of adding an extra filter? I'm not convinced but its an interesting idea.
It especially raises the question of whether the noise issue that makes people rate their cameras at 1000 and 8000 ISO is as necessary under daylight as tungsten.
Another question that comes to my mind however is if these sensors have always been blue sensitive, why in the world was it standard practice to make Betacam cameras with built in 85 filters for daylight? It would have just added noise. Was this simply because they were mimicking film? Seems odd for Sony engineering to go that route.