Why Bayer

[Dj Joofa]

Not meaning to hijack this thread, but the CMOS on my Sony A1 does this on an annoyingly regular basis when white balanced for indoors;

Yes, there is contamination of signal into other photosites, that is why in matrix-based white balancing you will see negative coefficients to correct for that.

Additionally, don't know if my following comment applies to A1, but since many times linear R,G and B is used for internal processing, such as white balancing, then at least in theory, the color fidelity near saturated (primary) colors suffers.

 

[Matthew Greene]

FIT CCDs do not suffer from vertical smear like IT CCDs do. 700A, F900, 750, 790 and f23 all use FIT CCDs. Some cheaper HDCAMs used IT sensors.

Well... let me add that an FIT CCD has lower vertical smear levels but a very bright light can still cause smear. FT CCDs (with mechanical shutter) is the only CCD design that has absolutely no smear, without the shutter it's all smear.

 

[Pawel Achtel]

Well... let me add that an FIT CCD has lower vertical smear levels but a very bright light can still cause smear. FT CCDs (with mechanical shutter) is the only CCD design that has absolutely no smear, without the shutter it's all smear.

Correct, but practically it is not a problem with FIT CCDs unless you point into the rising or setting sun. :matrix: Even then, it is subtle and quite acceptable in most circumstances. Therefore I would not list it as a disadvantage.

 

[Craig W. Bickerstaff]

If saving private ryan has taught me anything it's that vertical smear can be a good thing.

 

[GlennChan]

Additionally, don't know if my following comment applies to A1, but since many times non linear R,G and B is used for internal processing, such as white balancing, then at least in theory, the color fidelity near saturated (primary) colors suffers.

It's when you apply a color matrix that color fidelity suffers- saturated colors will be too bright if increasing saturation via the matrix and vice versa. (This is only if you apply the color matrix on gamma-corrected/not-linear-light values.)
This effect is pretty subtle.

For white balance, apply the signal processing on gamma-corrected values doesn't matter if you are using the Rec. 601 transfer function. This is because the power function / exponent is commutative or something like that (I forget the exact math term).
if you have f(x) = constant * (x ^ 2.2)
you can go... f(x) = (constant ^ 1/2.2)^2.2 * (x^2.2)
= (constant^1/2.2 * x) ^ 2.2
you can hop that constant in/out, and it's adjusting that constant that lets you white balance.

If you use the Rec. 709 transfer function then that doesn't work, but it's close enough to 601 that you can probably cheat it. How well that works doesn't depend on how saturated the colors are.

-------
So, uh, back to Red. They don't do the color matrix cheating or the white balance cheating. So you don't get the subtle color inaccuracy where saturated colors are too bright/dark.

 

[Matthew Greene]

If saving private ryan has taught me anything it's that vertical smear can be a good thing.

Huh? SPR was shot on film, no smear. This pic shows smear which would otherwise just be the light.

 

[David Mullen ASC]

SPR had vertical highlight smearing due to the deliberate use of a mistimed shutter in the film camera, causing hot spots to streak upwards vertically in the frame. Not in every shot though, only was done selectively.

 

[Craig W. Bickerstaff]

SPR had vertical highlight smearing due to the deliberate use of a mistimed shutter in the film camera, causing hot spots to streak upwards vertically in the frame. Not in every shot though, only was done selectively.

I had wondered how they pulled that off.
Though I probably should have added "In the right circumstances" at the end of my statement.

 

[Craig W. Bickerstaff]

Going back to the Bayer discussion apparently Panavision went out and had a discussion about Digital cinema cameras turns out they aren't too big on the whole 4k bayer thing.

An important question to ask is how many photo sites does the imager (or sensor) have that contribute to the output? A so-called “4K camera” that has an imager with only 8 or 10 million photo sites but outputs 25 or 32 megapixels is interpolating data without adding any image performance. Moreover, it creates a four-fold increase in post-production digital data to be stored and processed that is both unnecessary and expensive.
Marketing pixels come at a price!

By contrast, Genesis has a 12.4 million photo site sensor, produces content superior to what is commonly called a “4K” camera, but through “super-sampling” our final output which is only 6 megapixels (or an 8 megabyte DPX frame). In other words: Same or better picture quality with far less overhead.

http://www.panavision.com/in_frame_detail.php?spotid=20&PHPSESSID=oljifidvtmibe2trdn5arhtfj0

 

[Graeme Nattress]

That's funny - it's like saying that you don't get a higher resolution, nicer image from a 12mp DLSR than you do from a Genesis video camera

And it takes me all the way back to the beginning of this thread - what do you do with your photosites?

Graeme

 

[Gunleik Groven]

Going back to the Bayer discussion apparently Panavision went out and had a discussion about Digital cinema cameras turns out they aren't too big on the whole 4k bayer thing.

An important question to ask is how many photo sites does the imager (or sensor) have that contribute to the output? A so-called “4K camera” that has an imager with only 8 or 10 million photo sites but outputs 25 or 32 megapixels is interpolating data without adding any image performance. Moreover, it creates a four-fold increase in post-production digital data to be stored and processed that is both unnecessary and expensive.
Marketing pixels come at a price!

By contrast, Genesis has a 12.4 million photo site sensor, produces content superior to what is commonly called a “4K” camera, but through “super-sampling” our final output which is only 6 megapixels (or an 8 megabyte DPX frame). In other words: Same or better picture quality with far less overhead.

http://www.panavision.com/in_frame_detail.php?spotid=20&PHPSESSID=oljifidvtmibe2trdn5arhtfj0

Oh crap, I've been fooled! Do you think if I sell my RED for 2,5k$, (it's more or less worthless now, after reading this....) I could get a panavision instead?

 

[Graeme Nattress]

I'd certainly encourage you to go out and by a Panavision video camera instead. Ok, you can't. Sorry.

Graeme

 

[Dj Joofa]

Additionally, don't know if my following comment applies to A1, but since many times non linear R,G and B is used for internal processing, such as white balancing, then at least in theory, the color fidelity near saturated (primary) colors suffers.

There was a typo in my original message, and I meant linear instead of non-linear. I have corrected my original message to reflect that.

It's when you apply a color matrix that color fidelity suffers- saturated colors will be too bright if increasing saturation via the matrix and vice versa. (This is only if you apply the color matrix on gamma-corrected/not-linear-light values.)
This effect is pretty subtle.

For white balance, apply the signal processing on gamma-corrected values doesn't matter if you are using the Rec. 601 transfer function.

Gamma corrected signals have a better response near saturated (primary) colors as far as color fidelity is concerned. On the other hand gamma applied to linear R, G, B, instead of non-linear R, G, and B, has better reproduction of luminance detail as you move out of the white point on a CIE xy diagram.

 

[Matthew Greene]

SPR had vertical highlight smearing due to the deliberate use of a mistimed shutter in the film camera, causing hot spots to streak upwards vertically in the frame. Not in every shot though, only was done selectively.

Oh, I remember that shutter effect now. I used a camera at film school that had no shutter (and not on purpose).

There was an old Ikegami video camera (perhaps a CCD model in the ITC series) that had horizontal smear instead. I find smear on video cameras extremely unpleasant while the horizontal streaks on anamorphic lenses don't really bug me. My best guess is that the physical (optical) defect is more pleasing (at least to me) than the electronic overloading of a colum of pixels.

 

[Jorge Díaz-Amador]

Mysterium Sensor Pixel Pitch, Active Area and Focus

Mysterium Sensor Pixel Pitch, Active Area and Focus

Graeme,

I just wrapped up my first week shooting on a feature with RED One 274 (rented). One more week to go. I'm working as the 1st AC for DP Philip Friesenbichler who expecting his RED One soon. I'll also be doing some operating on our two-camera days.

Prior to the shoot I had set up Philip's lenses (Zeiss T2.1 Standard Primes) on my collimator and projector so I knew they were calibrated. During the camera prep I set up the adjustable focus on the PL mount using a Siemens star target.

On the first day I knew something was off between my measured focus and eye focus. We were in a hot set (no ventilation - the things I get myself into) and the camera was running warm (but no overheating). I suspected expansion of the aluminum components of the lens mount, or something similar (the prep was in a cool sound stage). Once we had a break in the action, I wanted to correct the problem.

I had brought along my old Richter autocollimator with a LED Maglite as a light source. I tired using it as a target and focusing via the monitor but the image was too small. So I looked through it and to my surprise, I was able to see the pixel grid and focus off it.

So I set up the PL mount that way, and then checked the focus of each lens with a Siemens star set at 6 feet. Now all the lenses were dead on.

I flipped the Century focus target around to the resolution side and shot a quick test (this chart is meant to be shot at 50x the focal length). Using a Zeiss Planar 50mm at f2 / T2.1 I got 56 lp/mm equivalent observed resolution at the image plane. The 80 lp/mm target was not resolved.

That's very good for "delivered" resolution. About 2.7K horizontal (and the target was not perfectly flat, lens was wide open, etc.) But I'm trying to equate this to film resolution, and to the maximum resolution of the camera for purposes of evaluating focus, lens optical quality, etc.

Here's where I get confused: The Mysterium sensor has 4520 horizontal "active" pixels and 4900 horiz. "full" pixels. Are the pixels outside the 4520H area contributing to the "outside the framelines" image (not recorded but you can see the boom mic coming in), or are they "dark"?

In other words: Is the 4096 pixel wide 4K image created by a 4096 pixel wide "window" in the sensor area, or is it created by downsampling 4520 -> 4096? And is it the active or the full pixel area that is 24.4mm wide?

Any help you can provide will be appreciated. The reason I get into this minutiae is that I need to have standards so that I can assure the people I am working with, or my clients, that they are getting the best possible quality they can get. It's also important to know the sampled/captured image area dimensions in mm for lens coverage issues.

BTW, I'm sure you've heard this many times, but it bears repeating: At the end of each day the Director DP and I gather around the computer to view some clips of what we created that day. I am SHOCKED at the quality of the images we are getting.

 

[chuck colburn]

Jorge Diaz-Amador,

Finally someone (you) says that they can see the bounce back from an autocollimator off the RED sensor. And that infact one can field collimate with the portable Richter. This is info I've been asking for awhile . I've heard different things from different people including that it's NOT possible to see the reflected image off the sensor plane. I know that you know what you are up to as a lens tech. Were you able to rack focus enough on a lens too see the cover window on the chip? I'm wondering what that looked like. When you were doing your prep did you pop an optical flat into the PL mount to check parallelism of the lens seat to the film plane...errr excuse me sensor surface? This is great that you are posting what you have found.

cheers,

 

[Davide B.]

Hi Graeme,

What do you think of Leica's decision to not use an OLPF on their M8 camera and use software anti-aliasing instead? Could this be done on a digital cinema camera as well? Or do you think it's a bad idea for any type of camera?

davide

 

[Jorge Díaz-Amador]

Calibrating the PL Mount using Autocollimators

Calibrating the PL Mount using Autocollimators

Chuck,

Thanks! Actually I didn't think it would be possible to use an autocollimator or see the pixel grid. Steve from Du-All camera had posted that he used a Richter to set up the PL mount on the RED and I had trouble believing it. And here's the really interesting part: It looks different when the camera is on! When I tried to check it with the camera off, I could not see the grid as easily.

I actually took the Richter along as a last minute "what the hell" kind of decision. I have been using a Möller-Wedel collimator (what the Chrosziel and Gecko Cam collimators are based on) for the last two years and not using my Richter as much. I wanted to see what the Richter would be like with an LED (white) for the light source in portable mode.

I'm confident that I was seeing the image reflection off the image plane surface because once it was set up, all the prime lenses eye focus / measured focus matched very well. And it actually looks similar to the reflection from a ARRI fiber optic viewing screen (16SR ground glass).

BTW, the adjustment on the RED PL mount is VERY fine. The thread does not turn easily meaning a very fine pitch thread with tight tolerance.

As for the optical flat, it would have to rest directly on the surface you are checking, and it would test for flatness of the surface. I could check the flatness of the PL mount with a depth gage and teflon tip, but that would make the camera owners too nervous.

I do have concerns that the aluminum (I'm assuming it anodized Al) PL mount flange will get damaged more easily and go out of flatness. If a clumsy AC smacks part of the back of a lens into the flange. If it's treated right it will last a long time, especially if it's hard coated. But when I did my 16SR Super 16 conversions, I used Stainless for that reason, and for the improved temperature stability. Also a steel flange can be lapped if it gets scratched or dented and then you just add some shims to make up for the metal you removed.

 

[Derek Wiesehahn]

I am just wondering. The very first tests which were released by Red had a somewhat sharper look to them. I am referring to the Bubble Girl and the Milk Girls footage. Was there an olpf built into the prototype cameras when these tests were shot?

Also, would there theoretically be some way to increase sharpness in post by down-rezing to 1080p in a way similar to the Panavision scheme, or does the Bayer design of the Red make this impossible?

Finally (still on topic?), would it be possible to add in-camera sharpening to the 720P preview output as tagged metadata to the Raw file, just as is currently done with white balance and ASA?

 

[David Mullen ASC]

In other words: Is the 4096 pixel wide 4K image created by a 4096 pixel wide "window" in the sensor area, or is it created by downsampling 4520 -> 4096? And is it the active or the full pixel area that is 24.4mm wide?

The 4096 pixel frame is windowed / cropped from the full 4520 pixel sensor, so the 4096 pixel area is not 24.4mm wide, it's a little smaller (but still a little bigger than 35mm Academy/1.85 in width -- which is 22mm wide on the negative and 20.96mm wide when projected). Simple math says that if 4520 = 24.4mm, then 4096 = 22.11mm.

I don't know about any more pixels beyond 4520... probably not usable for picture information if they exist.