So about this new OLPF... Does it behave like an ND filter?

[Jason Tan]

I tried searching but couldn't find this info:

If the new OLPF/calibration adds noise, that must mean it's blocking some light. Does anyone know or have any guesses as to how much light it's blocking? I mean is it like having an ND.3 or .6 on all the time?

 

[ericyoung]

It doesn't block significantly more or less light. Its optical structure interacts with the sensor in different ways to the old OLPF which the debayering algorithms and signal processing has to take into account. Those have to be debugged and optimised to get the best performance.

 

[Lauri Kettunen]

Hmm, I think it is not technically correct to say the optical structure of the filter interacts with the sensor although the filter is designed to match with the sensor. Jason's question does have a point. An ordinary ND filter "blocks" light "linearly" (it's not really about linearity, affine is the right term). That is, what ever is the amount/intensity of light, an ND filter always blocks the same amount of light –and Eric is probably pointing to this fact. What has been explained about the new Dragon OPLF gives the impression it acts as like a non-linear filter; it is designed to block an increasing amount of light near the clipping level of the sensor, and this creates the nice roll-off of light to white. But, it seems like some light is indeed blocked meaning without the sensor one had a bigger signal and less noise. So, in tis respect I find Jason's question does make sense.

 

[Jason Tan]

The reason I'm asking is that the new OLPF obviously has stronger IR filtering which is important in certain circumstances, but by now I'm sure we have all added IR ND's or Pancro mirrors to our kit. And personally I think if you are planning to light a scene with only candles then you have to accept a bit of magenta in the flares and skin tone, some people might even argue that it's part of the look.

anyway if that extra IR filtering is costing us exposure than it would be better to let the DP add IR filtering in front of the lens when necessary.

 

[Terry VerHaar]

Doesn't the term - Optical Low Pass Filter - describe what it actually does? I always understood it as an analog to audio terms. That is, it "selectively" filters certain wavelengths of light so as to reduce moire and, in general terms, yield a more pleasing image.

In that sense, it is "blocking" some light but not, it would seem, exactly like an ND.

Have I been laboring under miscomprehension? :huh:http://www.pinterest.com/pin/create/extension/http://www.pinterest.com/pin/create/extension/

 

[Michael Tiemann]

Doesn't the term - Optical Low Pass Filter - describe what it actually does? I always understood it as an analog to audio terms. That is, it "selectively" filters certain wavelengths of light so as to reduce moire and, in general terms, yield a more pleasing image.

In that sense, it is "blocking" some light but not, it would seem, exactly like an ND.

Have I been laboring under miscomprehension? :huh:http://www.pinterest.com/pin/create/extension/http://www.pinterest.com/pin/create/extension/

lo-pass can mean many things.

I believe it's primary purpose is to combat moire, which means it lo-passes *spatial* frequencies. This is the principle function of both the old and new OLPFs.

We know that both UV and IR rejection is important, so we know that it's not a lo-pass filter according to *wavelength* frequencies. If anything, it is a band-pass filter (see http://en.wikipedia.org/wiki/Band-pass_filter)

This is a reference to a novel material that passes low-energy light while blocking high-intensity light: http://newsletter.blogs.wesleyan.edu/2014/04/18/kottosresearch/
The novelty of this approach leads me to believe it is not a part of either the new or the old OLPF designs in the DRAGON.

Thinking of the OLPF as an ND filter is a bit like saying that the additional air resistance of a car's windshield is really part of the braking system. It is not. It may add air resistance, but that's incidental to its real job, which is keeping the driver's eyes and mouth free of the many insects that fly across our roads day and night.

 

[Terry VerHaar]

lo-pass can mean many things.

I believe it's primary purpose is to combat moire, which means it lo-passes *spatial* frequencies. This is the principle function of both the old and new OLPFs.

We know that both UV and IR rejection is important, so we know that it's not a lo-pass filter according to *wavelength* frequencies. If anything, it is a band-pass filter (see http://en.wikipedia.org/wiki/Band-pass_filter)

This is a reference to a novel material that passes low-energy light while blocking high-intensity light: http://newsletter.blogs.wesleyan.edu/2014/04/18/kottosresearch/
The novelty of this approach leads me to believe it is not a part of either the new or the old OLPF designs in the DRAGON.

Thinking of the OLPF as an ND filter is a bit like saying that the additional air resistance of a car's windshield is really part of the braking system. It is not. It may add air resistance, but that's incidental to its real job, which is keeping the driver's eyes and mouth free of the many insects that fly across our roads day and night.

Haha - Michael, your example is very visual. LOL

I am curious about the "spatial" frequency thing. I am not sure I understand how one would filter light other than by wavelength. (No, I am obviously not a physicist!) UV and IR are certain wavelength bands, no? I see what you are saying with the low energy/high energy thing but, if I understand you, your suggesting that is NOT what is happening here. Can you give some example or explanation of spatial frequency filtering? Thanks.http://www.pinterest.com/pin/create/extension/http://www.pinterest.com/pin/create/extension/

 

[Michael Tiemann]

Haha - Michael, your example is very visual. LOL

I am curious about the "spatial" frequency thing. I am not sure I understand how one would filter light other than by wavelength.

Spatial resolution (aka spatial frequency).

If the above was tl;dr, then high spatial resolution is an in-focus lens with lots of sharp detail and low spatial resolution is an out-of-focus lens. An OLPF introduces just enough spatial blur that bayer patterns don't exhibit excessive moire when the spatial resolution of the lens goes above the pixel resolution of the sensor. RED gets this right. So does Panasonic. Black Magic get this horribly, horribly wrong in the BMCC and BMPCC. It is supposedly not as bad in the BMPC.

 

[Dr. Sassi]

It allows to filter the amount of details (see above - Spatial resolution) in a very precise way, precision means here, it is designed for a specific sensor only.
The frequency of the sensor and its OLPF (in terms of details) follows the Nyquist idea that the sampling frequency must be at least twice as high as the details it likes to represent. As lenses can produce too much details, this needs to be limited, otherwise you get problems.

A good definition can be found in the book, Digital Cinematography, "David Stump, ASC". Page 117 for example.

... have a look here:
http://www.red.com/learn/red-101/resolution-aliasing-motion-capture

 

[Terry VerHaar]

OK guys, getting it now. Thanks for the help and clarification. I understood the "blur" thing but didn't connect it to the term "spatial" resolution.http://www.pinterest.com/pin/create/extension/http://www.pinterest.com/pin/create/extension/

 

[Jason Tan]

Just wanted to follow up because this thread has gotten away from what I was asking and is potentially confusing. Jared has just confirmed that the new OLPF does in fact block about a stop of light. So without arguing about what an ND does, it is essentially like having an ND.3 on the sensor. Which means the camera is as noisy now at 1000iso as it was at 2000 with the old OLPF.

This is really not that bad. Being able to shoot at 1000ASA all the time is pretty amazing actually.

But now that we seem to be getting plug and play OLPFs to choose from I think it makes sense to design one that doesn't trade off visible light for better IR protection. Because it's easy to add additional IR filtering in front of the lens when needed.

 

[Will Keir]

It's a .3 ND basically. One stop of light.

The new OLPF looses about a stop in the low end ( i.e. more noise ) for the benefit of better color science and increased latitude in highlights. Some people love it.. some people hate it. D.E.B makes the noise in the new OLPF much better.. but you still are starting from an optical difference.

I tried searching but couldn't find this info:

If the new OLPF/calibration adds noise, that must mean it's blocking some light. Does anyone know or have any guesses as to how much light it's blocking? I mean is it like having an ND.3 or .6 on all the time?