Jump to content

RED BUILD 14 FULL TEST


Zac Halberd

Recommended Posts

Evangelos, thanks for your appreciation. My name is Dapan Joofa, but I prefer to go by just Joofa. I am an engineer.

 

I have followed some of your and Graeme Nattress comments. I agree with many points of both of you.

 

I agree with Graeme Nattress that "down stream" stuff such as gamma, etc., should not be used for dynamic range calculation.

 

On the other hand, I also agree with you that Jim Jannard should not have provided ImaTest as an example of dynamic range of Red One Camera, if Graeme Nattress has reservations about it.

Link to comment
Share on other sites

  • Replies 112
  • Created
  • Last Reply

Top Posters In This Topic

We've only ever relied on our own way of doing things. The production of the Imatest screen was to show exactly what I've been talking about - that by mucking around with curves it gives different results, and hence is not reliable for us.

 

Joofa - you're right that dynamic range should be on linear light data as that's the only common reference. Then you're actually measuring how the sensor respond to light, it's linearity, and the noise as it does so. With the funky curves that some cameras burn into their data, you have absolutely no idea what the actual light level that code value refers to, which was fine back in the video age, but is not so cool in the data centric digital cinema age.

Link to comment
Share on other sites

Certainly you can discard part of the dynamic range, but there's no real way to gain more range.

 

It is possible to extract more dynamic range if DR measure is defined in terms of SNR -- not by increasing signal, but by decreasing the noise levels. For uncorrelated noise with the same mean and standard deviation at different sample locations, the standard deviation of the average of noise decreases. Chebyshev inequality can be used to verify that; basically, the average of a measurement is likely to be closer to the actual value, even if each of the measurement involves more or less error. The signal characteristics are assumed to stay the same; cross-correlation between signal and noise is not present by assumption. Hence, SNR should increase and therefore, DR.

Link to comment
Share on other sites

It is possible to extract more dynamic range if DR measure is defined in terms of SNR -- not by increasing signal, but by decreasing the noise levels. For uncorrelated noise with the same mean and standard deviation at different sample locations, the standard deviation of the average of noise decreases. Chebyshev inequality can be used to verify that; basically, the average of a measurement is likely to be closer to the actual value, even if each of the measurement involves more or less error. The signal characteristics are assumed to stay the same; cross-correlation between signal and noise is not present by assumption. Hence, SNR should increase and therefore, DR.

 

At last someone that understands.

 

The question is, for a cinematographer, yes I have 11 stops but which of them are really useful?

 

Signal to Noise Ratio, S.N.R. is the analysis of Imatest but expressed to F-stops not to db?s?

 

And expressed to F-stops to be easily understood from cinematographers!

 

That?s why the noise reduction and the correlation is resulting to better SNR or more useful range.

 

Some where I read that the sensor is hard balanced to 5000 Kelvin, that can be another cause of problems, since interiors with 3200K will be more noisy? unbelievable!

Link to comment
Share on other sites

  • Premium Member
Some where I read that the sensor is hard balanced to 5000 Kelvin, that can be another cause of problems, since interiors with 3200K will be more noisy? unbelievable!

 

This seems to be common with a few other CMOS cameras. They really are a pain to work with sometimes because the only way to get a decent image out of them is to gel the lights and get the color right otherwise when you're shooting with tungsten lights and you try to correct in post you start to pull too much red out and add too much blue and your image gets very noisy and contrasty. This is a problem I run into with several High Speed CMOS cameras, it can be fixed but you just need alot more light... umm almost 5 times more light to be exact (because of the CC gels).

Link to comment
Share on other sites

  • Premium Member
It is possible to extract more dynamic range if DR measure is defined in terms of SNR -- not by increasing signal, but by decreasing the noise levels.

How do you do that with real world images? If you have the leaves of distant trees in the background, the grains of a stucco wall in crosslight, stuff like that would have to be distinguished from thermal noise in the chip. Can you do that?

 

 

 

-- J.S.

Link to comment
Share on other sites

How do you do that with real world images? If you have the leaves of distant trees in the background, the grains of a stucco wall in crosslight, stuff like that would have to be distinguished from thermal noise in the chip. Can you do that?

 

-- J.S.

 

Your point is well-taken.

 

Leaves of trees are difficult candidate as they may cause Moire-type artifacts if the image variation becomes smaller than Bayer pattern spacing of individual Red, Green1, Green2, and Blue channels for the Red camera. Even if they don't cause artifacts, in the maximum case where the variation of image is just matched, or close to the spacing of the color channels in the Bayer pattern, simple deBayering techniques that do signal averaging per channel may not benefit from the scenario I was describing in my posts earlier.

 

However, for images where tree leaves are not significant portion of the whole image, then I would tend to believe that for the rest of scene the per-channel signal variation in the immediate neighborhood of a given location in the Bayer grid may be less than the noise variation in the corresponding positions. Hence, if we are able to reduce relative noise variation more than signal variation, then we may get a cleaner signal because of a greater reduced noise.

Link to comment
Share on other sites

  • Premium Member
However, for images where tree leaves are not significant portion of the whole image, then I would tend to believe that for the rest of scene the per-channel signal variation in the immediate neighborhood of a given location in the Bayer grid may be less than the noise variation in the corresponding positions. Hence, if we are able to reduce relative noise variation more than signal variation, then we may get a cleaner signal because of a greater reduced noise.

Ah, but we'd have to production design that sort of detail out of the show -- much like we learned what to avoid with interlace and chroma subcarrier aliasing in NTSC. A lot of us like the look of film grain, which is the same kind of detail.

 

Couldn't pretty much the same noise reduction be applied to any camera, making it a rising tide that lifts all boats? To have an apples to apples comparison, wouldn't we have to get test images without any noise reduction from each camera? Is that possible?

 

 

 

 

-- J.S.

Link to comment
Share on other sites

The question is, for a cinematographer, yes I have 11 stops but which of them are really useful?

 

Signal to Noise Ratio, S.N.R. is the analysis of Imatest but expressed to F-stops not to db?s?

 

And expressed to F-stops to be easily understood from cinematographers!

 

I agree with you that an SNR based analysis via a software is an acceptable and legal measure of DR. I think the point of contention was that Graeme Nattress is of the view that ImaTest is not predicting the applied gamma accurately enough. Therefore, the discussion boils down to should SNR be done before gamma or after gamma -- and it is difficult to have a unanimous opinion on this.

Link to comment
Share on other sites

And that assessing DR based on stops would, to me necessitate it being calculated back to a reference of linear light data, and if that is not done accurately, the measured scale of the noise will be incorrect.

 

We do capturing a calibrated step chart (like Danes Picta TS28D or Stouffer 4110) that is ACCURATE REFERENCE TO LINEAR LIGHT and to emphasis on that, I own an X-rite 310 status A transmittance densitometer and I measure every step and correlated with the certificate from the manufacture and the error is less than 0,01D.

 

In Imatest we are stating what chart we used exactly by selecting the chart type.

 

So the target is ABSOLUTE PERFECT AND IS BEING CALCULATED BACK TO THE LINEAR LIGHT REFERENCE.-

 

As for gamma curve... lets just use what the user will use in a real life project...

 

And Graeme all the cameras are writing data either on tape or on disks, RED, because of the sensor, writes bayer data thats all.

 

So please no more "data centric digital cinema age" marketing arguments...

 

Again we are not playing marketing games here!

 

Work with your camera to make it better, accept positive criticism, like the one thats on that thread, with no excessive aggressiveness from your side... Try to be humble, at the end it will pay back.

Link to comment
Share on other sites

So what we actually analyze is the dark steps. These steps because of the nature of the measurement are very sensitive to noise reduction.

 

I agree with noise reduction in darker image areas. I have seen it many times how it improves the overall perceived image quality.

Link to comment
Share on other sites

Couldn't pretty much the same noise reduction be applied to any camera, making it a rising tide that lifts all boats? To have an apples to apples comparison, wouldn't we have to get test images without any noise reduction from each camera? Is that possible?

 

 

I was talking about noise reduction which could inherently result from naive deBayering techniques that do simple averaging. I am sure noise reduction is employed by different cameras at many other levels.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Forum Sponsors

Metropolis Post

New Pro Video - New and Used Equipment

Gamma Ray Digital Inc

Broadcast Solutions Inc

Visual Products

Film Gears

CINELEASE

BOKEH RENTALS

CineLab

Cinematography Books and Gear



×
×
  • Create New...