Mysterium-X...

[Michael Neborak]

There continues to be a misunderstand of "native" ISO.

Jim

I know I got it wrong.... I am sorry, not really native... but what its capable of doing and still having the same noise as 320 on the red... is that better? :emote_couch:

 

[Felix K.]

I know that with film stock you can destinguish between film speed setting and speed rating (Exposure Index). Speed setting is the sensitivity to light determined by the film emulsion and is rated e.g. with the ISO standard.
The speed rating is how this film stock is treated. A ISO 400 film could be used with an IE of 800. Then (as the the film is 1 stop underexposed - it's only half as sensitive to light as it is actually treated while filming) the picture would be too dark if it was developed for ISO 400, so they "push process" it to compensate. The picture will then of course not look as if it was shot on ISO 800. The grading will be different. (The last sentence is what I took from wiki without fully understanding it)

Now, how could you compare that to a digital sensor in general and Mysterium and Mysterium-X in particular?

 

[Gavin Greenwalt]

I know that with film stock you can destinguish between film speed setting and speed rating (Exposure Index). Speed setting is the sensitivity to light determined by the film emulsion and is rated e.g. with the ISO standard.
The speed rating is how this film stock is treated. A ISO 400 film could be used with an IE of 800. Then (as the the film is 1 stop underexposed - it's only half as sensitive to light as it is actually treated while filming) the picture would be too dark if it was developed for ISO 400, so they "push process" it to compensate. The picture will then of course not look as if it was shot on ISO 800. The grading will be different. (The last sentence is what I took from wiki without fully understanding it)

Now, how could you compare that to a digital sensor in general and Mysterium and Mysterium-X in particular?

Pretty much exactly the same. Except usually more room to work with in film. And the various unique characteristics of grain and noise when pushed a stop.

 

[Felix K.]

Pretty much exactly the same. Except usually more room to work with in film. And the various unique characteristics of grain and noise when pushed a stop.

Well it can't be that simple:

Wikipedia(Film Speed): "The usual design is that the camera's parameters for interpreting the sensor data values into sRGB values are fixed, and a number of different EI choices are accommodated by varying the sensor's signal gain in the analog realm, prior to conversion to digital. Some camera designs provide at least some EI choices by adjusting the sensor's signal gain in the digital realm. A few camera designs also provide EI adjustment through a choice of lightness parameters for the interpretation of sensor data values into sRGB."

 

[MichaelHalsell]

There continues to be a misunderstand of "native" ISO.

Jim

This sounds like a good first topic for a possible RED EDUCATION SECTION on RED.com similar to the Mythbusters section.

 

[Jim McKinney]

Pardon me if this has been asked before, but what's the difference between the "Mysterium-X" and the "Mysterium-X Select?"

 

[Jarred Land]

This sounds like a good first topic for a possible RED EDUCATION SECTION on RED.com similar to the Mythbusters section.

Great Idea Michael...

 

[J.P. Scott]

This sounds like a good first topic for a possible RED EDUCATION SECTION on RED.com similar to the Mythbusters section.

It could be called REDUCATION

 

[I Bloom]

I know that with film stock you can destinguish between film speed setting and speed rating (Exposure Index). Speed setting is the sensitivity to light determined by the film emulsion and is rated e.g. with the ISO standard.
The speed rating is how this film stock is treated. A ISO 400 film could be used with an IE of 800. Then (as the the film is 1 stop underexposed - it's only half as sensitive to light as it is actually treated while filming) the picture would be too dark if it was developed for ISO 400, so they "push process" it to compensate. The picture will then of course not look as if it was shot on ISO 800. The grading will be different. (The last sentence is what I took from wiki without fully understanding it)

Now, how could you compare that to a digital sensor in general and Mysterium and Mysterium-X in particular?

Well, in the case of film there are basically two chemical steps to get to a developed negative, the first happens when the film is exposed to light and the second happens when the film is processed. Using a more sensitive negative adds grain. Pushing a negative during processing also adds grain. But because there is natural contrast that exists in the film emulsion (the response curve) and more contrast that is added or removed by pushing or pulling a negative, the choice of what speed of film stock to use and how to process it is important. Each different process not only affects the amount of grain, but also the "look" of the final outcome. Great film photographers, like Ansel Adams, were masters of the outcome of this chemistry.

If you think about a chemical reaction in the emulsion of the film, imagine that photons are hitting the various molecules in the emulsion. Sometimes when they hit one they knock off an electron and cause a molecule to bond with another nearby molecule. Later all these molecules that bonded to other molecules form an image on the negative. As the molecules get hit and they react, there are fewer and fewer of them left that can still react. At that point the photons basically have to look around for a molecule that hasn't been hit. But even though there are very few left there are always just a few more that could give you just a little more information in your image. This is why film has such nice "roll off" in the highlights, because it gets exponentially harder and harder to cause absolutely all of the emulsion to react.

In the case of digital, the world is completely different. There is only one step in the process, so to speak, and that is when photons hit the photosites on the sensor and knock off electrons producing measurable charge in the circuitry of the sensor. The sensor then measures the charge which you can think of as basically counting the electrons. Think of this like marbles falling into a cone. The vast majority of the marbles, will revolve around and eventually go to the bottom of the cone where we can count them. Just a few will get flung out and land somewhere else.

Because we are literally measuring the quantity of charge in each photosite the process doesn't introduce any roll off. (It's not getting harder and harder as with film) You can see from the chart the natural response curve of the sensor is basically a straight line. A film-like response curve is something we add later using software so that the resulting image is pleasing to our eyes.

The only real difference between various sensors is how much noise is introduced to the image and how precisely it can measure charge. In other words some sensors might count electrons by the thousands and others by the hundreds; the latter giving you more fine grained images with more tonality. The same sensor might count electrons by the hundreds but it's count might be off by give or take a 500 electrons while another is off by give or take 50 electrons; the latter giving you an image with a lot less noise.

So when they say there is no "native" ISO. What they mean is "Hey, we just count electrons here. Starting at ZERO. The more accurate the count of actual electrons the better our sensor. The ISO is all you." With film we selected an ISO because it helped us get similar contrast with various amounts of available light. In digital there is no need to make that selection because the selection of ISO doesn't affect contrast, it only affects the amount of noise we pull out of the image and the amount of rolloff we can simulate.

The curves that are applied to the image to emulate different ISO settings must amplify the blacks in the image more or less depending on how much of all the possible values you would like to perceive as black. In a higher ISO we have less of the possible values that we will perceive as black, so we have to expand that part of the image more to achieve the same contrast. When we expand it more, we see more of the noise.

The question of what ISO is acceptable (i.e. native) is always a matter of opinion depending on how much noise is acceptable to your eyes.

I hope this is helpful, obviously I'm glossing over a few things here.

IBloom

 

[Joseph Hutson]

It could be called REDUCATION

Oh, what a novel idear. :smilielol5:

 

[Stephen Williams]

There continues to be a misunderstand of "native" ISO.

Jim

Hi Jim,

I would understand native ISO to be the point where you have the maximum exposure latitude. Are you saying that the 13 stops of dynamic range is not fixed in the same way as on the current sensor.

Best

Stephen

 

[Jannard]

Hi Jim,

I would understand native ISO to be the point where you have the maximum exposure latitude. Are you saying that the 13 stops of dynamic range is not fixed in the same way as on the current sensor.

Best

Stephen

I don't think I understand your definition nor did anyone say 13 stops. Maybe you should read Graeme's comment about native ISO?

Jim

 

[Shawn Nelson]

Just got back from 4 days of using Red in the nether regions of Oregon, just read through the thread and see my question from page 4 never got cracked at, let me re-phrase it. Since you're saying M-X 800 is to Red's 320 (whilst firmly denying the existance of a native iso), I'm confused. I'm trying to practically understand if it will be harder for me to expose broad daylight situations. Red is such a double-edged sword for me, on one end I'm trying desperately to get it to be more low-light capable (when i'm deep underground with only torches) but then in the high desert I'm at a f11 with a ND1.2 in!! Will I need even more ND on M-X to get the same highlight protection in the same seen as Red currently is at? Also, will I be able to turn down the ISO without penalty? It's been said repeatedly that there is a penalty for going below 320 on Red, hopefully that's not the case on an even more sensitive M-X, because I really need to be able to tone down the electronic sensitivity of that sensor...or go buy extra matte box stages and a whole new range of ND

 

[Shawn Nelson]

Btw, as a tangent, after backpacking yet again with Red, I think my favorite two features of Epic are vastly lighter configurations and 2 sec boot time

 

[Gavin Greenwalt]

Well it can't be that simple:

Wikipedia(Film Speed): "The usual design is that the camera's parameters for interpreting the sensor data values into sRGB values are fixed, and a number of different EI choices are accommodated by varying the sensor's signal gain in the analog realm, prior to conversion to digital. Some camera designs provide at least some EI choices by adjusting the sensor's signal gain in the digital realm. A few camera designs also provide EI adjustment through a choice of lightness parameters for the interpretation of sensor data values into sRGB."

Digital gain adjustment can be done with film or digital source so in that respect manipulating your footage is pretty much exactly the same except for the nature of noise vs grain. The particulars of which were well explained above.

Red has no analog gain so in that respect it's exactly the same as film, all manipulation of the image has to happen on the footage. If you think of r3d files as a processed negative everything becomes pretty simple.

The only time I've ever found camera side gain adjustment to be advantageous is when the image is about to be heavily compressed. Setting the gain before hand allows the compression to just try and store the range of color I need before reducing the bit depth.

 

[Sanjin Jukic]

Just got back from 4 days of using Red in the nether regions of Oregon, just read through the thread and see my question from page 4 never got cracked at, let me re-phrase it. Since you're saying M-X 800 is to Red's 320 (whilst firmly denying the existance of a native iso), I'm confused. I'm trying to practically understand if it will be harder for me to expose broad daylight situations. Red is such a double-edged sword for me, on one end I'm trying desperately to get it to be more low-light capable (when i'm deep underground with only torches) but then in the high desert I'm at a f11 with a ND1.2 in!! Will I need even more ND on M-X to get the same highlight protection in the same seen as Red currently is at? Also, will I be able to turn down the ISO without penalty? It's been said repeatedly that there is a penalty for going below 320 on Red, hopefully that's not the case on an even more sensitive M-X, because I really need to be able to tone down the electronic sensitivity of that sensor...or go buy extra matte box stages and a whole new range of ND

I wouldn't say that you will need more NDs at daylight with M-X sensor.

That would be in a way bit "ridiculous".

Following "my simple logic" the new M-X sensor would give you much much better low light/available light performance

(more "headroom" to choose your own "noise tolerance" adjusting or "pushing" ISO higher in a RAW file after that at post)

and that's all what we were asking from the beginning on of these sort of discussions and speculations here.

 

[C_Skaff]

Got a Q for you guys...

I came across a clip shot on Cuba on RedOne.
The used a ND-9 filter and was still forced to set the ISO to 100.
This caused some problems, the green in the palmfeafs was gone, they looked brown in some shots.

Someone explained it by saying that a ND-9 doesn't filter out the UV-light witch made the image get correct exposure EXCEPT in the UV-spectra. Witch resulted in the weired non-green-leafs-look.
Apperently they should hae been using a filter called "Hot Glimmer" or something like that, to get both ND-9 + UV filter.

Apperently ArriD21 has that "built in" on the sensor...

If the new M-X sensor is ISO 800 that might be more of a problem since it will demand a lot more ND being used...
Will it have a UV filter on the sensor?


(I could be very wrong with the filter name and even if it is actually UV they were talking about, sorry for that, but its not the specific terms that are important, more the theory of the issue. Hope you understand what I mean

/Carl

 

[Hans von Sonntag]

Also, will I be able to turn down the ISO without penalty? It's been said repeatedly that there is a penalty for going below 320 on Red, hopefully that's not the case on an even more sensitive M-X, because I really need to be able to tone down the electronic sensitivity of that sensor...or go buy extra matte box stages and a whole new range of ND

This my take:

I guess that M-X will be a sensor like the current M with broader dynamic range and higher sensitivity. Like with M there will a sweet spot where maximum dynamic range meets minimal noise. This sweet spot can be a peak or a relatively broad plateau of 1 stop or so. Currently for high contrast scenes this is 320 metadata ASA. With M-X, this is going to be 800 metadata ASA.

Since M-X has a broader dynamic range it will accept lower ASA rates, say 200 ASA, without significant loss; especially if you point the sensor to a low-contrast scene such as an overcasted panorama or a controlled studio environment. With Red's tools (histogram, false colour, etc...) it is easy to decide where the sweet spot for a given scene is. Not always maximum dynamic range is required, hence overexposing to a certain extend is possible.

In the case you want the highest dynamic range 800 metadata ASA is going to be the equivalent to the 320 metadata ASA we have in the moment. Shooting a high cotrast scene in the desert requires the highest possible dynamic range because you want both: maximum information in the shadows AND the highlights. Thus, the purchase of more NDs is meaningful.

The big question is: How deep can you underexpose M-X? Will it be capable to expose skintones at 1600 metadata ASA with acceptable noise? Will M-X be able to shoot easily at 800 conventional, meter ASA?

Anyway, Epic-X will be a huge step.

Hans

 

[Hans von Sonntag]

This caused some problems, the green in the palmfeafs was gone, they looked brown in some shots.

/Carl

Carl, this is an infra red attack. CMOS sensors like the RedOne are very sensitive to red and infra red. A ND filter does not filter the infra red part of the spectrum. By a given amount of infra red (IR) the IR part will increase dramtically the more ND you stack in front of the lens. Interestingly the amount of IR pollution is changing from hardly any to severe. Only in rare events it becomes so drastic like in your case. There are Hot Mirrors available which stop the IR. A combination filter of ND9 plus Hot Mirror is a meaningful, basic filter for the RedOne.

Hans

 

[Bob Torrance]

Hi Jim,

I would understand native ISO to be the point where you have the maximum exposure latitude. Are you saying that the 13 stops of dynamic range is not fixed in the same way as on the current sensor.

Best

Stephen

What if you set the exposure where you have no highlight clip at all then see how far down into the shadows you can go. When you hit too much noise, there is the sweet spot. If you open up, you see more shadows but have clipped your highlights. With pure white, not hits off chrome, of course. When you max out sensor white, wouldn't this be "native ISO" or sweet spot? Of course you can choose to let highlights clip (or make it look intentional).

Bob Torrance