Dragon Vs Alexa Low Light Tests - Someone?

[David Mullen ASC]

If you want to shoot out in blackness and get blackness in return, then most cameras can accommodate that!

As sensors get more sensitive and less noisy, obviously you can shoot in lower and lower levels of light and get decent results.

 

[Bob Gundu]

I joke with Directors and say "Remember when night actually looked like night time"?

 

[Mark L. Pederson]

I joke with Directors and say "Remember when night actually looked like night time"?

So true!

 

[Greg M]

My wish for 2014:

Let somehow people understand that the only thing that affects "exposure" and noise, is the number of photons hitting the sensor.

ISO (on RED cameras) does not alter the noise or "cleanness"/Fidelity of an image. You can rate it at "any" ISO and get identical results when you develop from "any" ISO.

You cannot change the amount of light you used in post, though…

There is no magic involved. Just physics...

I wish I had a dollar for every time I argued with a DP over this.

Happy New Year Gunleik!

 

[AndreeMarkefors]

I joke with Directors and say "Remember when night actually looked like night time"?

Hehe... which is the exact reason that my DfN grade of a Dragon frame a while back looked like this:

Not saying it's perfect, but it takes some determination to sacrifice all that detail and clarity for the sake of some realism.

I remember seeing the Troll scene in the first The Hobbit and I was jolted out of my seat on account of how fake it (night time) looked. Brightness levels were basically same as day exterior, only with a blue cast.

I understand that much of the responsibility has to do with it being a 3D release, but the 2D transfer looks the same.

 

[Alexandros Tsoupras]

In the mx sensor when you toggle raw the exposure is the same when you use iso 320. In the dragon sensor at what iso is the same exposure with raw ?

 

[Gunleik Groven]

In the mx sensor when you toggle raw the exposure is the same when you use iso 320. In the dragon sensor at what iso is the same exposure with raw ?

I can check now…


Have not, because that is not really how I use the camera. I just "get friendly" with its quirks. But I know that is good info for many.

Gimme an hour for breakfast, and I will go out and shoot something.

G

 

[Mark L. Pederson]

In the mx sensor when you toggle raw the exposure is the same when you use iso 320. In the dragon sensor at what iso is the same exposure with raw ?

I think it would be between 500-640 - BUT - I'm not convinced that the way the false color values are calculated is the RAW exposure mode is the same as false color being calculated at ANY ISO in "the non RAW aka Video" mode - that is what I need to confirm.

 

[David Battistella]

RAW is still a sort of LUT. It's just the metadata RED gives it so that you do not have to look at Liniar light (which is actually raw). I find the RAW has both changed color, tint, but not really ISO over the years. Switching between RG3, RC3 and RAW and getting the luminance(brightness of the image) to look pretty much the same, would typically be an ISO 320 look in RG3, RC3 mode, ISO 800 would look "brighter than RAW".

The RAW histogram provides the best way of knowing when the sensor is clipping.

Still even RAW is just metadata as set by red to distinguish between base sensitivity and whatever metadata look you are applying at the time of capture, but you will notice there is no RAW setting in Redcine X, but you can set to camera metadata.

If mark is right at 500-640, then the Dragon would be about 2/3 to one stop "faster" than what we are used to on MX.

It seems already though that the Dragon soaks up light much more efficiently as evidence by how far you can push a "properly" (read adequate light levels to saturate the sensors photsites with enough light for lower noise images) so much farther.

I think the real point is, if you light it right, it will look even better and you have more DR and elasticity to the image, much more play to push it around in post a bigger sweet spot, whatever. If you don't light at all then it will not be as noisy as it may have been on MX.

I am equating it to the difference between positive and negative film stocks. Say MX is more like a positive stock in that there is less margin for exposure error (a smaller sweet spot) and in Dragon, I am seeing it's much more like a negative film stock in terms of exposure room.

Nice. We can let things blow out with a bit more pleasant effect than before.

Disclaimer (this is from the wealth of great R3D being posted by a very generous few here, like Gunliek, mark, phil , peter, etc) as I have still to shoot some of my own tests in various lighting conditions.

Oh...and...Happy New Year everyone!

David

 

[Mark L. Pederson]

RAW is still a sort of LUT. It's just the metadata RED gives it so that you do not have to look at Liniar light (which is actually raw).

If mark is right at 500-640, then the Dragon would be about 2/3 to one stop "faster" than what we are used to on MX.

David

David - I agree that the display of the image in RAW mode is obviously an RGB interpretation. I don't really care what that interpretation looks like (since we are not supposed to have the ability to view a RAW image without processing anyway) - what I'm interested in proving (or disproving) is that the False Color exposure is the most accurate way to see what, if anything, is clipping or approaching and/or risking noise in the shadows - REGARDLESS of monitor out or ISO setting.

So ... again, I'm working under the theory that when in RAW mode - ISO is truly meaningless to the calculation and visual representation of colors in false color that indicate exposure levels. Further, I would NOT agree with your comment that because the RGB representation of the image in RAW mode is equal to somewhere between 500-640 in RC3/RG3- that Dragon is "2/3 to one stop faster than MX".

I could very well be wrong, but I think how the camera is calculating the color values in false color - is independent of the overall RGB interpretation.

I will dig deeper into this as soon as I can -

 

[Martin Stevens]

RAW is still a sort of LUT.

Yes, RAW view still has to apply a LUT. Since the Epic only has one sensor, true Raw would look like
a black and white un-debayered image. Something I have never seen.

 

[Martin Stevens]

David - I agree that the display of the image in RAW mode is obviously an RGB interpretation. I don't really care what that interpretation looks like (since we are not supposed to have the ability to view a RAW image without processing anyway) - what I'm interested in proving (or disproving) is that the False Color exposure is the most accurate way to see what, if anything, is clipping or approaching and/or risking noise in the shadows - REGARDLESS of monitor out or ISO setting.

Mark,

The Epic Manual confirms that the EXPOSURE tool only uses RAW data, and that ISO etc. do not
effect it.

 

[Mark L. Pederson]

Mark,

The Epic Manual confirms that the EXPOSURE tool only uses RAW data, and that ISO etc. do not
effect it.

Yup - I know that. There just seems like there might be a little confusion and/or misunderstanding with respect to how the overall image is interpreted for display vs. how the false color exposure tool indicates over/under exposure. The comments about RAW mode in MX being "same as ISO 320", etc.

 

[David Mullen ASC]

Yup - I know that. There just seems like there might be a little confusion and/or misunderstanding with respect to how the overall image is interpreted for display vs. how the false color exposure tool indicates over/under exposure. The comments about RAW mode in MX being "same as ISO 320", etc.

Back in the early days of the Red One and Epic, I found that by toggling between RAW VIEW mode and ISO settings that 320 ISO gave you something similar in exposure, hence where this comment tends to come from. Of course, RAW VIEW isn't really raw, it's processed into RGB.

 

[Mark L. Pederson]

Back in the early days of the Red One and Epic, I found that by toggling between RAW VIEW mode and ISO settings that 320 ISO gave you something similar in exposure, hence where this comment tends to come from. Of course, RAW VIEW isn't really raw, it's processed into RGB.

Understood. Based on my initial testing and playtime with Dragon my gut is telling me that false color in RAW VIEW might be much, much more relevant now as an exposure assist tool on Dragon because of the extended dynamic range and how limiting the histogram now feels. Anyway, more testing to be done. One thing for sure - you can rate this camera at very different ISO's than the MX and exposure options are different.

David are you in NYC right now?

 

[Kemalettin Sert]

14 Pages.
Where is the actual test?

 

[AndreeMarkefors]

14 Pages.
Where is the actual test?

I realize you're not asking but: under profile settings you can have 40 posts show up per page. I have only got 4 pages in this thread.

I remember the '33 pages threads' annoying me to no end.

 

[David Mullen ASC]

I'm back in L.A.

 

[David Battistella]

Mark,

i should rephrase the "faster" comment. It's more like the base viewing may have changed.

The more we use all the tools available for exposure the better the exposure and images and the less we have to worry about ISO values. ISO is a handy way of measuring light for those that use a meter and want it their lighting tuned to the relative speed of the film stock/sensor.

I don't care how RAW view looks as much as I care about having a base ISO to meter from (and I know that there will be a range just as there is a range with film stocks)

Most everyone will figure out what is best within their lighting style and the look they want to achieve.

The very best thing I am seeing is that one can rate and treat dragon as fast or slow, daylight or tungsten stock and add as little or as much light to suit the story needs and their noise tolerance threshold.

Battistella

 

[Stephen Matthews]

RAW view should be renamed to highlight check or something, which if I remember correctly is what Graeme said it was for, it would avoid a lot of confusion.