Dragon dynamic range, profiling for Sekonic

[Stephen Williams]

took the data from the site of Ryan Walter.

Finally,
Sekonic on Arri Alexa, are between 12 and 13 stop

[/URL][/IMG]

Why is 18.9 stops shown as the Dynamic range when Dragon shows 16.3 in the other graphs?

 

[John Marchant]

That Alexa graph is wrong... the upper exposure has had the wrong exif/meta applied which is why there's a whole stop gap in the centre... The right side of the graph needs to slide a stop to the left. Also the DR is being measured nearly three stops beyond the clip point. Tech testing well is a pig of a job.

 

[Gunleik Groven]

Tech testing well is a pig of a job.

Yup

 

[luigivaltulini]

That Alexa graph is wrong... the upper exposure has had the wrong exif/meta applied which is why there's a whole stop gap in the centre... The right side of the graph needs to slide a stop to the left. Also the DR is being measured nearly three stops beyond the clip point. Tech testing well is a pig of a job.

John,
is still in the last test Ryan Walter Alexa was 16 1/3, usable 13 1/6.

Data downloaded from here

http://www.ryanewalters.com/Blog/blog.php?id=8375868975844575706

http://www.ryanewalters.com/SP/sekonicprofiles.html


Alexa
- Dynamic Range (DR) at EI 800 As Profiled In Log Using the Sekonic Profile Target II [8-Bit Photoshop values]

• Overexposure (Total) = 7 2/3 [243]
• Overexposure (Safe) = 7 1/3 [235]
• At Exposure = 0 [106]
• Underexposure (Safe) = -5 5/6 [31]
• Underexposure (Total) = - 8 5/6 [22]
• Total Measurable DR = 16 1/3
• Usable DR = 13 1/6

EPIC
- Dynamic Range (DR) at EI 800 As Profiled In Log Using the Sekonic Profile Target II [8-Bit Photoshop values]

• Overexposure (Total) = 5 1/3 [237]
• Overexposure (Safe) = 4 2/3 [228]
• At Exposure = 0 [100]
• Underexposure (Safe) = -5 1/6 [16]
• Underexposure (Total) = - 8 1/3 [11]
• Total Measurable DR = 13 2/3
• Usable DR = 9 5/6

 

[Bill Anderson]

What light source did you use for this test, John?

 

[John Marchant]

This is tungsten. 2 x Arri 650 fresnels hitting the chart. I could have gained a little chunk of DR shooting the chart daylight, or filtering the lens, but this was done the 'wasteful' way by doing wb to the R3D. Worst case is what I wanted to know.

 

[Bill Anderson]

Good call, and even more impressive.

 

[Ergin Ozturk]

Thank you for the great test John.

if we assume that earlier Ryan Walter's seconic DR test and John's latest test are comparible then we practically have following table;

and according to that table we may say;

While Alexa is still better in highlights Dragon is better in lowlights.
Both camera has almost same total DR (technical 16.5 and usable DR 13.5)

As an Epic owner and dragon candidate i am super happy with that results.

 

[Martin Stevens]

Thank you for the great test John.

if we assume that earlier Ryan Walter's seconic DR test and John's latest test are comparible then we practically have following table;

and according to that table we may say;

While Alexa is still better in highlights Dragon is better in lowlights.
Both camera has almost same total DR (technical 16.5 and usable DR 13.5)

As an Epic owner and dragon candidate i am super happy with that results.

Not quite. The Dragon has MORE Dynamic Range than the Alex and Film. The charts are not quite correct.

 

[Ergin Ozturk]

Not quite. The Dragon has MORE Dynamic Range than the Alex and Film. The charts are not quite correct.

why you think that charts are not quite correct? any specific info?

 

[paulherrin]

why you think that charts are not quite correct? any specific info?

^see above...

That Alexa graph is wrong... the upper exposure has had the wrong exif/meta applied which is why there's a whole stop gap in the centre... The right side of the graph needs to slide a stop to the left. Also the DR is being measured nearly three stops beyond the clip point. Tech testing well is a pig of a job.

 

[paulherrin]

So with Dragon we would need DSC Labs latest high dynamic range test system - the Xyla-21 with 20 f-stops of dynamic range (21 steps)!!!

the Xyla-21 with 20 f-stops of dynamic range (21 steps)!!!

http://dsclabs.com/xyla.htm

you mean like this test from over a year ago?
i have a feeling there's still a bit more in there than what we're seeing in the current builds, remember the new color science has not yet landed...

http://www.reduser.net/forum/showthread.php?92158-Enter-the-Dragon

 

[Gavin Greenwalt]

That 20 stop chart has SOOOoooooooo much diffusion that it doesn't really accurately reflect true dynamic range. Remember the HVX days when everyone was rocking lowcon filters in order to eek out some extra DR?

 

[Phil Holland]

If you're curious this is how it is as of now:

Dynamic Range Summary of Digital Cinema Cameras as of 01.01.2014

Red Epic Dragon = 16.3 (w/ HDRx up to 22.3)
Arri Alexa = 14.1
Sony F65 = 14
Sony F55 = 14
Red Mysterium-X = 13.5 (w/ HDRx up to 19.5)
Canon C500 = 12

Results are "in stops" and measured at an "optimal exposure".

Note, that each camera responds slightly differently to light due to a variety of system unique properties, especially when taking into account various gamma spaces. Optimal exposure may be slightly different between each camera to get a perfectly balanced exposure based on 18% grey or mid tonal range (slightly different). A perfectly balanced exposure (the sweet spot found at what many would call the base sensitivity of a sensor rated in the EI with a ISO value) is often not needed depending on subject matter, lighting conditions, or desired targeted exposure.

Essentially each camera is it's own unique tool and optimal exposure is different across each system and might even be different for each shooting condition depending on where you would expose and rate your midtones.

Introducing a Low Contrast Filter of some sort can also increase the perceptual dynamic range, however, the measured and total possible captured dynamic range will remain the same within each camera system. However, by flattening the incoming image a bit sometimes "more" is captured and seen. The Alexa itself has a slight low con built into it. There is a tricky line to walk here as under certain conditions a low contrast filter may produce more image noise than expected.

Back to Red related things. How ISO Works with Dragon:

Red Dragon: ISO and FLUT Breakdown



The above graphic shows how ISO and FLUT work with the Red Dragon Sensor. The total Dynamic Range is always captured, however, when rating middle gray at a higher or lower ISO values the tonality is compressed or expanded across the captured dynamic range. The main advantage of this method being always having access to the maximum Dynamic Range and not inducing Clipping or Crushing when adjust ISO or FLUT in post.

In theory this gives you access to the total dynamic range at all times while working within the "recommended ISO range" of the sensor. In the case of Dragon, you're looking at a recommended range of ISO 250-2000 while maintaining separation, clarity, and distribution of tones with minimum visible signal noise. This can sometimes visually effect the perceptional dynamic range depending on what you do with the material itself in post. The totally dynamic range is always captured. The total potential is always there for maximum use. Depending on your desired exposure, aesthetic, processing, etc... you may not want to see "all" of that dynamic range however.

Going past recommended ISO values is sometimes "in the clear" depending on your overall exposure, processing, and where things land in the cameras histogram. In the case of digital cinema cameras you have a great deal of flexibility with values being exposed above the noise floor, essentially information that is exposed within the range of the workable latitude. However, if you find yourself with values in the noise floor and do increase the ISO to the "extremes", you will indeed see image noise on all of these cameras.

However, having access to the maximum dynamic range effects the workable latitude and what some would describe as overall "density" of the each camera as a medium. This is where the true allure is with these higher dynamic range cameras as most of us are chasing the subtle and intangible found inside this chromatic and tonal storm.

Canon and Sony are the interesting ducks in the room here as they have on board noise reduction which negatively effects total captured dynamic range and captured detail while positively effecting the "cleanliness" of an image. Many in high end motion picture production do not want to kill the detail found in the footage as it generally defeats the benefits of using high resolution capture in the first place. That's why the only true importance when it comes to pure image quality is what the actual usable, clean, and cinema ready ISO range is which is what the experienced big dogs look at in our industry. "Low Light King" is marketing hype in reality, though incredibly useful for certain applications where getting a clean image at high ISOs in camera is important. Post Noise Reduction can be done on the other cameras that don't have built in NR features via various software solutions if needed though.


I think that covers everything I can think of at the moment.

 

[paulherrin]

That 20 stop chart has SOOOoooooooo much diffusion that it doesn't really accurately reflect true dynamic range. Remember the HVX days when everyone was rocking lowcon filters in order to eek out some extra DR?

hahaha... but it's still sOo badass

phil, i'd think about striking 'native' iso and replacing that with 'recommended' or being precise with 250-2000

 

[Martin Stevens]

Thanks Phil. I love your detailed posts.

 

[Phil Holland]

Thanks Phil. I love your detailed posts.

Cheers Martin.

Something this illustrates, at least to me, is what a tremendous deal the Scarlet-X was at launch in comparison to other cameras at similar or significantly higher price points. Value in a total camera system is a hard thing to define, but from an image quality perspective (which is mostly what I'm focused on) we're doing rather well at the moment with Red's pricing structure across the board in comparison to other digital cinema cameras.

Still a big investment for gear at this level and that's exactly what it is.

hahaha... but it's still sOo badass

phil, i'd think about striking 'native' iso and replacing that with 'recommended' or being precise with 250-2000

There is a mathematical base sensitivity, which is a reality with sensors. What most refer to as Native ISO is tricky, though with John's measurements you can make your own conclusions to where that sweet spot is.

Red describes ISO 250-2000 as "cinema ready" performance. However, I just shot some footage all at ISO 3200 with no intention to do any noise reduction. I wanted the texture it provided. There's nothing stopping people from working realistically outside of that range. Personally I am shooting the most at ISO 250-800 as that gives me a lovely file to play with in a rather hard fashion.

Interesting enough there is something new and different about the overall higher ISO image noise with Dragon. I'm sure other will make note of it soon. It's a bit more pleasing, at least to me.

 

[Gunleik Groven]

Interesting enough there is something new and different about the overall higher ISO image noise with Dragon. I'm sure other will make note of it soon. It's a bit more pleasing, at least to me.

Agreed.

I think a huge part of that is that colors are saturated much longer "down".

 

[Robert Ruffo New]

Oh Clark, I've to hang my head in shame until Rob Ruffo tells me I can point the camera at anything but my shoes... sorry about that.

I'm not sure whether posting my actual Sekonic profile is helpful or not, as its specific to my meter... never tried sharing them.

John, I didn't like that one video you did, and I explained why. You did another sexy test on Dragon, with a girl on the beach, and it was in my opinion very charming, to give but one example of many lovely things I know you've worked on (you probably know me under another name, Rob Ruffo is my real name but not one I use as a director or DP very often, which is to say you may know me better than you think ;-) and I by no means meant a general condemnation of you as a person or a working creative, just a condemnation of specifically the acting choices (the visual aspects looked great) of one specific video, in a specific context.

 

[Gunleik Groven]

OK.

I feel kinda responsible here, as a new mod.
I should probably have cleaned up things before/when they got messy.

We are a bunch of engaged people here.

I hope we can leave this part of the discussion now