All the Epic/Dragon Vs. Alexa/Sony threads are belong to us...

[Matt Ryan]

Can someone confirm this? Given the credentials behind these test takers, I'd be shocked if something as rudimentary as white balancing wasn't controlled. If that was indeed the case then I'd begin to question just how much stock I'm giving these results as well.



So we're arguing semantics then? If we have allowed the distinction of latitude to become so ambiguous and immeasurable, shouldn't Alexa claim to have 18 stops of dynamic range?

Beautiful colors on the bottom two images btw, thanks for sharing.

I agree. From these tests if Alexa has 14 stops then Dragon has slightly less. You guys can throw around technical numbers all day but when you compare them side by side, Dragon has slightly less DR. If all the new OLPF did was move the calibration up slightly to give 1.3+ stops more in the highlights, but in return lose a bit on the low end, then were still less (practical) DR than Alexa. While all of Phil's and others theoretical DR numbers are nice, it's not practical at all. If Dragon is indeed a practical 16.3 stops, then it should have an obvious advantage over Alexa regardless of the test. Unfortunately, that does not seem to be the case yet. I do agree by far that Dragon is cleaner than the Alexa noise wise, but the DR isn't there yet from these tests...

Long story short is if Dragon has 2+ more usable stops over Alexa, then it would be more obvious in shootouts...

 

[Phil Holland]

Can someone confirm this? Given the credentials behind these test takers, I'd be shocked if something as rudimentary as white balancing wasn't controlled. If that was indeed the case then I'd begin to question just how much stock I'm giving these results as well.

So we're arguing semantics then? If we have allowed the distinction of latitude to become so ambiguous and immeasurable, shouldn't Alexa claim to have 18 stops of dynamic range?

Beautiful colors on the bottom two images btw, thanks for sharing.

Yep, from the horses mouth even.

No, all cameras were shot in preset 5600 where possible 5500 where not.

There was no tweaking of white balance or tint or anything!

Which I agree with in the scale and scope of this test as it gives us a general idea of how cameras look different right out of the gate.

Workable latitude is different from total captured dynamic range and I don't think we are arguing semantics. This is simply not a test that measures total dynamic range of a system. This UWE test does in fact compare cameras side by side and reveals how they are different.

The biggest thing that needs to be acknowledged is there are ways to shoot and "develop" all of the cameras to produce the highest possible level of image quality.

I agree. From these tests if Alexa has 14 stops then Dragon has slightly less. You guys can throw around technical numbers all day but when you compare them side by side, Dragon has slightly less DR. If all the new OLPF did was move the calibration up slightly to give 1.3+ stops more in the highlights, but in return lose a bit on the low end, then were still less (practical) DR than Alexa. While all of Phil's and others theoretical DR numbers are nice, it's not practical at all. If Dragon is indeed a practical 16.3 stops, then it should have an obvious advantage over Alexa regardless of the test. Unfortunately, that does not seem to be the case yet. I do agree by far that Dragon is cleaner than the Alexa noise wise, but the DR isn't there yet from these tests...

Long story short is if Dragon has 2+ more usable stops over Alexa, then it would be more obvious in shootouts...

I need to be explicitly clear. There is nothing theoretical about measuring stop by stop increments. That is hard data. How you use that measured range is entirely different conversation especially when applied to common gamma and looks.

Also, in the literal side by side with the old and new OLPF and Calibrations you can see that if anything, based on the local 18% Gray meter/exposure we have more range in the shadows as well as gains in the highlights. The New Calibration is not simply a "linear slide to the right". That's not how that was remapped at all according to the data. If I had to describe it exactly in some sort of metaphorical term I would say it's been stretched a bit across the range and the top clip has been reevaluated while still maintaining a nice "burn out". These images have been posted and that data is also "hard".

The thing that this UWE test doesn't show is exactly where that DR is placed. Moreover, it actually can't be measured in this particular setup accurately except for the differences at the top and bottom ends. It is a good side by side and "differences" comparison though.

I'll stress again, to get the best out of all of these cameras they can all be shot and worked with differently.

I do have to balance between practical shooting and technical shooting, but if you are indeed measuring total DR you better come up with a hard answer that can be repeatable under controlled conditions. I will say it does come down to what you can do with each of these cameras and that's the most important thing. Some may derive that from working with the system and "feeling out" the best ways to shoot. Some may base that off of numerical data. Doesn't matter if in the end you end up with good images.

 

[Gunleik Groven]

My enthusiasm for Dragon comes from how it works as an image-system.

I have seen the DR expand significantly since the proto and up until today, and I see the potential for more.
The test was focused on UHD/1080 delivery. I hope Geoff puts up the final test, which is all the cameras exposures "normalized" to base exposure at UHD and 1080 delivery.
He is planning to upload that to Vimeo, which obviously does not support UHD. What I have suggested is that he makes a 1080 center-crop of the UHD image to check how the images scale (which is a significant part of usability for the different cameras)
If he does that, the "original" files will be downloadable for you to have a look at. But you can basically do that test now with the files available.

I would urge you to do that.

If that is not done by him, I might. I just have a few other things I need to get done before I go into geekery again.

On RED and Alexa, WB'in the cameras prior to hit record would not make any difference. If you think it would, you could just WB them in post. Same image would have been captured in either case.

The highlight difference did not come as any surprise to me, after testing vs Epic, and knowing the Alexa.
Now. What tests like these does not show, and I did not "get" until having used the Dragon for a few days, is how differently Dragon captures lights and colors than any other system I have used.
I did capture a couple of "practical" scenes with the Dragon/F55/Alexa in ext daylight/high contrast, to play with.

I have not had time yet, but I will make them available, if Geoff does not. They were designed by me. Not by him. I see nothing wrong with how our main comparison was done as a side-by side and agreed with him on all camera-settings. If it is faulty, it is faulty in a way many will use these cameras... And thus I have not wanted to do competing analysis, as that would just create/add noise. I am more than happy to see that noise gone.
And I agree with what he sees. Hard not to. RAW files are available
Download the files and analyze them yourself. Nothing was done to cripple any cameras. And I can say that I found bigger surprises in other cameras behaviors than in Dragon and Alexa.
As far as I can see, and as Phil says, the tests are good for what we tested for.

Still I happen to agree with Phil on all he says. And I do not see them as contradictory statements.

I know this may sound fanboy'ish, and I might well be proven wrong, but I think I see how RED can re-calibrate up, and still be able to go further or similarly "down" with the OLPF and re-calibration. I hope the image stays as "fat" as it currently is even after that, and I think it can well stay that "fat". It is hard to explain how well the captured images performs when taken into post. I know nothing like it.

While I think DR is important, and that Dragon can/will/might match/surpass the Alexa in that regard, it is only ONE of many parameters for a system. Usability of the image, flexibility, and dare I say: price, are all important factors.

One thing not mentioned a lot here, which I saw clearly on this test, is how much the Alexa has improved over its 4 years.
And those are all actually good arguments for Dragon.

Alexa is arguably a "matured" camera compared to the Dragon, and is a lot better than it was when released. I do not see much more extra potential there, and resolution is not unimportant when delivering UHD.
The Dragon OTOH is in the beginning of its evolutionary cycles.

Part of the background for the test is that UHD delivery for TV will be/is a reality in many markets NOW. (whether we like that or think it's important is a totally different discussion).

Geoff has decided against publishing the resolution-tests, based on framing issues. And I can understand that. No need to build up conspiracies for that, but you can get a decent hunch if you take the base exposures released, and scale them to UHD/1080 and compare the noise/detail, as to which cameras performs well when shot in range and what kind of colors you can get from them.

Green cast, does it worry me?

Nope. It is a tint-adjustment. Tint-adjustents are non-destructive.

And I think from what we have seen from Phils examples, that "Dragon Color", is not a new gamma/RedColor, but a bottoms-up new sensor-calibration. New calibrations changes more than a new gamma can.

As a camera to take home from the UWE test, I a happy I have the Dragon.

And for all practical color/shooting/post situations, that stands true for me, and I think you should take Geoffs remarks about how he sees that the Dragon might surpass the Alexa as to DR as a good hunch, as to the potential he too sees in the Dragon images. Around these corners he has not been known to be much biased positively towards RED (even though he has practiaclly used the REDs more than many here.. )

Now.

Photography is not a pissing-contest on single technical parameters.
Many outstanding film-stocks has less DR than Dragon/Alexa has. They are still super-good.

As to exactly measuring DR, I guess Xyla/Phil/John are onto precise tech measurements. Which would affect other cameras, too if measured in the same way.

As to what happens when you shoot these cams side-by-side with the same settings and same scene/lights, I cannot see that what we did in Bristol is flawed. And I cannot see that the way of measuring practical DR is bad or worse than when I compared the Epic with Dragon in a similar way. It is a "practical use" example, where we pushed all the cameras over and under their capacity, whatever you measure the DR to.

As a side-note:
For those looking for a "center" ISO (where you have a similar amount of stops over/under), it is designed in a way that could give you a pretty good indication if you can interpolate the data available.

Then again, that will change. It is only relevant "now".

I wholeheartedly agree with Phil/Mark and others: To get the best out of each camera, they should NOT be shot in the same way. You would light/filter differently for all of them, and a good DoP can get great images in many situations for most/all of them.

The Dragon was black-shaded and shot at that temperature and within shutter/black-calibration range, except for the lowest exposure. If that affects the lowest exposure, it is my bad, not Geoffs! But still represents a practical use I might addhere to in the field.

I will not go into a discussion as to how you read the tests, or Geoffs analysis.

I was much more curious about the non-showing F65, than the Alexa results, and I look forward to seeing that. But as a practical "usable" camera, there is little comparison between F65 and Dragon.

(Side note:
Back in the day, Dalsa may have been a "better" camera under optimum conditions than the RED One, and thus won the James Bond VFX gig over RED, which pissed Jim off.
But Dalsa never made it into practical use, it seems also not through the Aaton partnership.
The same cannot be said about RED.

There is a bit of a distance between measurable "tests" and practical usability.)

Practically I see a lot of more "RED gigs" go to C-500/F55 than to Alexa.
I see the Alexa chosen over RED for a number of other parameters than just the images, which Mark P and I have been discussing a bit in other threads.
Basically "set-friendlyness" and "AC-friendlyness" for larger-set setups.

In many ways the C-500/F55/Epic/Dragon comes out as pretty similar practical-use scenarios for larger sets (as to the amount of kit and rods and "things" you have to add to make it "useable" for larger sets, and also having a basic small form-factor), and then, to me, the REDS are much more user-friendly. (Not to mention price when you add up RAW recorders and post-workflow)

I think that part of the test, is a bit under-focused here and other places.

I would choose the Epic over the Alexa for many uses. The Dragon has not really tilted that towards Arri...

I mean, for all those going raving over the C-500lowlight capacity...

Anyhow.

Just my 2,5 cents.

Tests like these will be done and will be published. I think this was performed in a practical and reproduceable fashion.
And gives some valid information for the scenarios tested.

And it does not change much on my view on Dragon, other than that it did about as expected.

At 9.500$ the Dragon upgrade is a no-brainer.
(compared to any other "upgrade" you can do.)

AT 50k$, the CF makes a lot of sense compared to any other available camera out there.
And I happen to think this test confirms that.

 

[Gunleik Groven]

One point which I have tried to hammer down since I got to test the prototype, is that it is not all about "measurable DR". Controlling DR in a scene to our liking is basically what we do as photographers with any camera or filmstock.

Even if/when Dragon "wins" the DR contest, it is not what determines if the image is "better".

How the image and colors perform "within" DR is what makes a camera good/better/not-so-good.

And it is all the image characteristics that and the image-flexibility which to me is why I love it.
It is just "juicy and yummy".

"Measurable DR" is not what wins many over (even if RED happens to win that contest in the end)
Image characteristics will. "How easy it is to get the image you pre-imagine, with the tool of choice".

Dragon is a massive achievement compared to anything in that regard.

Because it does not get "thin" in any part of its range, like most/any other camera (including Alexa) does in parts of its arguable technically measurable range.

How tonality and saturation holds up close to clip (measured on the cameras own clipping point) and bellow "optimum" exposure, has a lot of more meaning as to a cameras practical dynamic range for a cinematographer, than any luma-measurable-number can ever give you. We are used to control for those.


As said a number of times:

All tests are flawed. The ones I do included.
AND.
It is not ONLY about the data, but how you interpret them.

Thus you should do your own tests based on your own user needs.

Tests like the CML/UWE tests, the ones I do nd the ones Phil does, all just shows what happens if you do what we did.
Most probably you will want to do something else...

 

[Ergin Ozturk]

Phil "Putty in broad daylight." is beautiful image and colour.

 

[Gunleik Groven]

Phil "Putty in broad daylight." is beautiful image and colour.

And a very convincing argument for my points.

 

[James_Mills]

Yep, from the horses mouth even.

I'm confused, doesn't this mean they were white balanced to a standard Kelvin setting? Shouldn't that be more telling than a manual card white balancing for each? I'd like to be able to trust my Kelvin readings in the field, no?

Workable latitude is different from total captured dynamic range and I don't think we are arguing semantics. This is simply not a test that measures total dynamic range of a system. This UWE test does in fact compare cameras side by side and reveals how they are different.

The biggest thing that needs to be acknowledged is there are ways to shoot and "develop" all of the cameras to produce the highest possible level of image quality.

If this figure is concrete and the term "workable latitude" is indeed a discernible statistic, then why have we not used this figure in the past ten years of filmmaking when regarding to other cameras' latitudes? Why aren't 5Ds marketing themselves as 14 (workable) stops of latitude, and why isn't Alexa claiming to have 18+ stops of (workable) latitude? Had the C300 came out and claimed to have 15 stops of "workable latitude" before its release a year and a half ago, would you honestly not have taken offense to the claim?

As to exactly measuring DR, I guess Xyla/Phil/John are onto precise tech measurements. Which would affect other cameras, too if measured in the same way.

My point exactly. If Phil were to do these same "workable DR" tests with Alexa, what could possibly suggest that the Alexa wouldn't again surpass this 16.5 figure? Therefore I believe any educated filmmaker who looked at RED's claim of 16.5 stops of latitude objectively would come to the logical conclusion that it is a very specious claim.

I'm a proud owner of the Epic and an admitted fanboy, but making claims like these is the exact reason RED gets so much hate from a lot of filmmakers. The Dragon is amazing in so many other ways that it doesn't make much marketing sense to exaggerate or aggrandize anything (and especially something as critical as dynamic range, the very figure that many choose the Alexa because of).

Thank you guys both for the long and well thought out posts. I'm not trying to troll here, I know how much work you guys put into this and in no way am I trying to discredit any of your tests or previous claims. I'm merely perturbed at this 16.5 DR claim because I believe it hurts our company's integrity (that which is already being constantly incised by all the haters out there).

 

[Gunleik Groven]

I'm a proud owner of the Epic and an admitted fanboy, but making claims like these is the exact reason RED gets so much hate from a lot of filmmakers. The Dragon is amazing in so many other ways that it doesn't make much marketing sense to exaggerate or aggrandize anything (and especially something as critical as dynamic range, the very figure that many choose the Alexa because of).

Right on.

That is exactly why I did not comment on DR until quite late after getting the camera.
And not at all on the prototype.

As to Phils "workable latitude" term ,the things I mention on how it is approaching clip, how colors "hold up" in the last few stops and how it clips are very important parameters as to how the DR "feels". and practically works when you shoot it.

One thing I have mentioned a few times, is how the Dragon is reassuring when you shoot, in much the same way as the Alexa "feels" good on set.

As to getting good imagery out of them, I think that is only good. It feels reassuring to shoot them. The Dragon, too.

 

[Phil Holland]

Phil "Putty in broad daylight." is beautiful image and colour.

And a very convincing argument for my points.

Thank you Ergin and Gunleik. I'll have some new stuff in a couple weeks that I can show that should be "pretty" as well.

The color and the DR are there. What people expect to see and what people want to see is largely going to depend on their expectations. Dragon is pretty damn good "straight out of camera" and there's a great deal of control after the fact. That's also true for all digital cinema cameras.

I also should be clear that much of my opinion about Dragon's performance is based on my experiences with film, the Alexa, Genesis, and F65. Sometimes in direct comparison.

I'm confused, doesn't this mean they were white balanced to a standard Kelvin setting? Shouldn't that be more telling than a manual card white balancing for each? I'd like to be able to trust my Kelvin readings in the field, no?

Depends. Geoff's test is showing what each camera "thinks" 5600/5500K is and that's a useful info for what you are talking about. What each camera sees as 5600K is going to be slightly different. Always, and again, sometimes slightly different per each camera by the same manufacturer. There's a general baseline and what I would describe as a tolerance. Most tend to be within a couple hundred kelvin and a bit of tint adjustment away from each other. The real question of what is right and what is wrong is often a mixed answer. Mathematically when using several of the higher end charts we can properly expose them and find out (read profile) each camera under each lighting condition.

Which is another factor. The light.

Light sources that "are" daylight often fall on one side or another, contamination/tint due to diffusion, etc.... Practical Daylight shooting sometimes isn't due to the contamination of the sky down or stray ambiance and radiosity coming from various other locations.

Lenses themselves also provide a color cast. ND filters, them too.

One thing I'll say, at least when it comes to actual daylight and tungsten WB values under the new OLPF/Calibration is that Red has decided to favor a slightly warmer gray, which is nice.

If this figure is concrete and the term "workable latitude" is indeed a discernible statistic, then why have we not used this figure in the past ten years of filmmaking when regarding to other cameras' latitudes? Why aren't 5Ds marketing themselves as 14 (workable) stops of latitude, and why isn't Alexa claiming to have 18+ stops of (workable) latitude? Had the C300 came out and claimed to have 15 stops of "workable latitude" before its release a year and a half ago, would you honestly not have taken offense to the claim?

My point exactly. If Phil were to do these same "workable DR" tests with Alexa, what could possibly suggest that the Alexa wouldn't again surpass this 16.5 figure? Therefore I believe any educated filmmaker who looked at RED's claim of 16.5 stops of latitude objectively would come to the logical conclusion that it is a very specious claim.

I'll first point to this thread as this clarifies the difference:

http://www.cinematography.com/index.php?showtopic=51973

Workable latitude is how far you can move things without "hurting" the image quality. Image "salvagibility" (not a real word) is part of this equation. You simply can't move all of that dynamic range around under most exposure conditions and expect usable results. You'll see extreme artifacts using that logic. However, and this is a big one, if exposed properly and you are are above the noise floor on Dragon you have an very large workable latitude. Reading into more of the specs that have been published and more importantly many of the example clips and images that have been posted you can start seeing just how impressive that workable range is.

Dynamic Range however is a hard measurable number.

Typically speaking, the more Dynamic Range a system has the more freedom can be found in the workable latitude.

I'm a proud owner of the Epic and an admitted fanboy, but making claims like these is the exact reason RED gets so much hate from a lot of filmmakers. The Dragon is amazing in so many other ways that it doesn't make much marketing sense to exaggerate or aggrandize anything (and especially something as critical as dynamic range, the very figure that many choose the Alexa because of).

Thank you guys both for the long and well thought out posts. I'm not trying to troll here, I know how much work you guys put into this and in no way am I trying to discredit any of your tests or previous claims. I'm merely perturbed at this 16.5 DR claim because I believe it hurts our company's integrity (that which is already being constantly incised by all the haters out there).

I am not making a claim. I need to again be clear. I measured the system in stop increments. That is controlled data. There's actually more than 16+ stops that can be measured. However, you wouldn't use it "that way" and to me that's why Red is going with 16+ stops as that falls within common usage. It's about having access to that dynamic range and what the camera system and yourself does with it that's important.

Red posted that lovely image a long time ago of the "unwrangled" raw sensor output with the total dynamic range of the sensor, which was very broad. What we are measuring now is how they have decided to use it.

The next time I measure the systems, if the ruling overlords allow me, I will post all of the images in 1/3 or 1/4 stop increments and you can also see exactly what I am seeing and more importantly measuring.

 

[James_Mills]

Workable latitude is how far you can move things without "hurting" the image quality. Image "salvagibility" (not a real word) is part of this equation. You simply can't move all of that dynamic range around under most exposure conditions and expect usable results. You'll see extreme artifacts using that logic. However, and this is a big one, if exposed properly and you are are above the noise floor on Dragon you have an very large workable latitude. Reading into more of the specs that have been published and more importantly many of the example clips and images that have been posted you can start seeing just how impressive that workable range is.

This makes sense. We all know the level of noise that we deal with when pumping up areas in coloring, so if the Dragon as noise-free as everyone claims, I could see how it could have an advantage in total latitude against the Alexa after a few passes of power windows in post. Is this essentially the argument you're making? i'm not sure if I believe this yet, nor do I think it justifies the free usage of the claim "16.5+ stops", but I understand the argument.

I am not making a claim. I need to again be clear. I measured the system in stop increments. That is controlled data. There's actually more than 16+ stops that can be measured. However, you wouldn't use it "that way" and to me that's why Red is going with 16+ stops as that falls within common usage. It's about having access to that dynamic range and what the camera system and yourself does with it that's important.

The "claim" I was referring to in that post was not yours, but rather Red's original claims of having 16.5+ stops and "more dynamic range than 65mm film". I was never questioning any of your posts or observations, the conversation just sort of drifted your way and I apologize if it felt like I was attacking you particularly.

 

[Gunleik Groven]

I too stand by the 16+ thing as what I have measured in incremental stop tests. They have been done differently than the one Geoff did, and I can kinda see how we get different results, without seeing any flaw in the CML/UWE test.

Yup. I get the seemingly contradictory statement.
But I think that still holds true. We did it differently, thus we get different results.
The difference does not invalidate any of the results, though.

A test only gives you results for what you test for the way you test it.
And I think Geoffs is good.

+1 to all else Phil says.
He is the tech chief-of-staff...

 

[Niko Pueringer]

I'm wondering if someone could step forward and answer this question clearly;

The CML test shows about a stop in highlight tolerance in the Alexa over the Dragon. If I'm not mistaken they show the same cutoff point in the blacks. Does this mean the Alexa has one stop more of DR over the (old OLPF) Dragon?

There's a lot of discussion about DR taking place on here, but I'm having a hard time finding a clear answer. If someone more knowledgeable than me were to lay it out, that'd be awesome.

 

[Gunleik Groven]

Answer is quite simple:

If that is what you see, that is true.

Have you dl'ed the RAW files?

What I (and Phil, as I read him) say in addition to that, is that we kinda like how "wide" the practical useable DR is on the Dragon. That reflects how many stops we find useable for different uses.

From what Geoff says, it is about a stop difference in the top. To which I do not disagree, and was about what I expected to see.

 

[Phil Holland]

This makes sense. We all know the level of noise that we deal with when pumping up areas in coloring, so if the Dragon as noise-free as everyone claims, I could see how it could have an advantage in total latitude against the Alexa after a few passes of power windows in post. Is this essentially the argument you're making? i'm not sure if I believe this yet, nor do I think it justifies the free usage of the claim "16.5+ stops", but I understand the argument.

The "claim" I was referring to in that post was not yours, but rather Red's original claims of having 16.5+ stops and "more dynamic range than 65mm film". I was never questioning any of your posts or observations, the conversation just sort of drifted your way and I apologize if it felt like I was attacking you particularly.

No worries. I don't feel like it was attack, I just want to avoid confusion of terminology because a lot people read this stuff.

In terms of having more dynamic range than 65mm I can absolutely stand by that. In my years working with film I have never, ever been able to get more than 15 measured stops out of it. Never. More often than not, it's less than that.

I have a fairly boring image from from the Leica Store on a rainy shot on the Summicron-Cs with a bit of power window play

Rain? In Los Angeles?

+1 to all else Phil says.
He is the tech chief-of-staff...

Heh. Shhh. People won't think I'm cool if they know I'm a nerd.

 

[Niko Pueringer]

Thanks Gunleik! I haven't downloaded the raw files yet, as I'm more trying to figure out all the different ways people are qualifying dynamic range and latitude.

Bu the way, how do you pronounce your name?

 

[Gunleik Groven]

Hahahaha

Very Norwegian.

"Gun" is with the same vovel sound as "goon", but short. Sorta like "gun" (in english), but with the "oo" sound. Phonetikaly there is not really an equivalent English written word, as in English that sound is always "long"

"leik", is a bit like the english "like".

That has to be the OT post of the year!

 

[geoffboyle]

When using the UWE - CML tests please use the RAW files, anything else may have been affected by my work processes.

The RAW is the RAW and it's what you get from the camera.

 

[geoffboyle]

Can someone confirm this? Given the credentials behind these test takers, I'd be shocked if something as rudimentary as white balancing wasn't controlled. If that was indeed the case then I'd begin to question just how much stock I'm giving these results as well.

If we're talking about the UWE - CML tests the cameras were not white balanced, as it clearly states all cameras were set to zero tiny and 5600K where possible 5500 where that was not possible.

The light source was 5600K.

The whole point of this is to show variation in the cameras starting from a neutral point.

 

[Gunleik Groven]

If we're talking about the UWE - CML tests the cameras were not white balanced, as it clearly states all cameras were set to zero tiny and 5600K where possible 5500 where that was not possible.

The light source was 5600K.

The whole point of this is to show variation in the cameras starting from a neutral point.

And as we were shooting raw, that is the only sensible place to start IMHO.

 

[geoffboyle]

I just have to say that I fundamentally agree with Gunleik and Phil, we are looking at the same things.

I'm also looking forward to seeing Gunleiks other material as I really don't have the time to do much more, the time taken to keep everything in ACES space at original res and keep every iteration separately is taking an enormous amount of time and disk space.

I will show 4K DCP at BVE and put HD on Vimeo and that's it.