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

[Gunleik Groven]

i have four hunches....

dragon is excellent as it is.
it will be even wilder with the new olpf and colorscience.

new colorscience will also affect existing footage.

re-calibration is done to avoid stupid gut-reactions.

but these are all wild speculations. chill down. to me there were bigger surprises than the dragon vs alexa on the test. that was all pretty predictable...

 

[Matt Ryan]

Nope.

Geoffs analysis will be that, as I guess mine will be.
It is kinda hard to show what we see, without making that "permanent"...

The underlying raw files will be released for everyone to play with

If Geoff does not, I can. I will probably not reeae it ALL, but the extremes and the middles. We shot A LOT of cameras at A LOT of exposures...

Good to know. Geoff kept mentioning TIFF files so it's good to know RAW will be available.

No one has even criticized anything yet... Geoff seems to like Dragon color and the Dragon as a whole which is a step in the right direction. Hearing the way he assumes Dragon will get better is an even bigger step forward. I think his initial response towards RED was great considering his past comments. Looking forward to the test results.

 

[Sulekh Suman]

It is sort of redundant to pit Dragon against Alexa which is almost a 4 year old sensor with substantially less resolution. F65 sensor is it's closest competitor which was missing from this test. I guess one test won't serve all the curiosity.

 

[luigivaltulini]

to me there were bigger surprises than the dragon vs alexa on the test. that was all pretty predictable...

interesting ..
I'm curious to see ..

and

Gunleik a curiosity,
why we're seeing here on Dragon about 14 stops on reduser .. and in this test is below 1 stop from Alexa ..
According to what we have seen we should be equal ...
Why Alexa have 1 stop advantage ...?
I'm just trying to understand ...

 

[Gunleik Groven]

AFAIK I have not commented on difference in DR. At all.

Nick quoted Geoffs initial remarks. Those are Geoffs initial findings.

What I am not surprised about is that Alexa has more highlight tolerance than Dragon currently, as I was pretty sure about the difference between MX and Alexa, and could see how much more the Dragon handles than MX.
But I have not thrown around any numbers I am aware of yet.

What I know is that Geoff and I will be looking at the files in slightly different ways. If that will lead to similar or different results... I have no idea.

I will be getting the OLPF upgrade anyways.
But like I thinkk "k's" is a very singular way to measure a camera system, I think it is stupid to get hung up on other single parameters.

STanley will be running happily, with or without new OLPF. It is a great cam in any comparison, that I am not less confident about after visiting UK...

 

[Matt Ryan]

One question that keeps popping up in my head is how do tests that Phil and John calculate Dragon has 16.3 stops but in a comparison it appears to have slightly less than Alexa which has 14 stops DR? I know the test hasn't been posted yet but this is more of just a general question. I can't wrap my head around it..

 

[Gunleik Groven]

It comes down to methodology.

In my "Espen" test, I also came out with more than 16 stops. But I am not surprised that the Alexa manages a bit more in the top in the dragons current state.

I know that Geoff and I will be looking at the highlights and lowlights with diferent approaches.
And I also know that the Bristol test was shot side-by side with the same settings on all cams where that was sensible.

I have not yet looked into the material, as I can not until I get home.

Thing is that even if I should find "more" than him (not saying I will. This is the first shoot I have been involved with with the current firmware), it does not really matter too much.
Because his approach to the development of the images is as valid as mine.

And at least as common.

so the results we get, even if they differ, are both valid.

That is the beauty of releasing the raw files along with the tests.

As Graeme has stated, the new colorscience, calibration and olpf invalidates the test. I would say: To some extent.
As I think the Dragon stands up as it is.
And whatever extras we get, are only "even more".

I think tests like these have some value, independently of whether you, phil, geoff and I come up with four different results.
Because if that would be the case, it would be a good illustration of the workflow issues David and others have been talking about.

Approach has a lot of influence on the result.

As much as any other factor.

The difference in the images you and I developed from Luigis first test, is a pretty good illustration of that....

(it is almost tempting to repost those...)

 

[Mike P.]

Matt, I think that's a job for DragonGamma.

I never understood why RG3 didn't just map stops above and below base exposure more consistently (in the same way that Alexa does its ~7up and ~7down at baseISO800). Seeing as Dragon has so much more DR than Alexa, it's safe to say that whatever Alexa has been doing right, it's been doing really right. Now that Dragon maintains DR across ISO, methinks that's exactly what we'll see happen when the new science is out...

…It's going to be really frustrating if RED is waiting until NAB to show off exactly what Dragon can do with new Colour (and Gamma) science.

 

[Matt Ryan]

Matt, I think that's a job for DragonGamma.

I never understood why RG3 didn't just map stops above and below base exposure more consistently (in the same way that Alexa does its ~7up and ~7down at baseISO800). Seeing as Dragon has so much more DR than Alexa, it's safe to say that whatever Alexa has been doing right, it's been doing really right. Now that Dragon maintains DR across ISO, methinks that's exactly what we'll see happen when the new science is out...

…It's going to be really frustrating if RED is waiting until NAB to show off exactly what Dragon can do with new Colour (and Gamma) science.

Results are viewed with Redlogfilm for tests sake, so DragonGamma will have little to do with it other than give a better viewing gamma on set. Now, a new Dragon Log Film could make a real difference.

 

[Mike P.]

Good point.

The funny thing is that REDlogFilm is supposed to be the same as Cineon, whereas LogC is basically a made-up Log format by Arri (that apparently makes the best use of its DR). The irony is that, while it may not be "correct" it certainly places DR emphasis in the right spots directly out of the camera.

Something tells me whatever camera/sensor plays nicest/best/easiest with ACES, will be the camera most used in the next few years.

 

[Rohan B]

ain't dragon was supposed to have 2 stops more dr than the alexa?? keeping in mind that alexa has "just" 1 full stop more than "mx" and the dragon "full" 3 stops more than that, now another stop added with the new caliberation and olpf. So a total of 4 stops over mx and 3 stops over Alexa. Dragon was marketed as a camera that will have more DR than any other digital camera, way more than film actually, and now it is still a stop behind Alexa?? Its still trying to just "match" Alexa with the new olpf and new caliberation ??? Common this camera was to blow Alexa off!!!! Epic has a stop less than alexa and now dragon "ALSO" has a stop less than alexa?? WHAT THE "FUCK" IS GOING ON???????????

 

[Rohan B]

I know that all these number are subjective, 14 stops 15 stops 16 stops, but, the difference between two cameras is usually accurate, so if epic had a stop less than alexa then something that measures 3 or 4 stops over mx, should have at least 2 to 3 stops more than alexa, how the hell can it be a stop behind it??????

 

[Rohan B]

than what was the actual difference between an epic and alexa that even after making a new sensor that has 3 stops more than the mx sensor,its still a stops behind alexa???or was there something wrongly done in the test?? even without the new olpf and new caliberation it supposed to have 2 stops more than alexa and people are still saying that maybe it will just match alexa with the new olpf?? And the only quality it has over the alexa is that its a bit quieter/less noisier??????

 

[Phil Holland]

ain't dragon was supposed to have 2 stops more dr than the alexa?? keeping in mind that alexa has "just" 1 full stop more than "mx" and the dragon "full" 3 stops more than that, now another stop added with the new caliberation and olpf. So a total of 4 stops over mx and 3 stops over Alexa. Dragon was marketed as a camera that will have more DR than any other digital camera, way more than film actually, and now it is still a stop behind Alexa?? Its still trying to just "match" Alexa with the new olpf and new caliberation ??? Common this camera was to blow Alexa off!!!! Epic has a stop less than alexa and now dragon "ALSO" has a stop less than alexa?? WHAT THE "FUCK" IS GOING ON???????????

I need to be clear on this subject, because comments like this are what travel through the ears of people looking for reasons and obviously misinformation is already happening even before results have been published. This camera comparison has been done already and will be done again with the New OLPF and New Calibration, the CML test did not use these.

In relation to what what was said with these CML tests is that Dragon had one less stop in the highlights. That has nothing to do with the total captured dynamic range. That does have everything to do with where the previous calibration placed captured dynamic range based on a chosen exposure. This has now changed as of last week with approximately a 1.3 stop improvement in highlight performance. An improvement that this test will not show. The OLPF also plays a role here in terms of tonal "purity".

A few things.

- Dragon does have more captured and usable Dynamic Range than Alexa.
- Dragon does have more captured and usable Dynamic Range than Film.
- Dragon isn't trying to "match" the Alexa with the OLPF and Calibration.


Remember, don't assume the Dynamic Range increase is "all up top". It's everywhere. Shadows, Midtones, Highlights. In actual use how you expose to maximize this potential is important. The new calibration to me personally puts 18% gray at a more pleasing place across the captured range and also you will see improvements in the highlights and shadows due to this.

Give Geoff some time to generate his thoughts and generate his images. He seems to be approaching this with an open mind and knows of the New OLPF and Calibration. Which is more or less why he stated "I expect that pretty soon the Dragon will match and maybe surpass the Alexa for DR." Also of importance, "As far as colour goes I'm still checking but I'm really impressed with the Dragon so far."

 

[Nick Morrison]

Just to clarify, did Geoff test Dragon at 5KHD, or 6k?

 

[Mike P.]

I think Gunliek said 5.5KHD for lens compatibility.

 

[Rohan B]

I need to be clear on this subject, because comments like this are what travel through the ears of people looking for reasons and obviously misinformation is already happening even before results have been published. This camera comparison has been done already and will be done again with the New OLPF and New Calibration, the CML test did not use these.

In relation to what what was said with these CML tests is that Dragon had one less stop in the highlights. That has nothing to do with the total captured dynamic range. That does have everything to do with where the previous calibration placed captured dynamic range based on a chosen exposure. This has now changed as of last week with approximately a 1.3 stop improvement in highlight performance. An improvement that this test will not show. The OLPF also plays a role here in terms of tonal "purity".

A few things.

- Dragon does have more captured and usable Dynamic Range than Alexa.
- Dragon does have more captured and usable Dynamic Range than Film.
- Dragon isn't trying to "match" the Alexa with the OLPF and Calibration.


Remember, don't assume the Dynamic Range increase is "all up top". It's everywhere. Shadows, Midtones, Highlights. In actual use how you expose to maximize this potential is important. The new calibration to me personally puts 18% gray at a more pleasing place across the captured range and also you will see improvements in the highlights and shadows due to this.

Give Geoff some time to generate his thoughts and generate his images. He seems to be approaching this with an open mind and knows of the New OLPF and Calibration. Which is more or less why he stated "I expect that pretty soon the Dragon will match and maybe surpass the Alexa for DR." Also of importance, "As far as colour goes I'm still checking but I'm really impressed with the Dragon so far."

Thanks For the reply Phil!!!! Your Posts are always so informative and enlightening.....

 

[Phil Holland]

Just to clarify, did Geoff test Dragon at 5KHD, or 6k?

I think Gunliek said 5.5KHD for lens compatibility.

They apparently shot in 5K HD. Which I would imagine was done to emulate an equivalent FOV between the cameras they tested. It was stated they tested for UHD/1080 delivery.

I do hope they shot at a reasonably low REDCODE compression ratio.


This also should point out two things to people who look at this test. One, Red Epic Dragon is a multi-format camera and it's up to you how you use those larger formats and resolutions. Two, there are benefits both in perceptual noise/texture and also resolved detail finishing from higher resolutions.

Personally, I think they should have tested 6K HD and 5K HD so they can actually see the benefits and differences there. But clearly this is a vanilla S35 digital cinema camera test.

Thanks For the reply Phil!!!! Your Posts are always so informative and enlightening.....

Thank you Rohan. I have been fascinated by a few things over the years relating to Reds public image and perception. More or less in conversations with clients and shooters a lot of whom have formed extremely strong opinions which have delivered via word of mouth or by the internet largely from people who don't even shoot with these caliber cameras. However, there are valid criticisms and Red has clearly worked hard on Dragon to clear that murky water and also pave some ground on new fronts.

It's funny because there likely two other things that will effect the results from a test like this that Red is working on down the line. That points out one thing that they do that others do not. They update and enhance our cameras as they are able to integrate new features or develop new technology. And they aren't charging for that, which is a big thing for me.

I have a slight bias from an owner perspective, but more or less I'm fixated on what I can do with all of these cameras because that's part of the job I do in real life. Red is on the path that I can get behind. I have a love ARRI gear too. Some of it is designed and built for lifetime use, many shooters here would agree. I'm very curious to see what they have up their sleeves with their new camera even.

Right now Dragon sits atop a very small mountain of digital cinema cameras that can capture impressive images. However, to get the absolute most out of any of these cameras you really need to look deep.

 

[Gunleik Groven]

5.5k HD, lowest available compression

Reasonnfg is one I follow.

1. It is S35
2. It is 16:9
3. (like 4kHD was on the MX) is probably the new Dragon/TV format. (I am talking abut series/drame etc, not neccesarily commercials)

Fore res-test, 6k was shot


Point of te test is UHD (16:9) and 1080 delivery

 

[Phil Holland]

5.5k HD, lowest available compression

That's good news. I just was going by this:

From Geoff:

For the dynamic range sequences we shot everything at as close to a S35 image size as possible, this meant that the Dragon was shot at 5K HD and that the KineRAW had a smaller sensor area than the others as it appears to be the same size as a RED at 4K HD.