Meaning of 16.5 stops of Dynamic Range

[Ivan Kovax]

Show us. I don't mean that as a challenge. We just keep hearing people make these claims with no visual proof to back it up. And I don't mean this to sound harsh, but I personally don't want to hear what people think. I want to see what people know. I understand that people are excited and happy with their cameras. But excitement does not equal facts. Because all of the images and tests that have been posted so far have been contrary to these claims. And the testers themselves have stated the same. The Alexa still has more DR than Dragon.

And once again, I'll try to reiterate... No one is debating any of the other merits of the Dragon sensor. I'm sure that it's great in many ways. Or how may stops one needs, or whether color is more important than DR or anything else. This thread is about the 16.5+ stops claim. That's it. Just that number. And you even just posted yourself that you would rate it at slightly under 15-15.5 stops useable. That's still not 16.5+. And based on other tests it has less than the 14 stops claimed by Arri for the Alexa. So, like i've been saying, either Red needs to be as conservative as Arri and rate the Dragon at 13.5+ stops, or Arri needs to be as generous as Red and Rate the Alexa at 17.5+ stops.

You do realise that RED was clear every time they posted figures, and that they were rather clear on the original postings that they were lab figures. Jim Jannard went as far as to say that "whatever you rate the MX as, the dragon is 3-3.5 stops higher than that". They were know that these things people have different ways to measure (this is sadly true).

DXOMARK (from memory) rated dragon 14.7 stops. Which is more than film, practically, and the alexa too. So... Is that proof enough, or are you only focusing on the poor measurements that support your argument?

And this is where the issue lies: I rated the camera this way according to what I personally find usable. Others have their own method, and measure in different ways for specific reasons.

This is honestly a stupid argument simply for the reason that red predicted it would occur and augmented it with what I said above - 3 stops better. What is on the website is simply marketing, much like what is on arri's website (and yes, arri isn't entirely honest about some things on their website too... I have seen plenty of places rating the alexa at 13-13.5....).

So where does that leave us? We should all be out shooting movies, commercials and documentaries and not giving a damn about a stop or so, especially when all this tech we are talking about is matching or even BETTERING the DR of film. Am I the only one who finds this incredulous?

You know, Paris, zacuto will do one of their blind tests again this year. If it will satisfy your and everybody elses pixel peeping desire, then wait till then and let this dog lie and until somebody goes out and tests the crap out of all the cameras and gets an audience of people to judge it.

Or you could keep shaking your fist at red for improper marketing... whichever way, the choice is yours.

 

[Gunleik Groven]

Paris, on the old OLPF, old calibration, I had 1+ stops more in the top and 2+ stops more in the bottom on Dragon than Epic.

This is supposed to have improved with the new OLPF/Calibration.

Having had the chance to use the Dragon from late prototype and see the advances first hand, I would be surprised if that did not hold true.
Not least after seeing the latest colorscience..

But I am not sure that counts. As I guess I am a fanboy for seeing what RED claims

 

[Paris Remillard]

I'm not shaking my fist and I'm not trying to pick any fights here. If it seems that way, then I apologize. I'm simply going by the tests that I've seen, and by the words that Gunleik spoke himself. And several others. That Alexa had more on both the top and the bottom than Dragon.

I just find it frustrating to try to have an objective discussion on this forum when two posts after factual data is presented, someone contradicts it with speculation or opinion. And these opinions are often presented as facts. These threads tend to go in circles for this reason.

Not pointing fingers at anyone.

I just don't know why we can't agree that Dragon does not have 2.5+ stops more DR than Alexa. Which we've all been calling a 14 stop camera for a few years now.

It's been stated by various testers and shown in various tests.

 

[Phil Holland]

I just don't know why we can't agree that Dragon does not have 2.5+ stops more DR than Alexa. Which we've all been calling a 14 stop camera for a few years now.

It's been stated by various testers and shown in various tests.

Mostly because the tests presented that this "guesstimate" was made off of are flawed:

- Observing DR through any sort of gamma curve isn't exactly the best way to measure things.
- Previous side by sides referenced are literally not a precise way to measure dynamic range among camera systems.
- Using the older OLPF/Calibration (which makes a rather big difference here especially considering where tonal range falls now)

All of these tests need to be performed again, my own measurements included. Digging into the REDCODE RAW you will find more than 16 stops in reality, however you may not want such a flat image. How you use the total sum of that Dynamic Range is up to you.

I might measure it again soon against other systems. Just depends on free time really.

I'll stress hard you need to test the entire range of each camera system in terms of usable ISO to understand fully and truly about each platform. Then you can derive your ISO versus ISO on each system as well.

 

[Ivan Kovax]

Paris, on the old OLPF, old calibration, I had 1+ stops more in the top and 2+ stops more in the bottom on Dragon than Epic.

This is supposed to have improved with the new OLPF/Calibration.

Having had the chance to use the Dragon from late prototype and see the advances first hand, I would be surprised if that did not hold true.
Not least after seeing the latest colorscience..

But I am not sure that counts. As I guess I am a fanboy for seeing what RED claims

I get the exact same. I rated the old Epic MX at slightly less than 12 (same as zacuto did). New Dragon? 2 extra stops below. 1 above. Maybe slightly more than one... I never had the old olpf so I have no idea as to the before and after of that one.

And what did you actually say before Gunleik that everybody is yammering on about? People keep saying stuff about "Gunleik rating this or that sensor better/worse than... blah blah" .. .I have gone through this thread and can't work it out... maybe I am not reading hard enough...

 

[Gunleik Groven]

Not sure I have really said much else than the above.

I deliberately did not comment at all on DR on the proto.

I tested the Epic and Dragon over 15 stops in a pretty wide DR scene.

I also took part in the Bristol test, and have stated many times, that "nothing was done to harm any cameras involved".

And lastly, I think there is no real conflict between the DR statements, and that may seem strange, but it is sync with my general idea of testing:

You get results that are true for the test you did the way you did it.

Phil and John have technically measured the sensor.
Side-by sides have some relevance to me. Thus I have done some.

DxO was another tech measurement, but with a different methodology. The results when compared to other cameras being tested with the same methodology, still gave Dragon a decent outcome, would you not say?

If you need to test for something else, well, then do it.

There will be areas (on Dragon) particularily in the lower end that can be measured, but cannot neccesarily be used for much else than "richness".

What kinda surprises me a bit when people are attacking cameras for "less DR than", or "it does not see in the dark" or "it does not tolerate enough highlights", is that there seems to be some basic misconceptions on DR and how it works, and even more so, as Sergio is hinting at: Controlling light for the shot you want, or make the light you have useful for the shot you have to take.

Now... Shoot reborn Dragon!

Cheers!

 

[Paris Remillard]

I'm trying to convey that I'm not a pixel peeper or a flag waver or "yammering on" about things, but somehow Ivan keeps attacking me and misconstruing my intentions. If there's some way that I can show that I'm just trying relay information in relation to the OP and keep this on topic, please tell me. In all honesty.

Phil, I appreciate your continued politeness and tact. I'm not disagreeing with anything anyone is saying on these other points.

The OP was asking what the 16.5+ stop claim means when specifically the Hannover tests show the Alexa to have more on the top end and more on the bottom end. As I posted before, Adrian Jebef took the raw frames and tested them against each other and said that, yes, the Alexa had more range on both ends. Then Gunleik, who performed that test reiterated the same. No one is "yammering on" about anything Ivan. I'm simply quoting what two people said about those test results. DON'T SHOOT THE MESSENGER.

And I'm not quoting them to say this camera is better than that or out of some kind of rage that I can't get 16.5 stops. I personally don't really care. I've shot all formats, from mini DV to 8x10 Polaroids and literally everything in between. And I've been very happy with both my Red One and my Epic. I actually appreciate the limitations of different sensors and film stocks and know how to expose for what I want. But that's not the topic of this thread. I'm simply trying to answer the OPs question, which should be the point of this thread. Not to talk about the other merits of the camera. Doing so if off topic.

So to stay on topic, I'm trying to relay the information that has been gathered from a few threads on these discussions and tell the OP, no, the Dragon, at this point, does not have 2.5+ stops more DR than these cameras that we've come to know as 14 stop cameras. So , yes, there is a flaw in the way that the DR of all of these cameras is labeled.

That's all that we should be discussing in this thread based on the original post, in my opinion. All other discussions about the other merits of the camera can be had on numerous other threads.

And I really, genuinely want to be able to present this without being confrontational or rude. It's so hard to convey intent and tone when writing on a forum like this. Particularly when people's views are so passionate. But I find your aggressive and dismissive tone to be unproductive to the discussion Ivan. I really don't mean that as an attack. I just want to be able to have objective discussions about these topics. And I personally feel like the manner in which we communicate can either make that possible or impossible.

 

[Ivan Kovax]

I'm trying to convey that I'm not a pixel peeper or a flag waver or "yammering on" about things, but somehow Ivan keeps attacking me and misconstruing my intentions. If there's some way that I can show that I'm just trying relay information in relation to the OP and keep this on topic, please tell me. In all honesty.

Phil, I appreciate your continued politeness and tact. I'm not disagreeing with anything anyone is saying on these other points.

The OP was asking what the 16.5+ stop claim means when specifically the Hannover tests show the Alexa to have more on the top end and more on the bottom end. As I posted before, Adrian Jebef took the raw frames and tested them against each other and said that, yes, the Alexa had more range on both ends. Then Gunleik, who performed that test reiterated the same. No one is "yammering on" about anything Ivan. I'm simply quoting what two people said about those test results. DON'T SHOOT THE MESSENGER.

And I'm not quoting them to say this camera is better than that or out of some kind of rage that I can't get 16.5 stops. I personally don't really care. I've shot all formats, from mini DV to 8x10 Polaroids and literally everything in between. And I've been very happy with both my Red One and my Epic. I actually appreciate the limitations of different sensors and film stocks and know how to expose for what I want. But that's not the topic of this thread. I'm simply trying to answer the OPs question, which should be the point of this thread. Not to talk about the other merits of the camera. Doing so if off topic.

So to stay on topic, I'm trying to relay the information that has been gathered from a few threads on these discussions and tell the OP, no, the Dragon, at this point, does not have 2.5+ stops more DR than these cameras that we've come to know as 14 stop cameras. So , yes, there is a flaw in the way that the DR of all of these cameras is labeled.

That's all that we should be discussing in this thread based on the original post, in my opinion. All other discussions about the other merits of the camera can be had on numerous other threads.

And I really, genuinely want to be able to present this without being confrontational or rude. It's so hard to convey intent and tone when writing on a forum like this. Particularly when people's views are so passionate. But I find your aggressive and dismissive tone to be unproductive to the discussion Ivan. I really don't mean that as an attack. I just want to be able to have objective discussions about these topics. And I personally feel like the manner in which we communicate can either make that possible or impossible.

Paris. Red has already answered your questions. They have rated the dragon sensor 3-3.5 greater than the MX. SO whatever anybody rated MX, they do the math. They use 16.5 because that is the measurement they use. That is the only point I really was making.

You keep bringing up Gunleik. Not to drag him into this, but it seems I've missed something. I still can't find the page where Gunleik made that statement.

If I am being dismissive, or assertive, I apologise - but I really find the negative spin on this whole thing rather wasteful and pointless. Especially when in practice, I have seen quite different things.

 

[Mike P.]

I recently did tests "in the lab" between F55, C500, BM4k, Alexa, and Dragon. The differences in DR between Alexa and Dragon were minimal (believe it or not, I'd give the edge to Alexa)... Also, I hold my own personal experience above what others tell me (which is comparatively anecdotal).

That said, Alexa is still ~$80k, whereas Dragon is ~$25k... And that's excluding the huge resolution advantages of Dragon. I think the real winner here is Scarlet-Dragon, which could essentially be everything that Alexa is (with massive resolution and noise-freeness advantages) for a mere $14.5k.

 

[Phil Holland]

I'm trying to convey that I'm not a pixel peeper or a flag waver or "yammering on" about things, but somehow Ivan keeps attacking me and misconstruing my intentions. If there's some way that I can show that I'm just trying relay information in relation to the OP and keep this on topic, please tell me. In all honesty.

Phil, I appreciate your continued politeness and tact. I'm not disagreeing with anything anyone is saying on these other points.

The OP was asking what the 16.5+ stop claim means when specifically the Hannover tests show the Alexa to have more on the top end and more on the bottom end. As I posted before, Adrian Jebef took the raw frames and tested them against each other and said that, yes, the Alexa had more range on both ends. Then Gunleik, who performed that test reiterated the same. No one is "yammering on" about anything Ivan. I'm simply quoting what two people said about those test results. DON'T SHOOT THE MESSENGER.

Hopefully no attacks on the horizon from anybody here.

All I'm saying is the Hannover tests are extremely not relevant to the current camera people are getting when they order Dragon today. Even Gunleik's camera has the new tech now and it didn't for that test. The test more or less was "rushed", which is fine. I get the excitement and potential behind such a test, however I reiterate that's no way measure DR, though it works as a side by side comparison at those specific camera settings and gamma curve.

More or less my personal frustration stems from the concept that the camera tested then and the the Dragon that is "now" are different things. This is something a great deal of people in the outside world don't exactly understand because tests online exist not showing the new tech and differences between the two. Well, minus my thread with the info and differences that is.

On that note. I will say everybody who has seen frames off of the newer Dragon tech has been very impressed by the color and DR, which delivers what Red has stated and has been measured. Which is why Dragons are out in the wild working more right this second. Which is why I'm choosing to own and shoot with it. I'm very, very familiar with the images from Dragon's main two competitors in terms of application. It's just a matter of time for folks to see and unlock themselves what a few of us already know.

The bigger question for everybody overall is do you like what you see?

 

[Robert Ruffo New]

I recently did tests "in the lab" between F55, C500, BM4k, Alexa, and Dragon. The differences in DR between Alexa and Dragon were minimal (believe it or not, I'd give the edge to Alexa)... Also, I hold my own personal experience above what others tell me (which is comparatively anecdotal).

That said, Alexa is still ~$80k, whereas Dragon is ~$25k... And that's excluding the huge resolution advantages of Dragon. I think the real winner here is Scarlet-Dragon, which could essentially be everything that Alexa is (with massive resolution and noise-freeness advantages) for a mere $14.5k.

When you did your tests, did you use a single ISO and Redgamma 3 like 99% of tests I have seen?

 

[Nick Morrison]

I think the real winner here is Scarlet-Dragon, which could essentially be everything that Alexa is (with massive resolution and noise-freeness advantages) for a mere $14.5k.

Totally agree about Dragon Scarlet....

 

[luigivaltulini]

I recently did tests "in the lab" between F55, C500, BM4k, Alexa, and Dragon. The differences in DR between Alexa and Dragon were minimal (believe it or not, I'd give the edge to Alexa)... Also, I hold my own personal experience above what others tell me (which is comparatively anecdotal).

That said, Alexa is still ~$80k, whereas Dragon is ~$25k... And that's excluding the huge resolution advantages of Dragon. I think the real winner here is Scarlet-Dragon, which could essentially be everything that Alexa is (with massive resolution and noise-freeness advantages) for a mere $14.5k.

I agree with mike.p regarding the DR, when I made the test with the old OLPF I had the same result.
Now I have not tried with the new OLPF and the new calibration, surely we have something more than usable.
A good point as Phil says, you should try all the iso to see what happens, and even agree on his point of view and thought, maybe these tests will ruin ... once in front of the screen and see DRagon forget everything and you just want to shoot

Paris, I understand your thinking on the DR is just to understand how they are measured by various companies, but I do not understand why you are so insistent.
Now you know that + - Alexa and Dragon are very close to DR not you buy Dragon?

Guys I think now the only thought is what RED face quickly distribute updates and more cameras possible in a short time so we have more people that I can try this new sensor

 

[Dunkan Wolf]

I agree with mike.p regarding the DR, when I made the test with the old OLPF I had the same result.
Now I have not tried with the new OLPF and the new calibration, surely we have something more than usable.
A good point as Phil says, you should try all the iso to see what happens, and even agree on his point of view and thought, maybe these tests will ruin ... once in front of the screen and see DRagon forget everything and you just want to shoot

Paris, I understand your thinking on the DR is just to understand how they are measured by various companies, but I do not understand why you are so insistent.
Now you know that + - Alexa and Dragon are very close to DR not you buy Dragon?

Guys I think now the only thought is what RED face quickly distribute updates and more cameras possible in a short time so we have more people that I can try this new sensor

But how can red claim 16+ stops when 14 stops alexa shows better DR?

 

[Nick Morrison]

But how can red claim 16+ stops when 14 stops alexa shows better DR?

Maybe its visible vs usable? There appear to be many ways to measure DR. Not sure why this 16 number is becoming so important. It seems to matter more than the camera itself. Whatever the true meaning of it, is pretty clear Dragon gives you Alexa like performance with MUCH LESS NOISE, in a much SMALLER PACKAGE, at much HIGHER RESOLUTION, and at a much LOWER PRICE. Aren't those numbers, at this point, more important?

Oh yeah, and DRAGON COLOR is awesome too.

 

[Gunleik Groven]

Maybe its visible vs usable? There appear to be many ways to measure DR. Not sure why this 16 number s is becoming so important. It seems to matter more than the camera itself. Whatever the true meaning of it, is pretty clear Dragon gives you Alexa like performance with MUCH LESS NOISE, in a much SMALLER PACKAGE, at much HIGHER RESOLUTION, and at a much LOWER PRICE. Aren't those numbers, at this point, more important?

Agreed, and IMHO the colors in the Dragon is much better.
And THAT used to be the main Alexa domain IMHO.

 

[Nick Morrison]

Agreed, and IMHO the colors in the Dragon is much better.
And THAT used to be the main Alexa domain IMHO.

Amended post. Forgot about that!

 

[luigivaltulini]

But how can red claim 16+ stops when 14 stops alexa shows better DR?

For now, from what I've seen in tests with the new OLPF we DX0 14.8 usable, if we think that the old man was about 13.6 is a small but welcome gain.
So Alexa has 14 stops and Dragon 14.8.
I do not think he has less Dragon!
The DR is not only in the extreme white or in the shadows, but which one is the right test!
Hard to say ...

Well for now Dragon visually with my own eyes is better than Alexa in the most important things, this is what you notice right away

 

[Gunleik Groven]

Yup. That is what I said first time I had the Prototype and what Alexa had up until the Dragon:
I do not give a fuck what you can do in post, it "just looks better" out of the box.
And THEN comes all the post fun.

Whether it is DR, colors, pixels, colorscience or whatever that makes it that way... Honestly. I do not give a shit

 

[Mark Wuerthner]

Paris...I feel your pain. I think the problem is where you'e starting from. This is a Red forum, owned and run by them, with some of the purposes of this forum being marketing and sales. And why not? It's their thing, they can do what they want, and if others want to join in on that, that's fine. There is no disclaimer at the top of the page that says all info here is scientifically, absolutely verifiably true. This place is an opinion forum, with all opinions welcome (theoretically). So if people want to claim 16.5 stops of dynamic range, let them. Go out and look for your self and then form your own opinion. I happen to be in agreement with you in that I can't see how dragon has more dynamic range than alexa. However, when one states an opinion like that here it is quickly met with people saying this is not how to test for dynamic range, the tests are flawed, don't know how to process correctly, not the new olph (although the claims were made in regard to old olph), dragon 1/3 cost? 1/3 weight?, etc. Let it go. You'e not going to change their opinion. Just look at the test results out there and make your own decision. Overexpose latitude? Look at the overexposures that are attached:



Both of the images were taken at +6.5 stops overexposure. Underexposure latitude:



Both shot taken at -6 stops under.
Total dynamic range: (see attached files)

the forum won't let me attach the alexa sekonic profile so you'll have to take my word on that one. Both the CML tests and the Sekonic meter tests show the same thing. Dragon has +6.5 above mid gray, -7 below (at that point you're into the noise floor. There are others on here that will argue that there's more below but my "opinion" is that in any realistic grade there will be nothing there.) Alexa has +7.5 above and -7 below. So, with this info, you can come to your own conclusion. It may not match a lot of other forum member's opinions but so what. There's a lot of agenda going on in this forum. Hell, that's what makes it so fun to read!:redface: