Meaning of 16.5 stops of Dynamic Range

[Rob Castiglione]

Yes, without question, the camera always needed to be put in a larger body. It would address just so many issues.

 

[Matt Ryan]

Yes, without question, the camera always needed to be put in a larger body. It would address just so many issues.

I think it needs to be 2 separate models though. A lot of uses for the smaller Epic form factor. I'd call it the Dragon and the Dragon Studio. Studio would have everything built in a similar size, form factor and layout to the F55. On board Pro Res recording to CF cards as well. Built in ND's and global shutter. 4K out via SDI and HDMI, 4 xlr ins, powered lemo for power zooms like the Cabrio 19-90. Dual Lemo's for evf and monitor to be run at same time.

But this has nothing to do with the topic of dynamic range...

 

[Rob Castiglione]

All true Matt. You have created a dream camera but you are right ... we are off topic. When you consider the matter, everything is a compromise. Arri has its own compromises.

 

[Niko Pueringer]

Yeah, if only there was something like a modular system that would let you build a bigger camera body with those options...

 

[Matt Ryan]

Yeah, if only there was something like a modular system that would let you build a bigger camera body with those options...

We don't have modules that work well for that. Plus I don't want a bazooka on my shoulder. You would need the +1, Pro IO, and 4K broadcast module. That makes the camera extremely long.

 

[Gavin Greenwalt]

It would be nice to have an option to combine multiple ISO's to create a 16 stop image, and I'm not talking about HDRX. In RCX or any program if you lower ISO you can see a couple stops worth of info come back, same for lifting the iso and seeing a stop or so in the low end come back.

Start from linear. The ISO is just a (linear image * multiply ISO factor) -> RedGamma or Cineon lut. Neither RLF (cineon) nor REDGamma are clipping any of the dynamic range. But if you start from the linear you can roll your own custom gamma curve to do precisely what you want.

I also find it interesting that a growing number of ICONIC Director's seem to be trusting/embracing RED more.

I find it interesting that people care what camera a *director* chooses unless the title of Director is followed by "of Photography". Also every single one of your examples if a director who seems to have lost their way lately. Transformers was a disaster even according to Michael Bay. The Hobbit bombed on Rotten Tomatoes for good reason. Prometheus was one of the worst (although prettiest) films in recent years. And James Cameron if anything has lost faith in cameras period, he's just having his films rendered in PRMan.

They never claimed 20 stops of native dynamic range.

Splitting hairs. They did claim 18+ that to me means "18.5". And "maybe even a little more" to me means "19". So no they didn't claim 20, they claimed 19... and then 18... and then 16-17. Whether or not the sensor is great (it probably is) doesn't dismiss the fact that Dragon seems to have gone from "certainly more than any digital sensor known to man " to "about the same as Arri's 3 year old camera" which is quite the slide from earlier hype.

DXO Mark gives it 14.8 and everything that I have seen confirms that.

DXO seems extremely conservative in all of their ratings. They put the 5D MK2 at 11.9 where as I would put it at least a stop higher. If they rated the Alexa I doubt it would be 14 stops. MX and Alexa already were pretty close with Alexa having about an extra stop. Without DXO applying whatever standard they applied to Dragon to Alexa it's not a fair comparison to take a "absolute max measureable dynamic range" and compare it to DXO's more conservative 'usable' range.

 

[David Litchfield]

Can anyone explain how a small box effects the sensors ability to read the light hitting it?

 

[Nick Morrison]

I find it interesting that people care what camera a *director* chooses unless the title of Director is followed by "of Photography". Also every single one of your examples if a director who seems to have lost their way lately. Transformers was a disaster even according to Michael Bay. The Hobbit bombed on Rotten Tomatoes for good reason. Prometheus was one of the worst (although prettiest) films in recent years. And James Cameron if anything has lost faith in cameras period, he's just having his films rendered in PRMan.

I personally thought Prometheus was fantastic. And certainly would not say either Soderburg or Fincher have "lost their way" (Behind the Candelabra and House of Cards were massive achievements). And I'm actually kind of looking forward to Transformers 4.

I also forget to add Guillermo Del Toro, who shot Pacific Rim on RED as well.

Can anyone explain how a small box effects the sensors ability to read the light hitting it?

It seems pretty clear a small box has far more challenges in regards to heat and power distribution. I have to imagine there's also less space for boards, etc.

 

[Brian Boyer]

Splitting hairs. They did claim 18+ that to me means "18.5". And "maybe even a little more" to me means "19". So no they didn't claim 20, they claimed 19... and then 18... and then 16-17. Whether or not the sensor is great (it probably is) doesn't dismiss the fact that Dragon seems to have gone from "certainly more than any digital sensor known to man " to "about the same as Arri's 3 year old camera" which is quite the slide from earlier hype.

I always bring a micrometer to a debate. But, it seems you're no stranger to splitting hairs either, my friend. You expect us to believe whenever you've said to people, "X, maybe Y," and they ignored the X value and tried to hold you to the Y value, you didn't look at them like they were crazy?

With "X, probably Y," Y is fair game . "Maybe Y" is pushing it. Besides, it depends on how you state the numbers. Even if we overlook a helping of interpretational rounding up, 19 does seem really close to 20. But, in terms of stops, 19 is only half as much as 20, which doesn't sound nearly as close.

As for Dragon's DR going from "more than any digital sensor known to man" to being about the same as Arri's 3 year old camera, I'm going to play devil's advocate (with tongue planted semi-firmly in cheek) and say maybe RED can still rightfully make that claim (I know how much weight "maybe" carries with you).

If it were discovered the first-gen iPhone from 2007 had no more battery life in real-world usage than a Motorola StarTAC from the late 90s, can one realistically claim Apple hype without considering how much more the iPhone is doing with that same battery life?

If two runners have the exact same time for the 100 yard dash but one of them is wearing 5 lb. ankle weights on each leg, can you really claim an equivalency?

In the same manner, if Dragon has the same DR as Alexa's 3 year old sensor, given it has smaller and considerably more photosites, can't that also be considered more by virtue of producing the same outcome despite considerably higher demands?

You've already established that if Dragon has just "a little more" DR than Alexa, we can round that up to the nearest half stop, which would make RED's claim of having the most more legitimate.

And Phil stands by his numbers. So far, everyone refuting them has produced nary a test of their own.

 

[Gunleik Groven]

Tom me it is obvious that we talk about different things when talking about DR.
Which kinda makes everything true. And also pretty confusing...

I deliberatedly did not comment on DR when talking about the prototype. Just because the IMAGE was so insanely cool.
And I would much rather have that fatness, than insisting on pushing every theoretical bit of the sensor into the open.

I really digged the R1 M over the R1MX, because it was not pushed. To push the MX into the RED ONE (and getting better numbers), some of the fatness of the M seemingly had to get lost.

I am glad the Dragon was not an upgrade like that.

Usually we do not shoot numbers, but look at an image we have lit for and think:
Wow, this looks good

or:
This does not look so good.

FDragon is in the good part of things. And I dig it for that.

Sure, it can be broken down in numbers, too... But that does not tell me the feel of the thing.

 

[Tim Morten]

Bottom line is that Dragon has more DR and more resolution in a smaller package for less cost than Alexa. Dragon image looks outstanding. We can quibble the numbers (wasted time without a truly objective side by side test), but there is no question in my mind that Dragon has earned its crown. What the future brings only time will tell...

 

[David Litchfield]

It seems pretty clear a small box has far more challenges in regards to heat and power distribution. I have to imagine there's also less space for boards, etc.

Well yes I could have guessed that answer. What I'm interested in is why does heat negatively effect the sensitivity of the sensor. Surely the sensor reads a photon hit or it doesn't.

And Phil stands by his numbers. So far, everyone refuting them has produced nary a test of their own.

Well i took part in the UWE test and I would dispute the numbers. And before people mention old and new OLPF that's just shifting mid grey and not affecting total DR.

It wouldn't surprise me if the frankie sensor is different to what is going in Dragon now. Perhaps yields on an 18.5 sensor would be so low they'd never get any upgrades done. Perhaps a 16.5 sensor gets them a better yield? That makes more sense to me than crippling an 18.5 stop sensor in a matchbox.

Most DPs don't really care about the resolution race. But they do care about DR and colour. So far I don't think the Dragon does enough to tempt them away from their beloved Alexa's. If they haven't already they will move to XTs before Epic. That's why, IF the sensor is in fact capable of 18.5 stops, it is utterly ridiculous to not make that the priority and put it in a bigger matchbox. In one step you could undo all the gripes the camera dept have with your product and free it to be used on all the Oscar nominated movies and leave the 3D bullshit behind

David

 

[Rohan B]

One thing that i am still a bit confused about is that what was the difference between the mx epix and an alexa in terms of dynamic range?? Even though people are now saying that alexa is being very conservative when rating it's sensor 14 stops, but i also remember seeing many side by side tests between alexa and epic and they showed just 1 full stop difference. Dragon is claimed as a sensor that has 3 more stops than an epic and this claim is backed by numerous users who have tested it, then how is it that dragon barely equals alexa or maybe just gives half a stop more than that. Then were those tests flawed or has the alexa newer versions have more dr but that does not seems right,they never did any sensor developments its the same alev iii sensor in all their bodies, i have also seen some tests where the testers claimed that alexa has 2 stops more than epic,but that would still mean that the dragon would have a full stop more than alexa having 3 more stops over mx. So did alexa had 3 more stops over the epic mx,only that would make it logical but still i don't remember any test or any one claiming that, so i don't have anyone or anything to back up my conclusions.

Basically if we have to cut it all down to two basic questions, they would be....

1.What was the difference of DR in stops between Epic(MX) and Alexa?? (How many stops did alexa truly has over Epic MX??Just numbers 1 stop? 2 stops? 3 stops?)

2.How Many stops of DR does the EPIC DRAGON( complete version with new olpf and new caliberation)has over EPIC MX??

I think its a pretty decent and straight forward question and if one of the answers contradicts the other, then the answers are wrong and someone is not being truthful or confusing the stuff up.....

Looking forward to a good and meaningful reply that clears things up once and for all...

 

[Kevin Marshall]

Well yes I could have guessed that answer. What I'm interested in is why does heat negatively effect the sensitivity of the sensor. Surely the sensor reads a photon hit or it doesn't.

Heat can increase the noise in a system. So it's entirely possible for the S/N ratio to be different in a testbed compared to a final product.

 

[Sidney L. Plaut]

One thing that i am still a bit confused about is that what was the difference between the mx epix and an alexa in terms of dynamic range?? Even though people are now saying that alexa is being very conservative when rating it's sensor 14 stops, but i also remember seeing many side by side tests between alexa and epic and they showed just 1 full stop difference. Dragon is claimed as a sensor that has 3 more stops than an epic and this claim is backed by numerous users who have tested it, then how is it that dragon barely equals alexa or maybe just gives half a stop more than that. Then were those tests flawed or has the alexa newer versions have more dr but that does not seems right,they never did any sensor developments its the same alev iii sensor in all their bodies, i have also seen some tests where the testers claimed that alexa has 2 stops more than epic,but that would still mean that the dragon would have a full stop more than alexa having 3 more stops over mx. So did alexa had 3 more stops over the epic mx,only that would make it logical but still i don't remember any test or any one claiming that, so i don't have anyone or anything to back up my conclusions.

Basically if we have to cut it all down to two basic questions, they would be....

1.What was the difference of DR in stops between Epic(MX) and Alexa?? (How many stops did alexa truly has over Epic MX??Just numbers 1 stop? 2 stops? 3 stops?)

2.How Many stops of DR does the EPIC DRAGON( complete version with new olpf and new caliberation)has over EPIC MX??

I think its a pretty decent and straight forward question and if one of the answers contradicts the other, then the answers are wrong and someone is not being truthful or confusing the stuff up.....

Looking forward to a good and meaningful reply that clears things up once and for all...

Good questions Rohan. I also a little confused.

 

[Hrvoje Simic]

Even though people are now saying that alexa is being very conservative when rating it's sensor 14 stops, but i also remember seeing many side by side tests between alexa and epic and they showed just 1 full stop difference. Dragon is claimed as a sensor that has 3 more stops than an epic and this claim is backed by numerous users who have tested it, then how is it that dragon barely equals alexa or maybe just gives half a stop more than that.

The confusion is caused by not making a distinction between maximum measured range and subjectively perceived "usable" range.

3 stop gain is likely referred to using the same approach.

 

[Michael Epple]

One thing that i am still a bit confused about is that what was the difference between the mx epix and an alexa in terms of dynamic range?? Even though people are now saying that alexa is being very conservative when rating it's sensor 14 stops, but i also remember seeing many side by side tests between alexa and epic and they showed just 1 full stop difference. Dragon is claimed as a sensor that has 3 more stops than an epic and this claim is backed by numerous users who have tested it, then how is it that dragon barely equals alexa or maybe just gives half a stop more than that. Then were those tests flawed or has the alexa newer versions have more dr but that does not seems right,they never did any sensor developments its the same alev iii sensor in all their bodies, i have also seen some tests where the testers claimed that alexa has 2 stops more than epic,but that would still mean that the dragon would have a full stop more than alexa having 3 more stops over mx. So did alexa had 3 more stops over the epic mx,only that would make it logical but still i don't remember any test or any one claiming that, so i don't have anyone or anything to back up my conclusions.

Basically if we have to cut it all down to two basic questions, they would be....

1.What was the difference of DR in stops between Epic(MX) and Alexa?? (How many stops did alexa truly has over Epic MX??Just numbers 1 stop? 2 stops? 3 stops?)

2.How Many stops of DR does the EPIC DRAGON( complete version with new olpf and new caliberation)has over EPIC MX??

I think its a pretty decent and straight forward question and if one of the answers contradicts the other, then the answers are wrong and someone is not being truthful or confusing the stuff up.....

Looking forward to a good and meaningful reply that clears things up once and for all...

For one thing, I don't know anyone who has recently stated 3 stops more than MX and I definitely haven't seen any footage or comparison's that show more than a stop improvement on latitude over the MX and I also haven't seen anything that makes me believe that the Dragon has a lower noise floor, ie ISO. I have heard all the claims yet seen none of the comparisons. Nobody was denying that Dragon is better than MX, we're just trying to figure out how much. It doesn't matter much now, since I'll finally be able to check it out for myself at the local rental house which is finally getting theirs.

 

[Jan Reiff]

we did a 1:1 comparison today with my Dragon and a MX Epic ...

Highlights: Dragon has much more dynamic in highlights. Much and visible smoother Roll Off without (!) softening the image in general
Shadows: Dragon is much more clear, i don´t want to say cleaner (there is noise, but its finer and more like filmgrain), but clearer, sharper, better colors, when it´s "underexposed"
Colors: Richer, brighter overall. Overexposed white stays true color with the Dragon, where MX has its colorshifts.
Skintones on Dragon are perfect out of box
Dragon has ADD - and ADD is bit "magic" ...

Sum of today:
Dragon is worth every cent to do the upgrade in comparison to the MX.

(We had bit too less sun today, not sure about the IR / ND with Dragon, testing sunday ...)

 

[Bob Gundu]

Thanks for making it harder to wait Jan! :smile:

we did a 1:1 comparison today with my Dragon and a MX Epic ...

Highlights: Dragon has much more dynamic in highlights. Much and visible smoother Roll Off without (!) softening the image in general
Shadows: Dragon is much more clear, i don´t want to say cleaner (there is noise, but its finer and more like filmgrain), but clearer, sharper, better colors, when it´s "underexposed"
Colors: Richer, brighter overall. Overexposed white stays true color with the Dragon, where MX has its colorshifts.
Skintones on Dragon are perfect out of box
Dragon has ADD - and ADD is bit "magic" ...

Sum of today:
Dragon is worth every cent to do the upgrade in comparison to the MX.

(We had bit too less sun today, not sure about the IR / ND with Dragon, testing sunday ...)

 

[Gunleik Groven]

For one thing, I don't know anyone who has recently stated 3 stops more than MX and I definitely haven't seen any footage or comparison's that show more than a stop improvement on latitude over the MX and I also haven't seen anything that makes me believe that the Dragon has a lower noise floor, ie ISO. I have heard all the claims yet seen none of the comparisons. Nobody was denying that Dragon is better than MX, we're just trying to figure out how much. It doesn't matter much now, since I'll finally be able to check it out for myself at the local rental house which is finally getting theirs.

Well, I have (on the old OLPF/old calibration).
Have not checked the new, but I would be wildly surprised if it is not "more" in that.