Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Dragon Dynamic Range Test - 21 stop chart

I really hope that 4 years from now we're not still saying whoa that camera doesn't really
have 19 stops of DR, it really only has 17...... but I'm sure we will. And I get it, but
15 stops of usable DR is heaven compared to what digital use to be like.....many years ago.
 
Martin Stevens said:
I really hope that 4 years from now we're not still saying whoa that camera doesn't really
have 19 stops of DR, it really only has 17...... but I'm sure we will. And I get it, but
15 stops of usable DR is heaven compared to what digital use to be like.....many years ago.
Click to expand...

Personally, Dragon has plenty of DR for me. My beef is with everyone claiming Dragon has more DR than Alexa which in all the tests posted here are proven untrue. I don't care whether it's less, the same, or more, I just want the honest truth. A standardized DR rating for every camera system would be nice so we have a perfect comparison across the board. Everyone measures it differently so there are so many variables. I'm happy shooting Dragon, the low end color cast and noise will most likely be fixed in a firmware update and all will be good. If I seriously get called an Alexa fanboy again kiss Reduser goodbye. I've used the camera on 2 shoots in my career and the rest on RED. Just here to hopefully point things out that can be improved on.
 
Matt Ryan said:
Personally, Dragon has plenty of DR for me. My beef is with everyone claiming Dragon has more DR than Alexa which in all the tests posted here are proven untrue. I don't care whether it's less, the same, or more, I just want the honest truth. A standardized DR rating for every camera system would be nice so we have a perfect comparison across the board. Everyone measures it differently so there are so many variables. I'm happy shooting Dragon, the low end color cast and noise will most likely be fixed in a firmware update and all will be good. If I seriously get called an Alexa fanboy again kiss Reduser goodbye. I've used the camera on 2 shoots in my career and the rest on RED. Just here to hopefully point things out that can be improved on.
Click to expand...

Matt, you are so well liked and admired on RU...and your contributions, and hunger for clarity always appreciated. Only by deconstructing Dragon's pros and cons..are we ever going to have a chance to improve it...and make it the BEST camera it can possibly be. These discussions are invaluable, and are what make RU such a compelling resource. Keep asking...the questions that need to be asked.
 
Very useful, Gavin -- I appreciate your approach, and I think all of what you say here makes sense.

Matt Ryan said:
Personally, Dragon has plenty of DR for me. My beef is with everyone claiming Dragon has more DR than Alexa which in all the tests posted here are proven untrue. I don't care whether it's less, the same, or more, I just want the honest truth. A standardized DR rating for every camera system would be nice so we have a perfect comparison across the board. Everyone measures it differently so there are so many variables.
Click to expand...
A greater man than me once sang, "all I want is the truth... just gimme some truth."

There's so many different ways of measuring the numbers and interpreting the data, I'm not sure if and how that can ever be done. In truth, the real world may show one thing, and the charts may show another; I know this happened with the newer generations of Kodak Vision3 film in the last ten years, as one example.
 
Matt is a very balanced and credible voice on RU. I read his posts in their entirety since I learn from each one. So much of RU is blind faith support which does not serve either RED nor the individual. Only from penetrating questions comes understanding and necessary competitive improvement. There can become a bubble effect within any forum. You need those who have the qualifications and balls to question the answers to keep any group relevant and vibrant.

As far as grey scale test charts - don't see much of an application since, imho, it has little relevance other than academic discussion. Give me a well rendered, detailed and noise free Elk instead ;-)
 
Matt Ryan said:
Personally, Dragon has plenty of DR for me. My beef is with everyone claiming Dragon has more DR than Alexa which in all the tests posted here are proven untrue. I don't care whether it's less, the same, or more, I just want the honest truth. A standardized DR rating for every camera system would be nice so we have a perfect comparison across the board. Everyone measures it differently so there are so many variables. I'm happy shooting Dragon, the low end color cast and noise will most likely be fixed in a firmware update and all will be good. If I seriously get called an Alexa fanboy again kiss Reduser goodbye. I've used the camera on 2 shoots in my career and the rest on RED. Just here to hopefully point things out that can be improved on.
Click to expand...

Yes. and yes some more. I am a bit troubled by Florian's comparison test as I really expected Dragon DR and noise to be better than Alexa, which has been available for years now. Dragon still looks great and I hope the low end noise is solved via software. We can all agree that Dragon is better than Epic and I've been very happy with the images I've achieved out of that camera for the past couple years. I've also been very happy with the images I've shot on Alexa. But it IS so valuable to know how these tools compare, where their strength and weaknesses are so that we can make better informed decisions in and out of the field!
 
Zeb B said:
As far as grey scale test charts - don't see much of an application since, imho, it has little relevance other than academic discussion. Give me a well rendered, detailed and noise free Elk instead ;-)
Click to expand...

Hehe. Well, the usefulness of the grayscale tests is in relative comparison. It's a very consistent and predictable test for comparison between sensors e.g. MX and Dragon. If we could point an Alexa at the chart and burn off 72 frames we would be able to get a really good look into its closet of skeletons as well. Everyone already has a number in their head for what MX does, so they can compare the two and see what more they're getting with Dragon.
 
Long and short answer to all of this chaos.

Dragon right now at any given point in ISO is always capturing 18-19 stops of visible and measurable stops of Dynamic Range.

Which is significantly more than film btw.

If you want "access" to a 16+ stop Dynamic Range image you will want to rate and expose properly for Dragon in the realm of ISO 800-2000 with a subtle curve via DRAGONcolor/REDlogFilm. A more aggressive curve like say REDgamma4 will pretty much chop that up and ISO 2000 can reveal that amount still.

When shooting at ISO 250 however, you may only be "revealing" 13/14 stops of visible Dynamic Range via REDgamma4. With a less aggressive curve favoring shadow detail you can see more.

However, no matter what you are always capturing 18-19 stops.

How you use these stops is up to you. How you choose to expose for a given scene is up to you. How many stops in a given scene is up to you or the scene itself. The additional stops provide a very flexible latitude in post both in tonal and chromatic manipulation.

Alexa doesn't see past 15-16 stops on this chart btw. At ISO 800 you are getting a usable 14 and likely "curving to less". But hey, the last time I gave that a good measure was in 2009 and then again in 2010.


The real question happening behind closed doors is how much Dynamic Range do we actually need to capture and where do we go from here? That's going to provide some of the info for future technology. More Dynamic Range and solid color science does indeed provide more workable latitude after the fact. If we hit 20-25 ever, that's going to be scary.


Also, if you really want to have some fun, slap a 1/8th Digicon or Lowcon on Dragon. Or as Evin recommended/enjoyed a 1/16th Schneider HD Classic Soft. You know, because that's pretty much how the Alexa sees light and how everybody has measured it minus the few who measured without that low con in there. I mean ON I think published it was a "true 14" without that filtration in there and the low con is spreading it out that addition bit more.
 
In all these comparisons with Alexa one thing keeps to be ignored - the resolution.
It goes without saying that downscaling image increases both the dynamic range (while decreasing noise level) and colour depth.
This is also why comparing Dragon with MX at the same resolution makes no sense, much less doing so with Alexa...
RED has long figured out that you need to start with higher res then you want to end up with...
So take the 5K/6K R3D's - properly process them down to 2K/HD of the Alexa and then compare them...

Ultimately this is not about comparing RED to Alexa. This is about what you can deliver with RED.
So what I would like to see is properly finished 4K files (even from these chart tests) from the 5K/6K R3D's with applied sharpness (that RED does not do in-camera unlike other manufacturers), etc...

That is what matters...

Just my 2 yen...

:sifone: Peter
 
Peter Majtan said:
In all these comparisons with Alexa one thing keeps to be ignored - the resolution....
Click to expand...


Except for the comparison tests where you can download the Alexa and Dragon raw files... which is pretty much all of them.
Such as this one here: http://www.reduser.net/forum/showthread.php?117070-First-Day-with-Dragon
or this one here: http://cinematography.net/UWE/index.html

So you're free to compare them at 1080p/2k or any resolution. And people do. And the results and conclusions are still the same.

So I'm not sure what you're talking about...
 
Matt Ryan said:
While it is nice to see 18 stops, there is no way in hell they are all usable.
Click to expand...

It is beneficial pointing out that the "DR usability falloff" is normal. :)

Elsie N said:
The beauty of that chart to me is that 17 and 18 are easily recoverable so they are useable for many things.
Click to expand...

With that in mind it is beneficial pointing out that stops at the upper and lower extremes are most suitable for the appearance of natural falloff, and "recovered" at the cost of that falloff.



Hats off to Gavin and Stacey for the tests.
 
Juan Melara said:
So you're free to compare them at 1080p/2k or any resolution. And people do. And the results and conclusions are still the same.
Click to expand...

YMMV...


Juan Melara said:
So I'm not sure what you're talking about...
Click to expand...

If you really want to compare - what I would like to see is the best image an experienced DP together with an experienced colorist can get of a specific camera and then try to match that image from the other cameras you want to compare to. In our case do it both with Alexa as a reference and then with Dragon as the reference.
I like how many of these test start with "The objective was very simple, a totally neutral and unbiased evaluation of currently available cameras."
That is the problem to begin with...

In the ideal world the "best test" would be:

Get an DP and colorist with experience, understanding and "affinity" for Alexa to shoot and process the best image from a given set setup.
Then get a DP and a colorist with experience, understanding and "affinity" for Dragon (which is the hardest part as the Dragon is still very much work in progress and very few have adequate experience and understanding) try to match that image from the same setup...

Then reverse the process with the Dragon as the reference...

The irony in all of that is that the final image will be open to "interpretation" and hence the YMMV...

So what is it that I am talking about? There simply is no objective, neutral and unbiased way to compare any cameras and any direct comparison of a particular spec is never going to give the answer some are searching for. Or put it in another way - anyone can find the answer THEY are looking for. The question really is - what are you looking for?

:sifone: Peter
 
Phil Holland said:
Long and short answer to all of this chaos.

Dragon right now at any given point in ISO is always capturing 18-19 stops of visible and measurable stops of Dynamic Range.

Which is significantly more than film btw.

If you want "access" to a 16+ stop Dynamic Range image you will want to rate and expose properly for Dragon in the realm of ISO 800-2000 with a subtle curve via DRAGONcolor/REDlogFilm. A more aggressive curve like say REDgamma4 will pretty much chop that up and ISO 2000 can reveal that amount still.

When shooting at ISO 250 however, you may only be "revealing" 13/14 stops of visible Dynamic Range via REDgamma4. With a less aggressive curve favoring shadow detail you can see more.

However, no matter what you are always capturing 18-19 stops.

How you use these stops is up to you. How you choose to expose for a given scene is up to you. How many stops in a given scene is up to you or the scene itself. The additional stops provide a very flexible latitude in post both in tonal and chromatic manipulation.

Alexa doesn't see past 15-16 stops on this chart btw. At ISO 800 you are getting a usable 14 and likely "curving to less". But hey, the last time I gave that a good measure was in 2009 and then again in 2010.


The real question happening behind closed doors is how much Dynamic Range do we actually need to capture and where do we go from here? That's going to provide some of the info for future technology. More Dynamic Range and solid color science does indeed provide more workable latitude after the fact. If we hit 20-25 ever, that's going to be scary.


Also, if you really want to have some fun, slap a 1/8th Digicon or Lowcon on Dragon. Or as Evin recommended/enjoyed a 1/16th Schneider HD Classic Soft. You know, because that's pretty much how the Alexa sees light and how everybody has measured it minus the few who measured without that low con in there. I mean ON I think published it was a "true 14" without that filtration in there and the low con is spreading it out that addition bit more.
Click to expand...
TRUTH!
 
Peter Majtan said:
In all these comparisons with Alexa one thing keeps to be ignored - the resolution.
It goes without saying that downscaling image increases both the dynamic range (while decreasing noise level) and colour depth.
Just my 2 yen...

:sifone: Peter
Click to expand...

Quite the contrary. We're not only supersampling the Dragon and Epic to 1080p thereby decreasing their noise but we're also temporally supersampling the noise to ensure we're seeing everything that the sensor can see (within reason: 72 frames).
 
Play nice, everyone -- stupid argument deleted...
 
Last edited:
Aaand with that one Jeff gets to 15K.
Way more than 645.

:)
 
a little DR real life scenario Dragon test. White car, stright into the sun to shadowside of a house. Could have exposed it more for the sky but still...


 
You must log in or register to reply here.
Back
Top