Just got our Dragon

[Jan Reiff]

don´t know, this low light frame looks too bad for dragon when i compare the posted footage before in other threads, ...
seems to be some bugs in the shooting set up ?

 

[Matt Ryan]

That depends. Lifted blacks can often compresses shadows and reduces shadow contrast which actually can reduce the appearance of noise. bringing the blacks down may actually increase contrast and increase the visibility of noise.

It is likely Dragon is better than MX for low light, but I'm a little worried it is only very slightly so, and MX isn't great in low light to begin with compared to most of the competition.

Maybe Marc will post an R3D trim so we can see.

 

[Carlos Reyes]

What is it that people are so worried about? I cant understand this obsession with "low light." Are people trying to light entire movies with a candle?

Take a look at Her, shot with Alexa wide open using ambient light and small LEDs. http://www.icgmagazine.com/wordpress/2014/01/02/the-way-she-hunts-my-dreams/

Dragon has been proven to AT LEAST match Alexa in DR though its probably more based on whats been written here by very respected members of the industry, there has been an obvious improvement in the noise floor, better compression schemes. The color science is much improved and the skin tones I've seen look phenomenal, and we all know things are being tweaked so we can expect even more improvement which is kind of mind blowing. The camera can shoot up to 100 FPS at 6K and even higher FPS as you window down to less resolution (Funny how I'm referring to 4K as less resolution). It has a bigger than S35 sized sensor at 6K which yields a different look and one many have been asking for, and for those complaining about lens coverage you get a S35 size sensor at 5K.

For $10K you get to upgrade your camera to AT LEAST the equivalent of a 6K Alexa, though I'd say when everything settles a bit I'd expect Dragon to outshine Alexa in at least a couple of ways, only because its new technology just as I'd expect the 4K Alexa to have obvious improvements over the original Alexa.

Some peoples expectations on here dont make any sense to me, and I guess hyping up the sensor for at least two years didnt help that situation.

You get an entirely new camera for $10K. If that doesnt blow your fucking mind I dont know what will. Buy a GH 4K for $2K which im sure will make some nice images and be done with it.

The Epic with all its quirks and perceived issues already has the ability of creating phenomenal images, many of which are playing at tens thousands of theaters world wide. Now Dragon kicks everything up a couple of notches and people are still complaining. There are no excuses anymore, the camera can't be blamed at this point.

Not to derail Marc's thread, if any one wants to discuss this further:
http://www.reduser.net/forum/showthread.php?111203-Dragon-Reality-Check

 

[Björn Benckert]

from jpg.

http://www.flickr.com/photos/107237384@N04/11955057874/

 

[Adam Ekblad]

Don't forget the compression from the Camera (we have no clue what ratio it was shot @), what settings it was exported with, and who ever is hosting the file.
It is probably all just compression, don't jump to conclusions until we have the .R3D.

 

[Niko Pueringer]

Carlos, telling us how incredible of a deal Dragon is has nothing to do with the question of what might have led to the increased amount of noise in that shot. I'm extremely passionate about RED's products as well, but it's important to discuss technical issues, limitations, and expectations without people feeling the need to sell the brand. RED's cameras already do a fine job of that.

 

[Marc Ruiz]

Thanks for posting a shot! Always nice to see extreme tests like this.

People have thrown around the idea that 2000 ISO is as clean as 800 ISO on MX, but this shot looks too noisy for that comparison. Any idea as to why? Black shading? Color/gamma space?

I think you are looking at the screen shot and that does not really give a full look at the potential of the camera - I was enjoying the sensitivity of a single match out of the box and sharing that really quick.

Marc

 

[Marc Ruiz]

Maybe Marc will post an R3D trim so we can see.

Certainly

 

[Carlos Reyes]

Carlos, telling us how incredible of a deal Dragon is has nothing to do with the question of what might have led to the increased amount of noise in that shot. I'm extremely passionate about RED's products as well, but it's important to discuss technical issues, limitations, and expectations without people feeling the need to sell the brand. RED's cameras already do a fine job of that.

Sure, but trying to draw conclusions about noise floor based on a compressed JPEG is not going to get us very far in understanding the cameras low light capabilities. The small amount of dragon footage is pretty frustrating but im sure the flood is coming sooner rather then later.

 

[Marc Ruiz]

So my quick out of the box test seems to have done more harm than good...

So I have done the blackshading to default and documented all the settings...

Without further adieu...

R3D of 2000 ISO with Match: https://www.dropbox.com/s/y3ekuymyaaqps4z/A001_C004_0114RD.0000176F.R3D

R3D of 2000 ISO with NO Match: https://www.dropbox.com/s/a0wu84n2539cpb8/A001_C004_0114RD.0000143F.R3D

JPEG with Settings: https://www.dropbox.com/s/aedea0jsqyqahpu/2000 iso.jpg

I hope this dispels some concerns and answers some questions

Thanks

Marc

 

[Marc Ruiz]

I am trying to post a larger JPEG as well for those who have trouble with the links...

 

[Matt Ryan]

Noise looks on par with current MX at 800ISO wouldn't you say? Just set wb and did a curve with Redlogfilm in RCX-Pro. Down Res this to 5K and match the MX and I think you'll see it's an improvement. Technically I probably should have crushed the blacks further for proper levels, but I don't think it looks bad. It's just a match light for Christ sake.

Dragon 2000ISO with WB adjustment and slight curve:

Just for fun, same look but dropped to 800ISO (Almost noise free if you ask me):

 

[Marc Ruiz]

I'm inclined to agree that the Dragon 2000 ISO looks on par or better than the MX 800.

I have another Epic still waiting for the upgrade I am tempted to do a side by side comparison.

 

[Matt Ryan]

I'm inclined to agree that the Dragon 2000 ISO looks on par or better than the MX 800.

I have another Epic still waiting for the upgrade I am tempted to do a side by side comparison.

Go for it. The more comparisons the better. Just make sure they are with the same lens, same settings, same distance to subject, shot at same time, etc.

 

[Denis Buhot]

I am tempted to do a side by side comparison.

Please do !! That would be a great input, really...

 

[Niko Pueringer]

I'm inclined to agree that the Dragon 2000 ISO looks on par or better than the MX 800.

I have another Epic still waiting for the upgrade I am tempted to do a side by side comparison.

That would be amazing if you did! Thanks for posting the R3Ds and chiming in with some info. It's exactly what I was hoping for. I'm just trying to learn as much as I can about the new sensor before our camera is called for the upgrade.

 

[Filip Orlandic]

A little bit dimmed and cinematic grade.
Highlights aren't blown and minimal presence of noise.

 

[Bob Gundu]

Wow. I'm blown away. Really. Dragon has such filmic grain.


ISO: 2000

ISO: 3200

ISO: 4000

ISO: 6400

 

[Brian Boyer]

I putzed around in RCX until I got something I thought was passable. I know it's sometimes hard to judge the quality and how objectionable the noise may be on a static image but I find this level of noise, under the circumstances, quite acceptable (on my uncalibrated monitors, at least).

The scene is lit with a match at ISO 2000. This amount of light is certainly within most lighting budgets.

I feel like if this were a scene from a movie and you're concentrating on anything but her face, you're not so much watching a movie as you are a 90 minute camera test.

 

[Carlos Reyes]

Wow. I'm blown away. Really. Dragon has such filmic grain.


ISO: 2000

ISO: 3200

ISO: 4000

ISO: 6400

These look phenomenal!