DRAGON vs EPIC MX Noise - NOT GOOD

[Gunleik Groven]

This is indeed pretty civil, as far as I can see (haven't read through the whole thing, but been brushing through it)

Keep calm, and document things well, and there is nothing but documentation happening.

To recap a bit...
People were complaining that the old calibration didn't hold highlights like the Alexa.
RED re-calibrates the sensor to hold more highlights.

Now, that's a nice response, methinks.
Thank you RED.

Unless they didn't start swapping the full architecture behind the sensor at the same time, I am not too surprised to see that that recalibration has some consequence on the lowlights. Even without testing, that kinda makes sense to me.

I haven't had much time with the Dragon after the OLPF upgrade, but the shootingdays I have had, have indeed not been disappointing. THe color-richness and fatness in the image vs MX, both in the monitor on set and in the files is still there.
They will act differently in the extremes.

It is interesting to see the difference in Phils and the other examples here.
I haven't had the chance to do side-by sides with several dragons, but from the reports I have, a newer CF Dragon and mine early -X perform similarily.
I know the Proto I used was a bit different from the production unit I now have. Also different to the old OLPF version.

I am not saying the issues are not there. You see what you see and documentation is only good.

I just wanted to put the discussion into a bit of a larger context.

Keep it civil, don't call names, and I am sure RED looks into any real issue.

 

[Brent Carter]

Guys, we are watching this thread. And have noted the people who have brought this to us individually as well. We are investigating internally at the highest levels of engineering. Until we have more data from our tests, premature for us to really comment aside from the fact that "we hear you and are looking into it with all the RED big guns". We will keep the community posted.

And, yes, this thread is what REDUSER is all about.

 

[luigivaltulini]

Luigi could you post a link after clicking on the post # in the post's top right corner? Page #s are random based on what people have their posts-per-page setting at. For instance I only have 17 pages on that thread.

Sorry Gavin
post is #509


I believe and I am sure that the new setting has brought substantial benefits to the colors and the high lights with the new OLPF, the other side of the coin and that we have lost something in low light.
I believe and I am sure that red is working on this, I always see small improvements in the new firmware.

 

[Adrian T.]

Sorry Gavin
post is #509

post numbers can also change when moderators delete and merge.
here's the link with the unique post id:
http://www.reduser.net/forum/showthread.php?p=1333204&viewfull=1#post1333204

 

[Jonathan Stevenson]

Here's a few more examples where I feel the noise levels are much higher than they should be. In these R3Ds, there is PLENTY of light hitting the sensor and yet I'm seeing heavy noise in certain areas.

R3D FILES
https://elevationvideo.box.com/s/155nfvoqpcs67t15atr5

https://elevationvideo.box.com/s/roq8u7wi7b0drnzok6sb

Here's some 100% screenshots of areas of the frame where the noise is much higher than expected.

 

[Ted Parkes]

Thanks for chiming in Brent. I am confident all of this will be address as quickly as Possible from the Red Team.
Cheers

 

[Brian Robbins]

mark you may never notice since you shoot everything at ISO 250.

but yeah if I had the old Olpf I'd probably hold off too...just my opinion based on this one dragon camera body. Hoping to test 2 together on Wednesday.

More than welcome to do a test in-house. Extra Dragons aren't a problem.

 

[David Battistella]

And, yes, this thread is what REDUSER is all about.

Amen brother!

Battistella

 

[Taylor Morrison]

"we hear you and are looking into it with all the RED big guns".

And thank you for this!

Thanks

 

[Brian Merlen]

Thanks for the heads up! Appreciate the insight and honesty! I think this is what RU is all about, geeking out to cameras, doing tests, and all of us striving as a community to always improve the system, the images, the workflow. It takes a family to build a cinema camera Those r3d files I posted have some under exposited spots if people who don't yet have Dragon want to check them out? I'll try and get some more up when possible

Guys, we are watching this thread. And have noted the people who have brought this to us individually as well. We are investigating internally at the highest levels of engineering. Until we have more data from our tests, premature for us to really comment aside from the fact that "we hear you and are looking into it with all the RED big guns". We will keep the community posted.

And, yes, this thread is what REDUSER is all about.

 

[Medavoym]

Here's a few more examples where I feel the noise levels are much higher than they should be. In these R3Ds, there is PLENTY of light hitting the sensor and yet I'm seeing heavy noise in certain areas.

R3D FILES
https://elevationvideo.box.com/s/155nfvoqpcs67t15atr5

https://elevationvideo.box.com/s/roq8u7wi7b0drnzok6sb

Here's some 100% screenshots of areas of the frame where the noise is much higher than expected.

This is EXACTLY what I've been seeing on my two recent DRAGON shoots.
The images are properly exposed in daylight/sunlight. However, the gradient/less bright areas in the picture exhibit a very noticeable red noise pattern other people have pointed out as well. As soon as I began reviewing the footage after the shoot, I was struck by it - the first second the footage was up. I also played with an earlier version of Dragon a few months ago, and the footage didn't have it (even in footage than was less than ideally exposed). That Dragon almost certainly had the old OLPF, as it was one of the first Dragons in the wild.

 

[Álex Montoya]

To recap a bit...
People were complaining that the old calibration didn't hold highlights like the Alexa.
RED re-calibrates the sensor to hold more highlights.

Now, that's a nice response, methinks.
Thank you RED.

Unless they didn't start swapping the full architecture behind the sensor at the same time, I am not too surprised to see that that recalibration has some consequence on the lowlights. Even without testing, that kinda makes sense to me.

What you're saying is more or less that they have changed the ISO. With the new calibration, they have moved the mid gray point, so in fact when it says 800, we're shooting more like 1600. Something like that?

 

[Elsie N]

I'm curious about something... what happens to the noise if you apply ADD to the image. Note: I am not suggesting using ADD as a solution to any noise problems... I'm just suggesting applying ADD in order to give the folks at RED more information to digest.

 

[Rob Macey]

Just for comparison and information. I received Dragon last week and could easily see how ramped up the noise became above 800 in blacks/shadows. However, I don't seem to be having the same noise issues.

Here are some Dragon clips I just shot along with their corresponding jpgs. I would normally shoot these shots at 250-400, but at 800 they're pretty clean.


These were all shot at:
6K 2:1
5:1
800 ISO
Dragon Color / RG3
Temp at 65
Xenar 25 at exposure or just under


https://www.dropbox.com/s/uhhepsakvr101ta/A001_C001_0609TZ.RDC
https://www.dropbox.com/s/hurpa10b3gkgdfe/A001_C002_06093X.RDC
https://www.dropbox.com/s/sivrnibbnf12pw7/A001_C003_0609D0.RDC
https://www.dropbox.com/s/v9vb46zzl0svv3b/A001_C004_0609BP.RDC

 

[Jeff Kilgroe]

Rob,

I haven't downloaded the clips themselves yet, but as far as the JPGs go, those are pretty crushed.

Anyway, nice to see that RED is looking into this.

 

[CJ Roy]

Just for comparison and information. I received Dragon last week and could easily see how ramped up the noise became above 800 in blacks/shadows. However, I don't seem to be having the same noise issues.

Here are some Dragon clips I just shot along with their corresponding jpgs. I would normally shoot these shots at 250-400, but at 800 they're pretty clean.


These were all shot at:
6K 2:1
5:1
800 ISO
Dragon Color / RG3
Temp at 65
Xenar 25 at exposure or just under


https://www.dropbox.com/s/uhhepsakvr101ta/A001_C001_0609TZ.RDC
https://www.dropbox.com/s/hurpa10b3gkgdfe/A001_C002_06093X.RDC
https://www.dropbox.com/s/sivrnibbnf12pw7/A001_C003_0609D0.RDC
https://www.dropbox.com/s/v9vb46zzl0svv3b/A001_C004_0609BP.RDC

These are showing up as 32kb. I think it's just the .rdc folder, not the .r3d file.

 

[Blair S. Paulsen]

Thanks for the post Brent. I know its always best to have an actual solution ready to roll before joining this kind of thread, but sometimes just assuring us that you guys are taking it seriously is very valuable to the tenor of the discussion.

My Dragon is definitely showing the red/orange noise in the shadows. Please include Epic-X #00019 in your investigation to see if it shares any components or other variables with other cameras reporting this issue. I'd be happy to bring it to Irvine or install beta FW and run tests here for uploading to engineering, etc if that would be of any help. I'm also happy to reach out to Phil and bring my Dragon to him for side by side testing if he, and you, think its worthwhile (and he has the time).

At your service - #19

 

[Gunleik Groven]

What you're saying is more or less that they have changed the ISO. With the new calibration, they have moved the mid gray point, so in fact when it says 800, we're shooting more like 1600. Something like that?

I think i will avoid speculating on exactly how they did it, but as the highlights allow 1+ stop more exposure before clip, something in either the optical path or whatever changes sensitivity on the sensor, is PART of what has happened.

As it seems people have non-uniform results, all other things are just stupid speculation. It would not equal the effect similar to change ISO on the old olpf, though

 

[Gunleik Groven]

Thanks for the post Brent. I know its always best to have an actual solution ready to roll before joining this kind of thread, but sometimes just assuring us that you guys are taking it seriously is very valuable to the tenor of the discussion.

My Dragon is definitely showing the red/orange noise in the shadows. Please include Epic-X #00019 in your investigation to see if it shares any components or other variables with other cameras reporting this issue. I'd be happy to bring it to Irvine or install beta FW and run tests here for uploading to engineering, etc if that would be of any help. I'm also happy to reach out to Phil and bring my Dragon to him for side by side testing if he, and you, think its worthwhile (and he has the time).

At your service - #19

RED/orange noise/blotchyness in lower exposures were a part of the old OLPF old ColorScience characteristics.

Haven't tested for that with new.
But I have a remote hunch that this possibly could be on the solvable part of the equation.

 

[Ted Parkes]

Seems the issue is real and is wide spread. The folks at Red are aware and I suspect working very hard at coming up with an explanation and a fix. So now all we can do is assist them in finding the solution as quickly as possible.