Weird banding on my Epic-W

[Gavin Greenwalt]

I'm prepping for a big shoot and RED took kindness upon my soul and gave me a beta to help them test. I can say that in my first test it's about 90% fixed already. So I'm sure they'll solve this for sure soon. I'm not sure if I can post comparisons but the green is totally gone on a lamp gradient.

 

[Aaron Lochert]

I'm prepping for a big shoot and RED took kindness upon my soul and gave me a beta to help them test. I can say that in my first test it's about 90% fixed already. So I'm sure they'll solve this for sure soon. I'm not sure if I can post comparisons but the green is totally gone on a lamp gradient.

Siiiick.
That's good to hear.

I guess I was correct in some way: "one could happen before the other"

Wasn't expecting the banding to be fixed first. That's awesome. Correct color processing can come later, I'm getting very pleasing results with my current workflow, even if it's not correct or accurate.

 

[Sabovic Adis]

How's the lows'n'highs? Clippish or roundish? :biggrin:

 

[Gavin Greenwalt]

The weird magenta smudgy highlights are gone as is the green halo. It isn't "rounded", it's properly linear. If anything they used to be slightly rounded and flattened out too much.

 

[Sabovic Adis]

That's great news, Gavin, glad to hear!
And the overall DR? About same, bigger, smaller? :biggrin:

 

[Gavin Greenwalt]

Considering how subtle the error is, I imagine little to no change, they just linearized the response slightly and removed some slight undulations in response that were non-linear gain.

 

[Jason Han]

This was my biggest problem with the Helium... I've been sending Graeme and tech my R3D's for weeks before everyone even noticed this. The green highlight shift you are seeing is a missed debayer process in camera. If you shoot RAW and do a full 1:1 resolution demosaic in post, it should debayer the missed process and you shouldn't have the green banding shift. If you shoot ProRes, you bake that green shift in, so never shoot ProRes until this is fixed. I was on a show where I had to shoot ProRes as it was requested, and everything we shot had this color shift baked in, we had to do an emergency camera swap after seeing that. I was pulled aside and told how much this will cost post to fix, I was extremely embarassed for bringing a faulty camera in. I did extensive tests before, but never seen this problem as I never shoot ProRes in camera. The good news is that I don't have this problem when shooting raw as it will debayer back in post. Red tech is aware and should be releasing a firmware soon is what I was told

 

[Sabovic Adis]

Ugh, the snapshots that we see in this thread are actually ProRes - snapshots? :rofl:

@Gavin
Thank You very much. :biggrin:

 

[Kazu Okuda]

This was my biggest problem with the Helium... I've been sending Graeme and tech my R3D's for weeks before everyone even noticed this. The green highlight shift you are seeing is a missed debayer process in camera. If you shoot RAW and do a full 1:1 resolution demosaic in post, it should debayer the missed process and you shouldn't have the green banding shift. If you shoot ProRes, you bake that green shift in, so never shoot ProRes until this is fixed. I was on a show where I had to shoot ProRes as it was requested, and everything we shot had this color shift baked in, we had to do an emergency camera swap after seeing that. I was pulled aside and told how much this will cost post to fix, I was extremely embarassed for bringing a faulty camera in. I did extensive tests before, but never seen this problem as I never shoot ProRes in camera. The good news is that I don't have this problem when shooting raw as it will debayer back in post. Red tech is aware and should be releasing a firmware soon is what I was told

Hey Jason,

Just curious how do you do the "full 1:1 resolution" can this be done in red cine?

 

[Jason Han]

Ugh, the snapshots that we see in this thread are actually ProRes - snapshots? :rofl:

@Gavin
Thank You very much. :biggrin:

No, I think they were snaps from R3D, but taking snaps from the player of the raw image won't debayer it. You need to actually bring it in a program like Davinci and set R3D's to full res.

Hey Jason,


Just curious how do you do the "full 1:1 resolution" can this be done in red cine?

In Davinci:
Project settings > Camera Raw > RED > Decode Quality: Full Res. Premium
And when you render go to advanced and tick "Force debayer to highest quality"

I only use davinci now so I can't say Premiere does the same, perhaps try setting playback quality to full and drop it in to see if it decodes it.

Graeme might be able to chime in on the Red Cine X, as he tested it with my R3D's that I sent weeks ago. We found that opening them in software full res fixes most of this problem

 

[Sabovic Adis]

Jason, what about exporting a .jpg or .tiff out of RCXP at full - debayer? Good? No good?

 

[Jason Han]

Jason, what about exporting a .jpg or .tiff out of RCXP at full - debayer? Good? No good?

That may work, I can't check as I can't seem to open RCXP anymore, crashes immediately

Here is an example of what I'm talking about, these are frames I sent to Graeme a while back. Bottom is opened through normal viewer and top is opened through Graeme's 1:1 res demosaic

After shooting with the Helium everyday in various lighting scenarios I also found that the CRI of the light source seems to affect the green shift the most. The above light source was the greenest light I could find, a normal office fluorescent. It's only the mid-highlights that you'll find this shift coming from:

F2.8

F4

Here is a ProRes snap lit with 95 CRI, as you can see it's a lot less prominent

 

[Luca Orlandi]

I'm sure that RED will solve @100% soon.
At the time the problem occurs also with natural sun light.

 

[Anthony Berenato Jr]

What lenses were you running with? Wondering if it's possible that the coating is also manipulating the image end-result (exaggerated by Helium still using a differing color science). For example, the Rokinon / Xeen set typically leans green in comparison to a lot of other lenses.

 

[Luca Orlandi]

Tried Zeiss CP2, Canon L, Sigma Art.
Same result. Same green banding.

 

[Kazu Okuda]

No, I think they were snaps from R3D, but taking snaps from the player of the raw image won't debayer it. You need to actually bring it in a program like Davinci and set R3D's to full res.



In Davinci:
Project settings > Camera Raw > RED > Decode Quality: Full Res. Premium
And when you render go to advanced and tick "Force debayer to highest quality"

I only use davinci now so I can't say Premiere does the same, perhaps try setting playback quality to full and drop it in to see if it decodes it.

Graeme might be able to chime in on the Red Cine X, as he tested it with my R3D's that I sent weeks ago. We found that opening them in software full res fixes most of this problem

WOW! Thanks so much Jason! Very excited to try this out!

 

[Sabovic Adis]

That may work, I can't check as I can't seem to open RCXP anymore, crashes immediately

Here is an example of what I'm talking about, these are frames I sent to Graeme a while back. Bottom is opened through normal viewer and top is opened through Graeme's 1:1 res demosaic

And that's the only difference - different viewers? Same ISO, Gamma space etc?
Bottom looks also a bit flatter to me.

 

[Jason Han]

What lenses were you running with? Wondering if it's possible that the coating is also manipulating the image end-result (exaggerated by Helium still using a differing color science). For example, the Rokinon / Xeen set typically leans green in comparison to a lot of other lenses.

Also forgot to mention this, I only noticed the problem when I first put on EF glass (Zeiss). I never noticed the problem when using Cooke and other PL's. I've tested nearly every PL and it was minimal or no green banding shift (apart from the -10 tint that I permanetly have on due to the green nature of the OLPF/sensor). The moment I tried EF glass I was literally saying to everyone wtf is this?

Tried Zeiss CP2, Canon L, Sigma Art.
Same result. Same green banding.

Can you try what I suggested? Bring the raw footage in post and debayer 1:1 full res, that should minimize the green shift

And that's the only difference - different viewers? Same ISO, Gamma space etc?
Bottom looks also a bit flatter to me.

Yeah it does look a little different, Graeme did not specify if he changed gamma, but I doubt he would correct the image as it was for testing purposes.
Please try what I suggested guys it works

 

[Luca Orlandi]

I don't think the problem is the mount... CP2 with PL or EF = same banding.
Also Angie EZ zoom show the problem.

Also forgot to mention this, I only noticed the problem when I first put on EF glass (Zeiss). I never noticed the problem when using Cooke and other PL's. I've tested nearly every PL and it was minimal or no green banding shift (apart from the -10 tint that I permanetly have on due to the green nature of the OLPF/sensor). The moment I tried EF glass I was literally saying to everyone wtf is this?


Can you try what I suggested? Bring the raw footage in post and debayer 1:1 full res, that should minimize the green shift


Yeah it does look a little different, Graeme did not specify if he changed gamma, but I doubt he would correct the image as it was for testing purposes.
Please try what I suggested guys it works

 

[Gavin Greenwalt]

It's not a debayer issue. I was reproducing the banding straight out of RCX at full debayer. But I guess since this is catching some traction again I'll go ahead and post the results of an alpha firmware fix.

This is with saturation jacked to +100 in photoshop to exagerate the improvement and the remaining quibbles (slight magenta shift now which can be corrected with a slight curve adjustment in your grading application). The color shift was annoying but really the banding was the irreperable problem before and that's fixed in the current WIP firmware, so they're really close to pushing out a solution.