Raven Grain

[Cory Vetter]

Hey guys!

I've shot a few projects lately and have noticed more and more that in grading my footage I've had a considerable amount of noise in my Raven footage. I dunno if the grain structure in the Raven is larger than the Scarlet/Epic?

I've attached a few stills for you all to see, and I just wanted to get your thoughts as to if this has been the case for you guys. I've kept a close eye on my histogram, and have exposed correctly, so none of these shots are shots that I've had to "save" in the grade. The midtone grain has been the most bothersome as there's a ton of chroma noise that's pretty unpleasant looking. Idk... any thoughts as to what I could be doing differently would be greatly appreciated.

FYI all of these have been graded in DragonColor 2 and REDlog with no luts and minimal pushing and pulling with exposure. Mainly adding some contrast.

Thanks!!

Cory

P.S. The compression on the uploaded photos kinda smooths out the grain, but for real... It's there.

 

[AndreeMarkefors]

Nothings stands out to me in these images (but I'm not a Raven owner).

I did want to say though, that you can provide a frame for reference, but for image analysis make sure to provide 100% crops.

Also, it normally doesn't take 3 posts until the first request for a .r3d shows up. Especially if you want people's opinion whether something is "normal".

In terms of grain structure, even if two sensors have the same pixel pitch, if one of them uses a larger area and greater resolution, the relative size and appearance of the 'grain' will look finer when both images are scaled down to a fixed size like 2k or HD. Same goes for shooting 4k on Raven compared to shooting at a higher resolution with another Dragon sensor and scaling down to 4k (of course).

 

[Aaron Lochert]

Also, it normally doesn't take 3 posts until the first request for a .r3d shows up. Especially if you want people's opinion whether something is "normal".

I'm comfortable with being "that guy".

Yo, Cory, upload r3d snapshots for those clips!

 

[Jake Wilganowski]

Its because with weapon and epic you are shooting 6k, the crop to 4k (Raven) is SIGNIFICANT, so you are basically "zooming in" on the noise structure. I knew folks would be in for a little surprise with raven. Expose to the right and use the least compression you can.

 

[Bob Gundu]

Post R3D snapshots.

 

[Nick Morrison]

Its because with weapon and epic you are shooting 6k, the crop to 4k (Raven) is SIGNIFICANT, so you are basically "zooming in" on the noise structure. I knew folks would be in for a little surprise with raven. Expose to the right and use the least compression you can.

What he said.

Post an R3D!

 

[jordan levy]

i actually had the opposite situation. shooting a raven at 4khd (barely cropping its own sensor) and a weapon at 4khd ( cropping way into its sensor) the dit on set informed me on one of my night scenes that the raven was coming out cleaner and the weapon was a bit grainier. all settings exactly the same and rating at 500 iso at 5:1.
i would of barely noticed until he punched in on his monitor. so we shot all our lower light scenes with the raven and its all being broadcasted and its all nice and clean.

 

[AndreeMarkefors]

shooting a raven at 4khd (barely cropping its own sensor) and a weapon at 4khd ( cropping way into its sensor)

If you're shooting the same sensor (Dragon) at the same resolution—4kHD—then you're using the same physical sensor area size and the results should be identical. The sensor crop has nothing to do with it. It's just a consequence of one of the sensors offering higher resolution (physical size) to begin with.

There will of course be variation due to black shading and other factors like the internal boards/circuitry being different. I'm not sure anymore if it's been officially stated that the actual sensor is the same in all Dragon generation cameras? Probably, only cut to different sizes.

 

[Jake Wilganowski]

i actually had the opposite situation. shooting a raven at 4khd (barely cropping its own sensor) and a weapon at 4khd ( cropping way into its sensor) the dit on set informed me on one of my night scenes that the raven was coming out cleaner and the weapon was a bit grainier. all settings exactly the same and rating at 500 iso at 5:1.
i would of barely noticed until he punched in on his monitor. so we shot all our lower light scenes with the raven and its all being broadcasted and its all nice and clean.

Black shade

If you experiment with different tempatures you will see how much it can vary the noise in the darks

Compression also

And finally, yes 4K crop from any dragon sensor with all things being equal should be the same noise, so some other variable was affecting it

 

[jordan levy]

both black shaded at prep, together, we were forced to because firmware updates.
the only thing i could think of is that the weapon was cropping the sensor much more.
but honestly i could not see the difference until the dit really zoomed in and showed me.
it was negligible to my eye at full resolution, but i listen to my DITs.

 

[Jake Wilganowski]

It is not possible for weapon to be cropping the sensor more at the same resolution, the resolution is the resolution with dragon sensor across different bodies

I'm not saying I don't believe you, fyi
I bet if you blackshaded weapon at a higher op temp you'd be cleaner

 

[Jake Wilganowski]

What about Olpf and lens? Did u have the standard Olpf in weapon? That's what's in raven

 

[jordan levy]

Standard in both.
I mean all cameras can vary a bit. I can never get two Alexas to match perfectly.
And this was almost a year ago when both bodies were brand new. So the dits best guess was that the weapon, starting out with a much bigger sensor has to crop more. Like when you crop to 2k and you get more grain.
And both were using light weight optimos.

 

[AndreeMarkefors]

So the dits best guess was that the weapon, starting out with a much bigger sensor has to crop more. Like when you crop to 2k and you get more grain.

You are partly right, but haven't thought it completely through.

It IS true what you say that when you crop to a lower resolution, the relative size of the 'grain' (noise) will increase and be more prominent. The noise will be most obvious at 1:1 pixel crops. An example of that would be using a FF Dragon sensor but setting it up to only use a 2k 2048x1080 center crop when shooting (10.24x5.40mm), and then delivering a 2k 2048x1080 file to look at. Compare this to using the much larger 6k size of the sensor (30.72x15.80mm) and then squeezing that resolution down to a 2k deliver file. The noise will become visually smaller and less prominent.

But in the Raven example above, when both cameras are shooting at the same resolution, 4kHD (19.20x10.80mm) it's the same as having "established the same crop". The resolution is the 'sensor wafer crop' (or pixel pitch crop) in that sense. There is ANOTHER crop of the sensor that simply refers to a 6k Dragon having a larger physical sensor to begin with. But it is technically irrelevant that the 6k Dragon has a lot of unused sensor that's sitting outside the chosen resolution crop.

I hope this makes sense. I'm only writing it in an effort to clear any uncertainties up.

 

[Jake Wilganowski]

Thank you I just didn't feel like writing a long post to explain but you're absolutely correct

 

[jordan levy]

yes, i completely understand, but all i was saying was that this was the dits effort to explain to me what might be happening. as i said, these cameras were brand new and all i can say is it was happening. realistically there are so many variables that go into making a sensor, or a lens, could of been less light transmission of the lens at a certain focal length, could of been the lack of exposure at a certain angle, could of been anything. like i said it was so little difference it was negligible….but the Raven was very clean.

 

[Mike Finegan]

unless same lenses and camera angles and field of view with identical histogram - wouldn't be apples and apples
assuming all else is same regarding red settings and crop.

 

[Nick Morrison]

unless same lenses and camera angles and field of view with identical histogram - wouldn't be apples and apples
assuming all else is same regarding red settings and crop.

That's a great point. If one camera has a more lowcon lens for example, the noise and exposure may be different - even if both cameras are at same 4K FOV. Optics will have a lot todo as well.

 

[Brent G Miller Jr]

Nothings stands out to me in these images (but I'm not a Raven owner).

In terms of grain structure, even if two sensors have the same pixel pitch, if one of them uses a larger area and greater resolution, the relative size and appearance of the 'grain' will look finer when both images are scaled down to a fixed size like 2k or HD. Same goes for shooting 4k on Raven compared to shooting at a higher resolution with another Dragon sensor and scaling down to 4k (of course).

What he said. The grain on the Raven will indeed appear "larger" with the 4K crop on the dragon sensor compared to down converting footage with higher native resolution. that's why we use Ravens as B-cams. When it comes to the dragon sensor, a 4K crop is a 4K crop, regardless of how big the sensor is and should have the same size of grain. So a 4K crop on a Weapon should yield the same grain as a Raven.

 

[Karim D. Ghantous]

They look pretty good to me!