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 :)

Hannover Hands On comparisons of RED F35 & D21

Status
Not open for further replies.
Geoff... thanks for clearing all this up and thanks to the group for toning it down.

I think where Graeme and I felt like a wrong conclusion had been drawn was around the "no need to go further" with RED, when the images had less noise than the D21 (as verified by our buddy Sanjin). The impression was that the RED could not keep pace with the others in this area, which is just not accurate.

Geoff certainly tells it like he sees it... and is a welcome addition around here.

Jim
 
OMG, Sanjin is going to flip his lid when he sees Jim quoting him. :sifone:
 
geoffboyle said:
I just wanted to let people get at the RAW data and then make their own conclusions.
Click to expand...

Hi Geoff, and thanks a lot for sharing the files.

I, like others, was able to open the files in Photoshop now. Excellent!

There's one more difference that grabbed my attention: i.e. in the contrast test, looking at the histograms, it can be seen easily where the other cameras' sensors clip: there's a peak in the values near, but clearly below full white. In the red shots, this peak it at pure white: there's no headroom whatsoever above assumed sensor clipping, and this makes it impossible to say if clipping has indeed happened in-camera, or during conversion.

Could you please do one more test, with just one image... take the source for i.e. RED-RAW-contrast-plus-one0255, and export it using lowest possible ISO, with exposure compensation set to minus two stops.

This should result in a very dark image, but one that certainly shouldn't be clipped in export. If the highlights of this version clip at the same point as in your currently available files, it would confirm that clipping indeed happens in-camera, not during export.

I know you've spent a lot of time on this already, but doing this test would be greatly appreciated, at least by me ;-)
 
Eki Halkka said:
Hi Geoff, and thanks a lot for sharing the files.

Could you please do one more test, with just one image... take the source for i.e. RED-RAW-contrast-plus-one0255, and export it using lowest possible ISO, with exposure compensation set to minus two stops.
Click to expand...


I'll certainly try that.

I'm also burning DVD's of the R3D files to send to RED and we'll see if they get anything better.
 
Jannard said:
I think where Graeme and I felt like a wrong conclusion had been drawn was around the "no need to go further" with RED, when the images had less noise than the D21
Click to expand...

Yeah, it wasn't the noise that I was looking at when I made that decision more the detail in the grill on the front of the amp.

Anyway, I've put them all up now :-)
 
Eki Halkka said:
Could you please do one more test, with just one image... take the source for i.e. RED-RAW-contrast-plus-one0255, and export it using lowest possible ISO,
Click to expand...


I just had a very quick look at that, there is no exposure compensation in the SG settings.

Where using 320 IS0 results in an 8 bit clip of 255, setting it to just 250 ISO give a maximum value of 242.

At that difference I'm not sure that you're losing anything.
 
Jay A. Kelley said:
Bottom line: You're good at what you do, no doubts there, and I would guess the people that will REALLY make use of your "Data" will never post, or have an opinion, they are most likely too busy working. I would guess they will look at what you have, file it away, and then check it against their own findings when it's time to shoot something.

The rest is cannon fodder.

Jay
Click to expand...

But we still skim the threads and read through all the BS. It's just that we have grown tired to post these days, since it seems like a waste of time, most of the time.

Your tests are highly appreciated Geoff, thank you.
 
Hi Geoff, just to know what we are comparing in the tests:
when we compare the dpx files from the different camera's, say at 'normal',
did all camera's had the same T-stop on the lens?
or was the 'normal' point for each camera determined with a graychart and spotmeter, iso value and set the each camera accordingly? ( or in red using false colors, searching for green on the graychart )
i know with the RED setting colorspace in redspace/rec709/raw can give different 'normal' values which makes comparing 'normal' different.

If the T values are absolute in the test, it could-will probably be that the 'normal' - or 'optimal' value for the different camera's will be 1stop or more shifted up or down - as the iso rating for the RED is different than for F35 or for the ARRI, and thus their optimal T-value for a given situation.

Also when working in reality, and having highlight as in your tests - which are very valid test situations - i and probably also you as a experienced RED user, would use only the lightmeter/18% gray as a starting point and then use false colors in RAW space to see when clipping in the highlights would occur and then see how much clipping i would accept on this image. This would then for this shot my 'normal' point, not necessarily the 18% graychart/lightmeter reading in the image.
Setting the T value this way takes only seconds on the RED, and from there on, it is usefull to know in steps of a half T or 1/3 T what happens when light changes in both directions.
 
Just to hijack this thread one more time for my own political motivations... sure would have been handy if Geoff could trim r3d files and only export a single frame....:toetap05:
 
Anyone noticed this in the blue channel on the F35 dpx's? I'm seeing a vertical pattern in the saturated colors. I see it very clearly on the Coke bottle label, but it's also in other parts of the scene. It's also visible on the Genesis DPX files available on CML.

Graeme
 

Attachments

  • blue.jpg
    blue.jpg
    53 KB · Views: 0
geoffboyle said:
I just had a very quick look at that, there is no exposure compensation in the SG settings.

Where using 320 IS0 results in an 8 bit clip of 255, setting it to just 250 ISO give a maximum value of 242.

At that difference I'm not sure that you're losing anything.
Click to expand...

Could you post the frame anyway? Though you're probably right, i'm one of those people who like to see stuff themselves ;-)
 
Graeme Nattress said:
Anyone noticed this in the blue channel on the F35 dpx's? I'm seeing a vertical pattern in the saturated colors. I see it very clearly on the Coke bottle label, but it's also in other parts of the scene. It's also visible on the Genesis DPX files available on CML.

Graeme
Click to expand...

Is it some kind of RGB stripe pattern artefact?
 
No idea. I've only seen it in the Genesis and F35 DPX's from CML. Not seen it on any other camera.

Graeme
 
Actually, I could be wrong about it being just F35 / Genesis. CML has a vast array of dpx files from a variety of cameras. I'm seeing the same thing in the blue channel of the F23 also.

Graeme
 
Aren't F23/35/Genesis virtually the same cameras?
 
By the way, I noticed an 8GB card was used.
 
Irri said:
Aren't F23/35/Genesis virtually the same cameras?
Click to expand...

Certainly the F35 / Genesis are very similar. The F23 may use some of the same electronics - that'd make sense that it's just it's sensor that's different?

Graeme
 
Status
Not open for further replies.
Back
Top