Welcome to our community

Be a part of something great, join today!

Register 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 :)

Camera testing and comparisons

Status
Not open for further replies.
geoffboyle

geoffboyle

Well-known member
Joined
Nov 10, 2007
Messages
266
Reaction score
0
Points
0
Age
76
Location
It varies daily
Website
www.gboyle.co.uk
I've just shot a whole bunch of dynamic range and skintone tests at the Gothenburg Film Festival, actually at the workshops organised by Camera Centre at the workshops.

I'm intending to make them available on cinematography.net in a week or so, it'll take me that long to process it all.

So, the big question....

I want to put them all out to 1920 * 1080 DPX uncompressed with much smaller reference files. That's the approach I've used for the last ten years of these tests and that seems to give people useful data to compare.

I know how to output the Alexa material and the SI material and the Panasonic material and the Canon 7D material but what do you want from the RED rushes?

If I put them out with a Rec709 conversion I'll get screamed at, and quite rightly, if I got Red Gamma2 and RedColor 2 I'll get screamed at, if I go Log, which I think is the right way, I'll get screamed at.

What do you think is the right approach?
 
With RED only valuable option is the original R3D (REDCODE RAW) file.

In RXC there is a nice and pretty useful option (a button you click) to output a snapshot in R3D file.

It takes between 1.2 MB to 1.4 MB per file (snapshot=1 frame) that depends if you have 4K or 4.5K resolution.

Then from R3D you can do in RCX, Storm or any other R3D enabled system or app (RCX, Storm, DaVinci Resolve, Scratch, Pablo, Apple Color, Filmlight-Baselight, Nucoda, Smoke, Autodesk Lustre, Mistika, etc,...)

from that R3D snapshot whatever you want.

It's a RAW and pretty much similar approach to still photo RAW file where you are able to play with all image parameters

like a WB (Kelvin), ISO, FLUT, exposure, shadow, brightness, contrast, saturation, RGB colors, etc,... and then convert and maybe resize to whatever you would need.

Hope this helps.
 

Attachments

  • RCX_example.jpg
    RCX_example.jpg
    87.7 KB · Views: 0
Uhm, I think Jim said Red Gamma2 and RedColor 2 was made to make it look as good as possible for most occasions, I would go with that.
 
Herein lies the basic conundrum with Raw cameras - and I assume you'll be capturing Alexa images to RAW as well - as soon as you bake in any given look to something as inflexible as a DPX, the whole point of RAW capture becomes moot. Sure you have to bake in at some time in any production but doing so before testing grading options and limits gives a false impression of the camera. I've seen many tests of Red footage falling apart alarmingly and each time, when I've traced back the workflow, the fault is revealed as DPX. For instance, rendering out mixed lighting to DPX and then trying to 'rescue' parts of the image back to a particular colour balance. Noise. Saving a DPX from a file with an area of the image at, say, zone 1 or 2 and then trying to lift the blacks in the DPX. Noise. However, make the same adjustments in RAW, before rendering and the image is revealed to be clean. Misleading. Sure, these cameras have limits but DPXs have more.

Perhaps, if you can't supply RAW files to the tests, you could try rendering several versions to DPX of a given shot, illustrating the actual limits of highlight, shadow, ISO, FLUT, white balance, tint, curves, sharpness etc. etc. RAW provides, not the limits of the DPX file. And all this applies equally to Alexa as Red.
 
Post as REDcolor2 and REDlogFilm... or single frame R3D.

What? No resolution tests? :-)

And thanks for asking.

Jim
 
Jannard said:
Post as REDcolor2 and REDlogFilm... or single frame R3D.

What? No resolution tests? :-)

And thanks for asking.

Jim
Click to expand...



Thanks Jim, that was sorta what I was heading towards!!

No resolution tests yet, maybe when I add the Epic and Scarlet to these tests....

Oh and the Alexa was recorded in Pro Res Log as the RAW option was not officially available when I shot the tests.

I try to make them "real world"
 
geoffboyle said:
Well, it helps and hinders!

It forces me to keep the images in a Red-centric world and the idea is to give people images to compare that reflect their use in the "real world"

Nobody delivers images in a RED format to broadcasters!!
Click to expand...

Why not let them create their own deliverables?

Roberto Lequeux said:
Why not R3Ds along with a link to the free software?
Click to expand...

Red shoots negatives. Till Red delivers a box that puts out prints, such as a ProRes module or what have you, then this seems to be the only logical way to examine Red.

Does that make sense Geoff?
 
I see. So you are hoping to find a way to deliver yourself something that "looks good" but doesn't eat at the edges of what is available in the negative, correct?

Then why not actually grade the best way you can, squeezing yourself as much as possible out of it while keeping trying to keep it looking "normal", and then include a note reminding people that red shoots negatives, that each production will need need to develop the footage themselves, that this was your "objective grade" provided for comparison purposes?
 
Roberto Lequeux said:
I see. So you are hoping to find a way to deliver yourself something that "looks good" but doesn't eat at the edges of what is available in the negative, correct?
Click to expand...


I'm trying to find something that lets DP's directors, Editors and Producers make approximate comparisons.

Nothing I do can be definitive, it's just to try and share info.

I can't go on making version after version.

So, I think that Jims suggestion is the right way to go, as this will let people take images into progs like Photoshop and Aperture and play with them.

With the log images they will have more room to play.
 
5391873611_5c2c672e10_z.jpg

5391873601_9cba0370ab_z.jpg

Göteborg exposure test with 7D and RED

Looking forward to get finally that RED REDCODE RAW R3D from your Göteborg digital camera comparison test and to play with it.

You just have to tell us for each shot in that test exact chosen aperture, shutter, lighting and we do the rest of "baking" image with

"RED-centric world" systems like a RCX, Storm, DaVinci Resolve, Scratch, Pablo, Apple Color, Filmlight-Baselight, Nucoda, Smoke, Autodesk Lustre, Mistika, etc,...
 
Geoff,

Did you shoot any film for your test? I would say, in addition to the great advice you already got, that you should treat it like film.

IN other words, nobody delivers the RAW film, but maybe they would deliver a color timed print of the film. IN this sense you could deliver a color timed DPX sequence.

If you shot film, you would have to go through a lab, so I suggest using a reputable "RED" lab that has processed a lot of RED footage. These people will know how to extract the best possible images from RED's R3D media file.

Whenever I have the kinds of questions you present, I just default to treating it with the same care and exactness you would treat 35mm negative film. Get it in the hands of people who know how to treat it and can get the most out of the images you recorded.

David
 
David Battistella said:
...so I suggest using a reputable "RED" lab that has processed a lot of RED footage. These people will know how to extract the best possible images from RED's R3D media file.

David
Click to expand...

This make me think of the Kodak image care programm... will there be a RED image care programm that certify the studio in the future?

...

Pat
 
geoffboyle said:
Because that assumes that the people looking at the images can, in most cases they wouldn't have a clue but want/need to have some idea.
Click to expand...

Smart!

redlogfilm/redcolor 2 would be the "common" log delivery for these cameras, as it is cineon and is "smart'ish" like the ARRI LogC in the sense that it tries to keep the top within range at "any" ISO setting.

that's where the comparisons will make most sense.

RLF is quite batchy, and will be used like that by most...

From many post and DP people RAW doesn't make much real sense.
Especially not in a comparative setting.


Agreed.
 
Status
Not open for further replies.
Back
Top