Clipfinder RA Color inconsistency

[bobbystone]

I find vast discrepancy between the color when using RA and clipfinder, and even within clipfinder. If I set the color using the Edit Look Settings window in clipfinder, the preview window looks completely different than the clipfinder browser window, even after making new proxies. And clipfinder and RA look completely different too. Is there something I'm missing here? I'd REALLY like to get a suitable workflow going that has less inconsistency.

Thanks!

[Gunleik Groven]

I find vast discrepancy between the color when using RA and clipfinder, and even within clipfinder. If I set the color using the Edit Look Settings window in clipfinder, the preview window looks completely different than the clipfinder browser window, even after making new proxies. And clipfinder and RA look completely different too. Is there something I'm missing here? I'd REALLY like to get a suitable workflow going that has less inconsistency.

Thanks!

Clipfinder can render both to the new-new color science, and the old - new colorscience anno summer 2009 (all terms in RED lingo)

RedAlert uses the old-new AFAIK

But it all depends on how you render the images...

If you render images with the render command in CF, the resulting files should look consistent with RedAlert with the same settings...

You have to look at the "Loupe" image to make any judgement, BTW.

Makes sense?

[Hans-Georg Daun]

I need to correct Gunleik:

Clipfinder cannot render with FLUT™™™™ color science (I'm waiting for the next version of REDline to make that possible).

Clipfinder can render with 2009 color science and original color science.

REDCINE-X, ROCKETCINE-X and Apple Color (with the beta plugin) display and render with FLUT™™™™ color science.

The latest version of REDAlert displays and renders with 2009 color science.

Clipfinder can display FLUT™™™™ color science (as well as 2009 color science and the original color science).

Clipfinder's main window displays off of the proxy files, i.e. via the RED Quicktime codec, which is still on 2009 color science and does not respect changes made to color space, gamma space and curves.

Clipfinder's loupe window displays via the latest SDK and should be consistent with REDCINE-X and ROCKETCINE-X (if set to FLUT™™™™ color science).

With Clipfinder 2.5 you can edit look settings with FLUT™™™™ color science, then send to ROCKETCINE-X for rendering, but the integration between the two is a little shaky and you need REDrocket.

You can also send to REDCINE-X from Clipfinder 2.5, but then you need to grade in REDCINE-X and build 158 has issues with spurious black frames in the render output and with incorrect timecodes unless you render the entire clip.

Interesting times indeed.

To summarize:

If you're happy to stick with 2009 color science, set Clipfinder to 2009 color science. It should then be consistent with REDAlert (in the loupe window).

If you have MX material or want to work with FLUT™™™™ color science you're living on the bleeding edge.

Cheers,
Hans

Update (April 2): REDCINE-X build 183 solves the aforementioned issues and is greatly recommended. You can also now edit your look in Clipfinder and REDCINE-X will read all parameters. Three cheers for Deanan !

[Damien Molineaux]

...

REDCINE-X, ROCKETCINE-X and Apple Color (with the beta plugin) display and render with FLUT™™™ color science.

...

Color works with FLUT color science ? Where's that beta plugin ?

I wasn't aware of that, and I can't find any info regarding this...

Cheers,
Damien

[bobbystone]

Thanks Hans for that very clear explanation. PLEASE RED let's get the new CS implemented!

[Adil Lahoulou]

Color works with FLUT™™™™™™™™™™™™ color science ? Where's that beta plugin ?

I wasn't aware of that, and I can't find any info regarding this...

Cheers,
Damien

I'm with Damien on this.... Is it a case of it uses the new color science but doesn't have the FLUT™™™™™™™™™™™ controls enabled???

Hans please elaborate.

thx

[Mikael Lubtchansky]

Clipfinder's main window displays off of the proxy files, i.e. via the RED Quicktime codec, which is still on 2009 color science and does not respect changes made to color space, gamma space and curves.

I always thought the QuickTime codec was still original color science and not the 2009 flavor of builds > 17



With Clipfinder 2.5 you can edit look settings with FLUT™™™™™ color science, then send to ROCKETCINE-X for rendering, but the integration between the two is a little shaky and you need REDRocket.

You can also send to REDCINE-X from Clipfinder 2.5, but then you need to grade in REDCINE-X and build 158 has issues with spurious black frames in the render output and with incorrect timecodes unless you render the entire clip.

Is this because ROCKETCINE-X can read the "FLUT™™ enabled .rsx" while REDCINE-X only reads .RMD files that grades cannot be sent from Clipfinder to REDCINE-X ? I hope you can soon implement .RMD support even if it is experimental / bleeding edge since the .RMD format is not finalized...

btw any plan to have a White Balance color picker in Clipfinder one day or will this always be a manual operation ?



If you have MX material or want to work with FLUT™™™™™ color science you're living on the bleeding edge.


truly exciting time and it will be even more awesome once REDline and QT codec are updated... while we all wait for ClipMeister :-)

Mikael

[Hans-Georg Daun]

Color plug-in beta info and download link are in this thread
http://www.reduser.net/forum/showthread.php?t=41389

and, yes, the Color plug-in does use the FLUT™™™™ color science but does not have FLUT™™™™ and Shadow controls.


I always thought the QuickTime codec was still original color science and not the 2009 flavor of builds > 17

Now that you mention it I'm less sure it is 2009 color science.
In any case, it is currently not useful for judging color.


Is this because ROCKETCINE-X can read the "FLUT™™™™™™™ enabled .rsx" while REDCINE-X only reads .RMD files that grades cannot be sent from Clipfinder to REDCINE-X ? I hope you can soon implement .RMD support even if it is experimental / bleeding edge since the .RMD format is not finalized...


REDCINE-X does read the RSX file as long as no RMD file exists, but in my tests did not accept REDcolor, REDgammma, FLUT™™™™ or Shadow settings from the RSX file.

RED have made it clear they do not currently want 3rd parties to generate or modify RMD files. I respect that.
Clipfinder will support the RMD format very soon after RED give the green light for developers.

Cheers,
Hans

[Damien Molineaux]

Thanks Hans-Georg,

I had done a quick search with no success.

Cheers,
Damien

[Gunleik Groven]

Hans Georg.

Sorry about any wrong info!

When one do an FLUT correction in CF and render out a still frame, what Color Science is the image rendered with?

[Nick Shaw]

When one do an FLUT™ correction in CF and render out a still frame, what Color Science is the image rendered with?

That would certainly appear to be done with an SDK decode, and so use whatever colour science is chosen in the prefs.

[Gunleik Groven]

OK. That's what I thought.

Then I haven't screwed up in my tests... -:)

Cheers!

G

[Nick Shaw]

It should be noted that the above is based on my observation, not hard info from Hans.

Also based on my observation, I believe the QuickTime codec v3.8 is based on the original colour science, and v3.9 is based on 2009 colour science.

[Hans-Georg Daun]

Yes, stills are rendered via the SDK based on the color science setting in the preferences.

Cheers,
Hans