Overly DARK/De-saturated RED footage in scratch

[Nathan L. Collett]

Hello,

I'm a bit in a panic and hopeing maybe to find some help here. I'm pretty ignorant about scratch.

I'm the director on a feature film shot on red and being color corrected on scratch in 2k for 35 output. And I'm a bit concerned about what i'm seeing.As I'm overseas, actually work in Kenya in Africa, I can't see the process day - to - day. Our co-producers sent us some quicktime outputs... but to me, in addition to other issues, they look quite dark on the skin tones, they are black faces. The blacks almost looked crushed. In the past i was told that on scratch it looks more de-saturated, and darker... then on 35. but is this true? I'm having doubts, as if what i see is what i get on 35, i'm not liking it.

I imagine there is some way that the look can be similar in scratch to approximate 35? or no? What is the best advice? I'm unable to go see the process directly, so what should i do to make sure my film doesn't look like sh*t?

Any and all suggestions appreciated.

Cheers.

Nathan Collett

 

[Mark L. Pederson]

Nathan -

There's lots of ways a quicktime output can wrong.

The first thing you need to do is make sure what you are seeing is what they are seeing. And quicktime - especially quicktime from windows - is a slippery slope.

Ask them to output some TIFF frames. Open them in Photoshop on the most calibrated display you can find.

Are they grading to a REC. 709, etc.?

If the RED material is exposed properly - Scratch is only going to "crush" what the operator crushes.

 

[Noah Kadner]

Yeah having close communication between DP, DIT and colorist is an absolute necessity with RED projects. They can look as amazing or as horrible as you can imagine with a few right or wrong clicks of a mouse.

Noah

 

[John Tissavary]

Of course it's possible that the grade is going in a direction you're not wanting. Scratch is a color correction tool just like DaVinci, Baselight, etc..., and has no 'tendency' to make footage look like something other than what the user is directing it to.

Yes, Red can have a tendency to fall off a bit quicker in the bottom shadow range than film, but it's largely correctable unless the footage is too underexposed.

When you compare to 35mm, perhaps you need to break that down a bit to the components that you are going after to communicate effectively with the colorist - if they're not sure what you want, it's really tough to dial in long distance. Perhaps you're asking them to bring up the shadows a bit without raising the actual black level?

However: as others have mentioned, if you're viewing Quicktimes you're going to have a very hard time because you're dealing with a 'moving target'.

Overly dark (conversely, washed out, elevated blacks) are common symptoms of quicktime gamma bug (perhaps Apple considers it a 'feature' since it's been there for over 10 years and not yet been fixed). There are some codecs that are better in consistency - ProRes, DNxHD come to mind - but .h264 is an absolute mess from one system to the next.

Ask the facility to send you tif or jpg images (or sequences) instead. While these will be much larger files than .h264, at least you can judge what you're looking at. Also, you'll need to make sure you're viewing on a calibrated monitor - this is another place that the viewing chain often breaks down.

I just spent an extra week on a job because the (remotely viewing) client wanted QT, and the turnaround cycle called for .h264. The ensuing cluster#$%^ let to major discrepancies between what the client was seeing, what we were seeing (even in the same QT), and what the final DPX output was.


cheers,

JT

 

[Bruce Allen]

Hey Nathan, old USC partner-in-crime... nice to see the Kenya thing going well! Not so nice that you're having problems.

Anyway, stop wasting time and send me the R3D file

Let me know if you need an FTP to upload to.

Bruce Allen
www.boacinema.com

 

[Fredrik Harreschou]

Nathan, can you please tell us something about what kind of QTs you are receiving in terms of codec, and what you are viewing them on - Mac or PC? Are the QTs directly output from Scratch?

You should do some kind of gamma test with the footage you are receiving. Ask them if they can send you a grayscale ramp in the same codec with a width of 1024 pixels. Make sure it is evenly spaced starting from 0 to 1024 (or 0 to 256 in 8 bit).

Down the line you should try out VLC if you haven´t. It is better at handling cross platform/software gamma issues. http://www.videolan.org/vlc/

There is no inherent property of the Red footage that would make it look desaturated or darker compared to 35 - unless you are viewing raw footage on a regular display. It has great latitude and color definition - but not as much as slow 35 film IMHO.

 

[Nathan L. Collett]

Hey Nathan, old USC partner-in-crime... nice to see the Kenya thing going well! Not so nice that you're having problems.

Anyway, stop wasting time and send me the R3D file

Let me know if you need an FTP to upload to.

Bruce Allen
www.boacinema.com

Hey Bruce. Great to hear from you. Thanks for that. Don't have the color corrected r3d files, just the low rest h264 quicktimes. Let me send you a link via email so you can download them and let me know what you think...It would be great to get your opinion on this.

Cheers. Nathan

 

[Krishna Rajan]

hey Nathan, i am going to be directing a feature this summer so this is something i also would like to know, while shooting did you not have an opportunity to see the daily footage? in other words did you get a chance and comment on what was being captured and was that footage looking ok? if so you could safely assume that its an operators error on the color correction.

good luck

-krish

 

[Nathan L. Collett]

Nathan -

There's lots of ways a quicktime output can wrong.

The first thing you need to do is make sure what you are seeing is what they are seeing. And quicktime - especially quicktime from windows - is a slippery slope.

Ask them to output some TIFF frames. Open them in Photoshop on the most calibrated display you can find.

Are they grading to a REC. 709, etc.?

If the RED material is exposed properly - Scratch is only going to "crush" what the operator crushes.

hanks Mark for that. I appreciate your response and at least feeling that support is out there. I've asked them for TIFF to get a better sense of what is going on...Maybe I'll be able to post them and get your feedback once they send me something. Thanks!

 

[Nathan L. Collett]

Of course it's possible that the grade is going in a direction you're not wanting. Scratch is a color correction tool just like DaVinci, Baselight, etc..., and has no 'tendency' to make footage look like something other than what the user is directing it to.

Yes, Red can have a tendency to fall off a bit quicker in the bottom shadow range than film, but it's largely correctable unless the footage is too underexposed.

When you compare to 35mm, perhaps you need to break that down a bit to the components that you are going after to communicate effectively with the colorist - if they're not sure what you want, it's really tough to dial in long distance. Perhaps you're asking them to bring up the shadows a bit without raising the actual black level?

However: as others have mentioned, if you're viewing Quicktimes you're going to have a very hard time because you're dealing with a 'moving target'.

Overly dark (conversely, washed out, elevated blacks) are common symptoms of quicktime gamma bug (perhaps Apple considers it a 'feature' since it's been there for over 10 years and not yet been fixed). There are some codecs that are better in consistency - ProRes, DNxHD come to mind - but .h264 is an absolute mess from one system to the next.

Ask the facility to send you tif or jpg images (or sequences) instead. While these will be much larger files than .h264, at least you can judge what you're looking at. Also, you'll need to make sure you're viewing on a calibrated monitor - this is another place that the viewing chain often breaks down.

I just spent an extra week on a job because the (remotely viewing) client wanted QT, and the turnaround cycle called for .h264. The ensuing cluster#$%^ let to major discrepancies between what the client was seeing, what we were seeing (even in the same QT), and what the final DPX output was.


cheers,

JT

thanks John for that. This is great info to have and echos what Mark is saying. I really need to get those TIFFs to get a better sense of what I'm seeing. Cheers.

 

[Nathan L. Collett]

hey Nathan, i am going to be directing a feature this summer so this is something i also would like to know, while shooting did you not have an opportunity to see the daily footage? in other words did you get a chance and comment on what was being captured and was that footage looking ok? if so you could safely assume that its an operators error on the color correction.

good luck

-krish

Hi Krish. I did see the dailies during the shoot. Some of which, even though not perfect, looked better than the color corrected footage i'm now seeing... so while i can't assume anything at this point.. i think its somehow in the color correction that things are going a miss.

All the best. Nathan

 

[Bruce Allen]

i think its somehow in the color correction that things are going a miss.

No shit!

Here's a quick test of an INSANELY BRIGHTER AND MORE SATURATED version. You know, to counter the "DARK/De-saturated" version you already have

For final grade you would choose a color palette, dial things back and color more selectively.

Just want to show that there's plenty of stuff there that can be dug out of the original footage:

Bruce Allen
www.boacinema.com

 

[Nathan L. Collett]

oh my... yes the saturation is out of control but they look great... all the info is there.. so my cameraman and the red didn't let me down...

as luck would have, i really like the 2nd from top, it looks just as i envisioned that scene.

thanks man.