Need Help For critical Feature's pre visualization

[Esmaile]

I'm going to shoot a feature and the feature will be printed to film.
how can i pre visualize the emulation of the film stack on set?
could it be possible by using the below stuff :
1.Native 1080p output (10bit LOG)
2.A Pro monitor like cine-tall with 3d LUT import
3.A default 3d cub (color managing system)like Arri cub or kodak

[liquidigital]

Ultimately if you have the resources to do so, I'd have some sample tests printed out. Do you have a post-house in mind yet?

[Esmaile]

I'm Just waiting for my Red cam
No but I'm looking for a pro post-house with an arri laser printer
Any suggestion?

[liquidigital]

I'm not certain what DuArt is using, but they made a print for Offhollywood's test recently. The footage was lacking a bit in saturation, but I think they (meaning Offhollywood) were under the gun time-wise with the LUTs. Mark's a busy man! But the footage looked great to me. Wish I could offer more info. Here's a link to DuArt...

http://www.duart.com/

[Paris Remillard]

Where are you, Esmaile? On the west coast, fotokem and efilm are both really good full service lab/post facilities (there are many others, too) . Both have Arri scanners and printers. If you are going that route, it wouldn't be a bad idea to talk to whichever lab you decide on first, and let them help you set up a workflow together. May help eliminate problems later. Good Luck.

[Esmaile]

thanks liquidigital & parisrem I'd call efilm , it'd be a good idea to ask them for a default workflow

[mmost]

I'm going to shoot a feature and the feature will be printed to film.
how can i pre visualize the emulation of the film stack on set?
could it be possible by using the below stuff :
1.Native 1080p output (10bit LOG)
2.A Pro monitor like cine-tall with 3d LUT import
3.A default 3d cub (color managing system)like Arri cub or kodak

If you go through a proper DI path with tight color calibration, there is no reason why the print should look significantly different than the digital image. The primary difference would be in the viewing environment, in other words, the look of a projected image vs. a directly viewed image on a video monitor. Some colors will change a bit, but the native colorimetry of the Red images is more similar to a film pallette than it is to a video pallette, at least in my experience so far. Video cameras have much higher native saturation and other characteristics that Red images do not seem to share. As a result, it is much easier to achieve the "Red Look" on a film print than it is from a video original. In fact, on all of our film DI's, we use film as our primary target, regardless of the origination. However, on Red DI's, I'm likely going to change that (at least for the moment) and use digital cinema (the DCI spec) as the primary target, as we've already developed an "inverse LUT" that yields basically the same image on the film print. Because we use Scratch as our DI system, we color correct directly from the R3d files with no need to go to log space prior to the render for film recording. So if you like what you're seeing on the digital image, there is really no need for a film preview LUT as you're already in linear space with an image that can be translated pretty faithfully, at least in our facility. But just to be sure, I can toggle between the two (the final digital image and the inverse LUT image viewed through a print preview LUT) to confirm that they're basically the same prior to doing the film render.

What I'm saying is that the need for preview LUTs is not really there when shooting with a device like the Red. With cameras like the Viper and the Genesis, the dynamic range of these cameras is preserved by encoding the output on a log curve and in the case of the Viper, avoiding any white balance adjustments in camera. This in turn leads to the need for a viewing LUT, as the light (and in the case of the Viper, somewhat green) image is not representative of what you're shooting, at least not when viewed directly. Red preserves its dynamic range by recording the RAW output of the Bayer sensor - which means, in part, that what you see is what you get, at least within the limits of Rec 709. You don't go through a lin to log process in order to achieve that.

One more thing - there is no "native 1080p output" on a Red. There is no 1080 output at all, only 720, at least at this point in time.

[Esmaile]

mmost,thanks for your cool information,

It's a good idea to have inverse LUT from DCI to film,but what would you do if you shoot a feature with digital camera and after blowup to film,saw the colors diden't that much good as you expected? I mean when i have a film LUT on set,I myself will choose the proper colors that will presented on screen with no false color,please corect me if i'm rung.

The other thing,
you sad"we color correct directly from the R3d files with no need to go to log space prior to the render for film recording"

Ok if so,than when you want to blowup your final to film,you have to convert it to log,if i'm not mistaken , because the laser printer dosen't support liner targets ,in this case you'll have some problems with dark areas and high lighted areas maybe.

Besides, there's a question comes up here :
when you finalize your feature to blow up than how 'd you covert it to log ? I mean by default presets ?
if so it dosen't make sense , other wise u send a tiff target to the DI company and let them decided how to convert it to log? , If not so than
what is your method to convert the final liner or R3ds to log?

BTW, thanks for your point,In case of 1080p, output

[mmost]

It's a good idea to have inverse LUT from DCI to film,but what would you do if you shoot a feature with digital camera and after blowup to film,saw the colors diden't that much good as you expected? I mean when i have a film LUT on set,I myself will choose the proper colors that will presented on screen with no false color,please corect me if i'm rung.

The difference in individual colors is of little consequence if the overall intent is maintained. Since basically no professional projects are distributed on only one format these days, it is virtually impossible to absolutely ensure that every color in the image is exactly the same across multiple disparate platforms, such as film, digital cinema, and video. So the general approach in DI work is to choose a specific target (most commonly film print) and use lookups to make all other deliverables look as close as possible to the selected target. However, Red is something of a unique case because the control of color is far greater when working directly from the RAW file than it is once the material is converted to an RGB format, regardless of whether you do it in log space or linear space. Using digital cinema color space is the least restrictive of all - it is based on CIE XYZ color space, which essentially encompasses all colors visible to the human eye. Of course, in practical use, that pallette is restricted by the display technology used, in our case, a DLP Cinema projector. And it is further restricted by the various distributable formats that is will ultimately be presented on. But is does yield the most control and therefore what is arguably the best result. The trick from that point on is to have properly done LUTs to convert that visual result into its closest equivalents on both HD video and film. In terms of production design, that is done through experience. If very specific colors are that important to your particular piece (is this a commercial, where the product or packaging colors are that critical?), you need to test them and see which ones work best. Film is never the only distributable format today, in fact, it is not even the most profitable or ubiquitous. One needs to consider all deliverable platforms in making such judgements.

The other thing, you sad"we color correct directly from the R3d files with no need to go to log space prior to the render for film recording"

Ok if so,than when you want to blowup your final to film,you have to convert it to log,...in this case you'll have some problems with dark areas and high lighted areas maybe.

Besides, there's a question comes up here :
when you finalize your feature to blow up than how 'd you covert it to log ? I mean by default presets ?
if so it dosen't make sense , other wise u send a tiff target to the DI company and let them decided how to convert it to log? , If not so than
what is your method to convert the final liner or R3ds to log?


As far as "the DI company", we ARE the DI company. And yes, you do convert to a Cineon log format for the files that are going to a film recorder. You also have to convert to a Rec 709 format for HD deliverables. In the case of the latter, we use a custom built DCI to 709 LUT, which deals with both the color temperature (white point) difference and the gamma difference between the two formats, as well as gamut mapping for any out of gamut colors. For the former, we convert to Cineon Log using an "inverse LUT" that is designed to maintain the greyscale as well as the color choices in the digital cinema version for a film print target. We confirm it by projecting the result using our standard print preview LUT (which is quite accurate, we use it for all of our film DI work) prior to film recording. So we are not guessing as to what the print will look like, we are accurately previewing it prior to creating the print and making any minor adjustments at that point.

One more thing: a film recording is not normally considered a "blowup," especially when you're starting with a 4K image - but even when you're using 2K, as the vast majority of DI's do.

[Esmaile]

ok perfect , one more question:
what workflow you will recommend for someone who whants to shoot with Redcam 4k raw or 2k raw and then go for film print,will he has to make a cineon for the Di company? if so how can he rich that perfect print preview LUT? it seams that no one can make a good convertion from
raw to cineon without it,so isn't a good idea to make a tiff 16 bit and send it to DI guys to convert it to cineon perfectly without any cliping?what do you thing mmost?
Besides,can i have your mail i like to ask you some thing about scratch ?
thanks agine for your cool notes