RED vs. Genesis/arri d-21 film look

[Nathan Garofalos]

Jamie, check your PM's. This is Nathan from FLF.

 

[Jason Murphy]

Thanks for all the great info.. I appreciate it.. The clipping is a result of poor cc in redalert... very fast preset.. I have posted another shot with a different cc from redalert here...

http://jamieheinrich.com/red_quicktests/better_cc_48th_shutter/

this was shot at 1/48th shutter speed....that still feels video to me.. the still on RED is amazing, but when motion begins it's got a video texture to the motion (bare with my explanation) thanks again

The CC on the shot is definitely better, but it looks as if the RAW file itself is very clipped in the sunlit areas. My guess is that if you exposed so that the sunlit area didn't clip, you could bring the rest of it up in RED Cine, while still retaining highlight detail. A properly exposed RAW file should have quite a bit of flexibility.

One (still) example:

http://www.reduser.net/forum/showthread.php?t=16365

Still having a hard time understanding the "video motion" problem here, but it may well be an issue of the monitor you're working with, like Deanan said.

Also, have you had a chance to shoot any tests on the Genesis yet? If so, do you have any clips for comparison's sake?

 

[Dan Hudgins]

TIF 4K REDLog sample files... please?

TIF 4K REDLog sample files... please?

I have been doing a lot of test shots with RED and cannot seem to crack that "video" look. The still images look amazing but something about the way it flows looks very video still.

[See note below about Global Release shutter]

I want to get that film look as seen with genesis footage. The emotion.. I've shot at various shutters speeds, lenses, frame rates, color correction and adding grain, and with shallow DOF and still feeling like "video". The stroby sharp edges and such are not helping..Is there a process in post or do I just wait for more builds? and what is the main difference between these cameras.. Super bad looks great with the genesis, especially the crappy looking scenes because it feels like bad film stock.... Does anyone else feel this way?

I can try to see if I can give your footage more of a "film look" in my systems CC, but you need to do some things:

1) Set to 24fps.

2) Set to 200 degrees.

3) Expose so that the background and general image on the camera monitor is very dark and the highlights are a little brighter than you would expect a 18% gray card to show on the monitor, then turn on fill lights that are just enough to make the subjects face a bit below 1/4 the monitor brightness.

4) Use backlight that is equal to a gray card reading of EI 320.

5) Have the actor look off camera into the main light so that his nose shadow goes a little away from the camera. The main light should measure 18% gray at EI 800 on an incident dome at the actors place dome into the light.

6) Put a bare bulb (not frosted) no fixture, just a bulb and socket, on the side of the camera the actor is facing, it should be bright enough to add a very small amount of fill and pick up a highlight in the actors eyes, maybe 25 watts depends on your f/ stop.

7) Set the white ballance to daylight and use an 80A filter in front of your lens. This is to give the most dynamic range to the highlights in the 4K REDLog TIF files you will be making since you will shoot under ONLY tungsten light (no mixed K light).

Take your R3D file through REDCINE and make 4K TIF REDLog files of the frames, and post some of the better keyframes as 4K TIF REDLog files that can be worked on and posted back here on REDUSER so people can see what can be done with them, maybe some others would fiddle with them as well, I am sure this topic is of general interest.

As for the motion blur, movie cameras have a feathered edge to the shutter shadow when the lens in wide open, the rolling shutter has a hard edge, so there is a difference between cameras thet shoot with "global reset" and a mechanical shutter, and ones that use a rolling shutter. The ARRI I think uses a global reset because of the optical finder based on their film cameras, this makes the RED ONE (tm) motion blur look a little different, perhaps.

I read somewhere that in India they shoot HD then make a film printing negative then scan the film then project the scans on video projectors, to give their movies a film look.

You cannot get that "film look" with just a 3D look up table, since the film layers do not have the same resolution and bleed (and have other artifacts) in different ways for the Yellow, Cyan, and Magenta.

My program can defocus and mask the RED, GREEN, and BLUE each with its own values and curves, this was done to correct film scans from negatives and positives, but can be used backwards to introduce color crossover and adjacentcy effects into digital images to make them have film like artifacts.

Anyway, if you really want to get your footage looking better or at least less like video, I am happy to look at your method and help you make changes to get closer to the goal.

Do not sharpen the images at any point in the process, just get the actors eyes in focus when you shoot.

Some images from the R3D seem to show an issue with the de-Bayer used to de-Bayer the R3D footage, there is some kind of edge filter on the chroma de-noise (?) that makes the color look detached from the luma in a very Digital Cinema camera non film like way, this is done to make the images look sharper and maybe to help with the compression noise, but I would like some way to turn off the chroma edge filter and just have a softer de-Bayer. Any kind of sharpen is un-film like since film has a rolling off MTF not a sharp one like a Video camera.

You can post the link to the TIF here and PM me if you want to see what I can do to your footage... If it is shot un-film like, then you will need to reshoot since like I have explained to my brother many many times, lab work is not a substitute for good camera work, filters, sets, actors, lights, costumes, make-up, and art direction. Once everything in front of the camera looks just right to your eye, and like a "real" move, then comes the camera artifacts.

The RED ONE (tm) is not a magic wand that turns whatever you point it at into a Oscar winning movie.

Anyway, I need well shot and exposed 4K REDLog TIF files for testing my system, and by way of that you can see what can be done with them, JPG or other files are of no use in this process.

 

[Deanan]

Still having a hard time understanding the "video motion" problem here, but it may well be an issue of the monitor you're working with, like Deanan said.

Given that exposure time is the same across cameras (ie. 1/48sec is the same regardless of the camera), the motion will be the same. Skew does not affect the cadence of the motion just the offset in space (that's why strobing is present in rotating vs rolling vs global shutter). The other factor is the response of the sensor. Linear sensors (genesis and d20/1 included) have a subtly different characteristic to the motion blur than film because of the way motion is integrated into the response curve of film at the toe and shoulder.

 

[Jason Murphy]

Given that exposure time is the same across cameras (ie. 1/48sec is the same regardless of the camera), the motion will be the same. Skew does not affect the cadence of the motion just the offset in space (that's why strobing is present in rotating vs rolling vs global shutter). The other factor is the response of the sensor. Linear sensors (genesis and d20/1 included) have a subtly different characteristic to the motion blur than film because of the way motion is integrated into the response curve of film at the toe and shoulder.

Hey Deanan-

Sorry, I wasn't clearer in my earlier post. Understanding the differences in motion characteristics between the different cameras, shutter types, shutter speeds, etc. is something I actually have no problem with. What I still didn't get was the specific nature of the motion in the 1/48sec footage Jamie posted that he objected to as specifically being "too video-y", and I had a hunch that the monitor might have something to do with it.

Thanks for the explanation, though. It's a good one.

 

[Deanan]

Sorry, I wasn't clear either. I quoted you in agreement

 

[Ivan Cortazar]

How you roll off the highlights also matters. My preference is to roll off the highlights well under 100% and sometimes bias towards a slight cream colored highlights.

Also extremely important is what you look at the footage on. If you're looking at it on an LCD monitor running at 60hz, you will see motion artifacts because you're missing accurate timing on displaying frames.

Deanan, that's really interesting. Does that mean that if I see these videos on a Mac G5, with quicktime and with a MAC LCD, even if quicktime info tabs says that is is playing at 24fps, we may see additional artifacts because of the 60hz?

 

[Deanan]

Yup. Ideally you want 24,48,72hz, etc. for viewing 24p footage.

 

[Dan Hudgins]

sync..

sync..

Yup. Ideally you want 24,48,72hz, etc. for viewing 24p footage.

And even then, the 24fps (23.976+ would be more of an issue) would need to be in sync with the vertical retrace, or you would get a rolling artifact in one of the monitor fields where a field would show parts of two frames.

Real live video, which does not exist any longer, did not have 180 degree shutter look, just the blanking time for the vertical sweep, so true video had more of a 360 degree feel (with interlace blur). You do not get the live video feel any longer with the compressed digital broadcast, like the TV screen is a window, film never had that feel. The interlace blur looked different with various tube cameras and the persistence of the phosphor. When you digitize video from sensor cameras the interlace artifact is more comb like in the digital frames than it would have been perceived viewing live video on a monitor.

If you view film scans in a compressed format on a computer they do not feel the same as viewing them with a movie projector. The "best" way to view motion in the RED ONE (tm) footage is uncompressed in a movie projector, since the movie projector is the image frame rate, and uncompressed. I remember the first time I watched some computer output onto film how much smoother the motion was on the screen with the film projection...

This frame to field sync issue, brings up another issue, what is the frame to field sync in a DLP theatre projector when you project 24.0000fps vs. 23.976+fps source footage, does the projector change the field rate, or will some frames show more fields, or is there a rolling or random image split on one or more field?

 

[Ivan Cortazar]

Very interesting. The hole world is going digital, but still the best way to see Red footage with motion is through a film output... Ironic.

 

[Christoffer Glans]

Lower the high areas in curves... I do not own a Red yet, but everytime I've worked with video and there's something overexposed I ususally lower the highest point. Keep the rest of the curve in whatever you like, but lower the highest point and you get a much smoother turnover into highlights.

Also, magenta doesn't look very nice on video. I usually put alot of wamrth into my footage. It's both an artistic choice but I also think it looks more film.

 

[Jason Murphy]

Very interesting. The hole world is going digital, but still the best way to see Red footage with motion is through a film output... Ironic.

A film output isn't necessarily the best way to view RED motion footage (though, sure, it would work), but a good digital projector designed for cinema projection would probably be just as good (if not better).

If you want that ever-elusive film look, though, probably best to project film.

 

[sergio arguello]

thats great that you posted examples. A couple things that have been stated and came to my attention right away is 1. the obvious shutter being short..ie 90 degrees 2. the over exposure 3. the amatureish camera operation this with 90 degrees will feel even more stacatto. 4. the flat lighting that is common when seeing video. no hint of any seperation, and those all combined might have vision3 looking akward...

 

[Jason Sinclair]

Very interesting. The hole world is going digital, but still the best way to see Red footage with motion is through a film output... Ironic.

well the audiences don't seem to mind, just the camera people. A lot of people i have spoken to say it is better than film.... But they are just average people with no technical knowledge...

 

[Phil Parmet]

Shoot film if you want a film look.

 

[Andrew Walker]

Shoot film if you want a film look.

Yes, then add dirt, scratches and some lateral movement and you've go the look you want. Film is great when you only watch answer prints all the time when there color has been timed right and the print has been struck off the inner negative, which is what an answer print is. I bet there are maybe a handful of people on this board that have seen an answer print of at least one movie. The general release prints are a joke. There are rated at around 1-1.5K. Mainly because they are printed on high speed printers just to get them out the door and into the theaters. Plus they are printed off a copy of a copy of a copy.

So if you want that film look, shoot at 24fps and you got it. Of course knowing a thing or two about how to color an image might help as well. Not to mention lighting and all that cinematography stuff. There's also the small matter of story. But now we are getting away from the look and into the realm good filmmaking.

 

[David Mullen ASC]

Yes, then add dirt, scratches and some lateral movement and you've go the look you want. Film is great when you only watch answer prints all the time when there color has been timed right and the print has been struck off the inner negative, which is what an answer print is.

You're forgetting the film can also be digitally projected, and looks quite good that way. And conversely, digital can be transferred to film and bad release prints made from dupes.

So I think we're confusing the film look issue if we start talking about the quality of release prints. It would be better to compare a digital original, let's say, to a scan of a film negative and talk about what's the same or different, or how they behave while color-correcting, etc.

Ultimately, I do believe that if is it absolutely critical that you achieve a film look, then you should shoot film. If you shoot 24P digital, you can come close with the right camera, the right lighting, the right color-correcting... but to some extent, you'll lose less sleep and hair if you accept that digital also carries its own visual qualities, just as reversal film is different from negative film, or Fuji is different from Kodak. Genesis footage looks like Genesis footage, RED looks like RED, they all may seem similar to film, but not quite, so think of them as also carrying their own textures and you'll be a little happier. All the while pushing your manufacturer to make the camera better over time...

 

[Luca Immesi]

This is my personal opinion but I've found some lenses are better than other with red to achieve a more film look. Very sharp lenses on a digital sensor are more video than softer lenses in my opinion (this on a computer or hd/sd broadcast monitor). Although some dop told me red footage shot with sharper lenses is better once projected on film.

 

[Zakaree Sandberg]

This is my personal opinion but I've found some lenses are better than other with red to achieve a more film look. Very sharp lenses on a digital sensor are more video than softer lenses in my opinion (this on a computer or hd/sd broadcast monitor). Although some dop told me red footage shot with sharper lenses is better once projected on film.

havent tested this.. so im talking out of my ass..
but doesnt digital to film soften the image a tad.. thus shooting sharper would get you to where you need to be?

 

[David Mullen ASC]

Yes, if you know you are going to do a film-out, which will add some grain and softness, you'll want to factor that in when shooting and in post.