Absolute Highest Quality Workflow if Time Irrelevant?

[Anson Fogel]

The question is - if quality matters and render/edit time does not, what apple workflow and applications get the sharpest, best end result? For my personal project, I've got time. Here are the details:

4K Capture, of course, with some 3K and 2K when forced for overcrank
Shooting remote location exteriors with a lot of movement and detail, all natural light
Output target BluRay for projection, and Dcinema large screen projection - probably 2K direct from a PC SDI or DVI
I have a calibrated 1080 SXRD projector and other displays in my post room, and I experience significant differences between well done commercial BluRay releases, and various output strategies from the various Apple/Red applications. Some of the questions:

Red Alert and Red Cine, red Laert v. redcine v FCP native render ( are they the same engines?)
full debayer renders to PR422 then into FCP>Color>Comp?
Native R3d injest/edit FCP > Color > Compressor > DPX?
Native R3d injest>FCP>Color>FCP>Crimson>FCP>Render>

And so forth. I've trolled these forums for hundreds of hours, done a lot of testing with my RED and the various apps - but no hard conclusions yet. Most posts are concerned at least to a degree about time. I'm not let's assume. What gets the absolute best output quality?

 

[Dan Hudgins]

Quality...

Quality...

The question is - if quality matters and render/edit time does not, what apple workflow and applications get the sharpest, best end result? For my personal project, I've got time.

TIF hold a little more tonal detail than DPX, and so can maybe get fewer Histogram gaps in the final result (4K filmout?). Every time the curves get bent then saved you can get some spread in the tonal values, so using 48bpp files makes the tonal values jump around less, both in the grading and the film recorder/projector etc. gamma adjustments.

There are PC options that can cost less and might give 1st class results. With no cost licence software you can run 20 or 40 computers in parallel for speed and quality at the same time, as well.

 

[Loyd Bateman]

Cool, so what would the cheap pc solutions be??

 

[M Most]

What gets the absolute best output quality?

Having the best cinematographer shooting the material, and an experienced colorist doing the color correction.

Seriously. Comparing high end feature work to material shot by cameramen with less experience, less equipment, and less time and blaming the differences on conversion and color correction software is not really being fair. Digital image manipulation is not rocket science, it is math. The same math is being applied in a DaVinci Resolve as it is in Shake. The differences are in the degree of control being given to the colorist, and, more importantly, the experience and skill of the colorist. But you must be starting with a pristine image to begin with, and that is determined not by the basic camera settings, but by the eye and skill of the cinematographer, as well as the equipment available to the cinematographer (I'm talking lighting and grip packages here, not necessarily camera). Not to mention the contributions of a good production designer. The truth is, what you're looking at as the source of the differences is likely the smallest contributor.

 

[Mick van Rossum NSC]

Having the best cinematographer shooting the material, and an experienced colorist doing the color correction.

Seriously. Comparing high end feature work to material shot by cameramen with less experience, less equipment, and less time and blaming the differences on conversion and color correction software is not really being fair. Digital image manipulation is not rocket science, it is math. The same math is being applied in a DaVinci Resolve as it is in Shake. The differences are in the degree of control being given to the colorist, and, more importantly, the experience and skill of the colorist. But you must be starting with a pristine image to begin with, and that is determined not by the basic camera settings, but by the eye and skill of the cinematographer, as well as the equipment available to the cinematographer (I'm talking lighting and grip packages here, not necessarily camera). Not to mention the contributions of a good production designer. The truth is, what you're looking at as the source of the differences is likely the smallest contributor.

Mike, with all respect, that does not seem to be the question. My interpretation is that with a given shot, what will give the best results going through all kinds of post possibilities. I just spoke to a post guy who has all kinds of equipment, including Color, and is experienced in Scratch. He actually liked the FCP L&T more than Redcine.

 

[Anson Fogel]

Yes Mike - while I fully understand MMost's point, your last line strikes at what I'm looking for - all other things being equal, what quality differences are people seeing between the various common Apple post tools and combinations. I've heard numerous comments to the effect that FCP L&T is getting noticeably cleaner results than Redcine. Is that true for Red Alert? When getting those results, what specific workflow was used?

 

[Uli Plank]

FCP is only doing a half (high) debayer.
REDAlert/REDrushes can do a full debayer at the expense of very long rendering times. IMHO only needed for very critical detail or low-light.

 

[M Most]

Mike, with all respect, that does not seem to be the question. My interpretation is that with a given shot, what will give the best results going through all kinds of post possibilities. I just spoke to a post guy who has all kinds of equipment, including Color, and is experienced in Scratch. He actually liked the FCP L&T more than Redcine.

Well, to some degree, it does seem to be the question. I quote:

I experience significant differences between well done commercial BluRay releases, and various output strategies from the various Apple/Red applications...

That statement implies that a comparison is being made between high end, professionally shot and posted material, most likely originating on film, and posted and compressed for BluRay by the best facilities in the world, to personal or client shot material on Red. So I think what I said did apply.

As far as different conversion software available for Red material, it is not a matter of "liking" one more than the other. The issue is understanding what each is doing, how it is doing it, and applying user settings that optimize that particular path. If you are using, for instance, Log and Transfer, you are not applying color correction (unless you have done custom settings in Red Alert) and thus not affecting the conversion parameters - as you probably are if you use, say, Redcine. Any and all color correction settings are to some degree destructive, so a conversion done using, say, Redline (or Red Rushes, same thing) that only alters the Exposure parameter, uses the Redspace color matrix, and the Redlog gamma LUT, is going to yield a different result than using Redcine and adjusting color balance, given that all other parameters are set to be equal - which in itself is difficult to achieve if you don't understand every one of them. That includes the debayer setting, but it also includes noise reduction, OLPF compensation, the resampling algorithm (i.e., Lanczos, Mitchell, Catmull-Rom, etc.), and any scaling that's involved. And finally, there is of course the simple fact that Redcine uses the GPU for many of its calculations, unlike the other options that use a software engine. The bottom line is that excellent results are available using any of these options, but you need to understand how to properly optimize each one. If you use them all on default settings, you are not comparing apples and apples.

 

[Barend Onneweer]

He actually liked the FCP L&T more than Redcine.

Which might say more about the experiences of this particular person than the potential of the tools. Which sort of ties in with what Mike is saying. But as long as we're talking optimal image quality, debayering in FCP isn't the best solution.

Bar3nd