Welcome to our community

Be a part of something great, join today!

Register Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

FLUT Science...

New color sience that matches the false color with the lightmeter, that's fantastic news.
I always have fights with the light-department about reds sensitivity and how to light it properly without being able to use the lightmeter the way it was ment to be.

That's why most pros say RED is a 160 ASA camera.
That will help bigtime.

Looking forward to the new stuff. It's gonna be a great year.

And yeah, Jim get some rest. It's christmas time for god sake.

Tim
 
Roberto Lequeux said:
Final LUT :)
Click to expand...

final LUT Pro? :)

but realisticly a final look is NEVER the look you shot on set. There is always room to improve the image and all films will still go through a grading process. The LUT system should give a consistent, but not final look, through production.

Even anything shot now, on film video, red, genisis, viper or whatever will be graded on delivery.

If the look a DP intended is critical they they will attend and supervise the final grading and be paid to do so. It's hard to imagine wanting to get your EXACT final image at the time of shooting.

Brook is right.

FLUT should equal consistent monitoring throughout the chain.

But how do you control color when such a wide variety of cheap tools and monitoring options are available. How do you keep an image consistent between a RED LCD, an EVF, an HD SDI monitor, HDMI signal and whatever the Professional grading facility has to offer? I doubt FLUT will do that because how is it going to be able to predict all of those standards.

I think the best we can hope for is reasonable continuity because this is the point of RAW. You still have to know your tools and have the flexibility to create your own sauce.

David
 
It's nice to be able to adjust an image with a DP, point to an accurate monitor and tell them that what is on that monitor is what will show up in post... period. Sure they'll be able to change it, massage it or get back from it, but when the footage leaves the loader's hands and the LUT leaves the DIT's, what we have created on this screen is what everybody will see. No, the phone won't ring because somebody processed in their own favorite color space or somebody "made a look!" on a laptop. No, the studio executive won't pull out his hair and fire you because dailies will be cranked from REDspace when the movie is supposed to be bleach bypass. This is the picture.

With RSX-based workflows, you cannot say this. With products like 3cp, you can. I'm hoping FLUT is somewhere in the middle.
 
Brook Willard said:
Took it for a run with the best possible intentions. :) While the output on the camera will always be one possible interpretation of RAW, being able to monitor with greater color and exposure accuracy [both of which are assumptions or inferences] will be of huge benefit to more HD-saavy DPs who are used to cameras like the Genesis and F35. If you thought I ran with it last time... here goes.

As a DIT, I'd love to be able to combine elements of different cameras.

When I do HD shows, I hate all the cable, tent, zillion dollar calibrated monitors, tape decks, etc. I have mixed feelings on live painting, but I do like things like the GDP that allow me to put an accurate, reproducible LUT onto the image. I don't like the wiring and nonsense surrounding it, but the simple concept of an accurate, reproducible and specific LUT is marvelous.

When I do RED shows, I love the simplicity of the camera setup, the untethered nature and the freeform style that stems from shooting RAW [light it, frame it, expose it, shoot it... do the rest later]. The in-camera painting [LUT preview, if you will] is limited by design, which is a mixed blessing. I love not having to use it [I always recommend shooting straight REDspace and grading afterwards], but I've been on shows where it is requested and there's no other workflow.

The bummer of that workflow [other than the difficulty of using the joystick to generate a LUT - something you guys have tackled with selectable looks on the RED ONE and the future options with the Epic] is the inherent inaccuracy. It's great as a broad-strokes tool in situations where people either want a basic LUT generated or the creative parties [DP/DIT] will be involved to the end of production and can have a final say in color correction... but there are situations where it doesn't work.

Since it lacks a "true" black point or a configurable white point [REDspace's shoulder when compared to linear, for example], it can be tricky to use for delivery [something that is often requested]. It also has a fairly coarse curve and basic color settings. The settings are also controlled on a computer monitor [instead of a truly calibrated broadcast monitor with hardware scopes] and is subject to all of the associated inaccuracies like gamma, viewing angle, 6-bit laptop displays, environment, etcetera. [note: while broadcast monitors are subject to some of these same obstacles, people tend to treat a $15,000 HD-SDI monitor differently than a MacBook]. Also keep in mind that different applications interpret the metadata/RSX/LUT/etc. data a little differently [particularly depending on what version of the SDK they're built around], sometimes resulting in potentially catastrophic gamma and color temperature shifts. Don't take this the wrong way at all, it makes perfect sense and is fantastic for what it is designed for! The bummer is that people hear "look on set" and "free" and run with it a little too far sometimes. It puts me in a tremendously difficult position as a DIT, because I'm asked to generate a LUT that may cause more problems than good if it's not handled properly... and I lack the types of control and accuracy checking I'm used to!

This type of approach works great on some projects as I've mentioned. The problem is when people who are used to another kind of on-set LUT expect the same iron-clad results... and they expect it for free on a laptop without the big equipment! On a commercial it's very tough, because the material leaves the DP's hands on the day with either no look or a relatively vague LUT that can be misinterpreted due to some of those variables. Commercials sort of are what they are though.

I've been involved in longer narrative projects [some very recently] that have elected to use that approach and in each case they had MAJOR post issues regarding color and communication from set to post. Now like I said in my post a few days ago, if you go into a project with the right knowledge, experience and workflow, you won't have a problem! The problem comes from people trying to stretch a workflow outside of what it was designed for... and expecting results/accurcy that can't be generated on a timeframe that wants yesterday on a budget that doesn't exist. :)

My current solution is to shoot in REDspace and generate accurate, reproducible LUTs on set [on a calibrated HD-SDI monitor using hardware or software scopes] using Gamma & Density's 3cp. REDspace serves as a terrific preview LUT at the camera and people accept it as a really awesome video tap. The real color action happens at the cart after the image has been shot. Each camera gets calibrated with a base LUT to remove any variation between sensors and very detailed LUTs are generated from within 3cp. You can create very complex looks on set and communicate them effectively, accurately and nondestructively to post. It can generate LUTs for just about any DI system, so it's all "standard" based and just sort of works. I've had wonderful luck with it

I've used this workflow to generate some very dramatic looks for certain shows. We used it on the second season of Southland [episodes will air on TNT in a few months] as an example. We actually expose "perfectly" at the camera most of the time. All of the harsh, gritty, desaturated color look comes from 3cp on set. Since that show was so freeform, we couldn't just use a few coarse LUTs and rely on them. Every setup was different and had a different challenge or goal. The sky, the lighting, the sun, the uniforms, you name it. The curve manipulation we did on that show [on set] was immensely detailed to get the look we intended. And at the end of the day, it saved us tons of money in post because we were able to bypass a supervised final color correction session because our looks were so effective on the day.

I've had success with that system on shows that have needed looks ranging from bleach bypass to saturated and sunny. It's pretty cool, and it just works. But my favorite part? No cables, no tethering, no tent, no live painting, no crap! We can shoot a show the way we used to do it with film and just spend a few minutes through the day dialing in a detailed LUT. And when it gets sent off with the "film break," it's nice to just know that what I saw on my monitor will show up at the post house, in editorial, on the DVDs and - eventually - on the screen.

[plus, since it works well with just about every digital camera out there including the Genesis and GDP, and has the means to calibrate a still camera to behave like Kodak and Fuji film stocks, it's a great tool for look management on multi-format shows... which so many shows are these days]

That's been my current approach to blend what I like about working as a HD DIT and what I love about working as a RED DIT. Take the pros of each and build a system that works. But the truth is that it doesn't need to work like that for every show. While I think longer narratives [features, television] can benefit greatly from that sort of workflow, shorter projects don't have the time to trust a relatively new workflow. They want a solution that works yesterday for free!

That's why I'm excited about what you're talking about. It sounds like a system that will bring some of the accuracies so many DPs and DITs are used to from the HD side of thing over to the RED side in a way that doesn't exist without stepping outside of the built-in RED workflow. These are all assumptions, so if I'm completely wrong... just take them as requests. But if I had to guess about what the purpose of FLUT [or perhaps the unnamed look system for the new cameras] is, I could only assume that it's designed to build upon what we already have. So to me, that would be a universal look that can travel from camera to computer and back again. It'd be a LUT that can stick with footage all the way through post. It'd be a LUT with more detail, more parameters and more consistency through the hardware and software. It'd be something between what I can do with RSX files [not a DIT's friend] and what I can do with 3cp. Knowing you guys, it'll probably be controllable via wifi from somewhere on set.

If it's anything like what I'm hoping for, it'll empower those shorter or less-budgeted shows with more accuracy and specificity! That's what I mean by combining the best elements of different cameras.

Even if I'm completely off the deep end with my assumptions [which this post is full of], take them as recommendations or merely the hopes of a RED and HD DIT who's been around the block with this camera a few times. :) :thumbup1:

Now if you'll excuse me, it's 4:11 in the morning and I've just done another one of those big Brook braindump posts. :)
Click to expand...

Right on. I work in a similar way with my RED DIT. I would LOVE more control over LUT curves (more points on the curve, etc). Ideally, I could sit at the computer and tweak the LUT and it would update wirelessly and instantaneously on the camera. Similar to a paintbox, but wireless and non-destructive due to the RAW workflow.
 
Jim -

If you guys have the next REDucation in Feb, I will definitely be there. I will bring some RAW footage to play with your new post creatures if it's okay. I'm so deadly excited that I might have higher blood presure than you do when playing Red :)
 
I don't understand you Brooks. You want more flexibility but you want to have it be completely fixed like Genesis? That doesn't make sense.

If you ask for a fixed output to add luts with a davio, cinetal, etc. that means removing all the options and flexibility. Does not compute how more controls would help that.

I think what you are instead asking for is to load up a 3D lut in the camera. But I don't see how that will keep someone from changing that during post.

RSX works quite well for us imho to communicate looks from dit into camera through post. I find it better than a lut because the parameters are consistent and known. If you need to tweak something it's very easy.
 
I'd add that 3D Luts can also be pretty innaccurate and more destructive in grading because when they get processed with those luts in rendering, you're baking in before you get to the grade. Then you throw away flexibility from r3d to get the best image.
 
Brook Willard said:
It's nice to be able to adjust an image with a DP, point to an accurate monitor and tell them that what is on that monitor is what will show up in post... period. Sure they'll be able to change it, massage it or get back from it, but when the footage leaves the loader's hands and the LUT leaves the DIT's, what we have created on this screen is what everybody will see. No, the phone won't ring because somebody processed in their own favorite color space or somebody "made a look!" on a laptop. No, the studio executive won't pull out his hair and fire you because dailies will be cranked from REDspace when the movie is supposed to be bleach bypass. This is the picture.

With RSX-based workflows, you cannot say this. With products like 3cp, you can. I'm hoping FLUT is somewhere in the middle.
Click to expand...

brook.

I totally hear you on this but everything you descibe is a calibration issue. No matter how well you do your job (and I would hire you in a second) you can never possibly make that happen. There are too many varibles with color science compared to audio for example, where tone and levels are a universal standard.

I think that the service you provide is most likely on a job per job basis and when you are working with a new facility you meet with them and the dp ahead if time to work this out.

I do agree that the system can and will be improved but I also like that RED gave us many options. I think that it's up to RED to provide us with a best practices, but people will still always demand what us not actually physically possible.

David
 
Carsten Fenstermacher said:
I don't understand you Brooks. You want more flexibility but you want to have it be completely fixed like Genesis? That doesn't make sense.

If you ask for a fixed output to add luts with a davio, cinetal, etc. that means removing all the options and flexibility. Does not compute how more controls would help that.

I think what you are instead asking for is to load up a 3D lut in the camera. But I don't see how that will keep someone from changing that during post.

RSX works quite well for us imho to communicate looks from dit into camera through post. I find it better than a lut because the parameters are consistent and known. If you need to tweak something it's very easy.
Click to expand...

I should clarify a bunch of stuff. I left some things out as well.

First, the Genesis workflow is not dramatically dissimilar from the RED workflow [bear with me, this is largely for those not familiar with it]. You always record "raw" first off. Now with the Genesis, it's not literally RAW... but it's as raw in the literary sense. It's a flat, desaturated log image called PanaLog. Certain parameters affect it [color temperature, for example], but the general idea is to leave it very untouched.

The Genesis Display Processor [GDP] takes a HD-SDI feed off of the camera and - via its ethernet interface with a Mac or PC using a GUI - modifies that signal. There is the default curve [think REDspace] that converts the PanaLog "raw" image into something meant for human consumption [just as REDspace modifies the RAW RED image into a similar place]. You can modify that [high resolution] curve or create your own in a nicely detailed interface with relatively in-depth controls and lots of options. That LUT gets sent to the box which modifies the PanaLog signal it's receiving from the camera and outputs the "pretty" image to a monitor.

You can save that LUT and pass it onto post [email, thumb drive, etc]. Post can apply that LUT in the rack-mounted post GDP [finicky] or ingest it into one of a few apps that will interpret it and modify PanaLog footage.

So fancy boxes, tape decks, cables and nonsense aside, the concept is fairly basic. Shoot "raw," make a soft LUT, look at it on the day and send it to post.

The point of describing this is not to suggest that I think it's a superior system... I don't. It is a very good system, but I much prefer RED's approach for a laundry list of reasons. The GDP system still has bugs and inconsistencies... and is a pain to use when compared to a RED. My original attraction to RED years ago when it was in development was actually due to my dissatisfaction with traditional or nontraditional HD cameras.

I'm describing it to point out three specific benefits to the system:

One - it's nondestructive. It's not like a more traditional HD camera that has to be "painted" on the day before you roll or else. While it's not as convenient and truly RAW as the RED system for a variety of reasons, it is similar in broad concept.

Two - it's accurate. We're talking about HD-SDI feeds in and out, real waveform monitors and vectorscopes, truly calibrated monitors set to a "standard" [if you will]. It's a reproducible system. What shows up on the GDP's output will show up in dailies.

Three: it's detailed. It's a high-resolution LUT with lots of controls, multiple curves and fine control. Also, since the GDP is hardware based, it is possible to overcome some of the limitations of Panavision's gamma software by using a third-party application [3cp comes to mind, as it's what I use].

There is a "Three and a half" pro that I'll point out in a bit.

Now I want to throw a few disclaimers in here before I go any further. Don't read this post as me saying the Genesis is superior, the RED sucks, etcetera. There are lots of systems out there that have pros and cons and these two are no different. This could be misinterpreted as a skewed representation, as I've just discussed only the pros of the Genesis system and I'm about to discuss only the cons of the RED system. Do not interpret it that way.

While I will discuss limitations of the current RED system, I want to make sure that it gets taken as constructive criticism and not an attack. It's amazing what RED has accomplished in such a short time and unbelievable that they've been able to make it available for such a low price. I don't want to discount any of that. All of this is written with the best intent.

So to round this back to your original response, now I'll talk about what I mean by flexibility. There are two major points to discuss.

The current RED look system is made up of a certain number of RAW development settings that we can adjust in camera or in R:A as an RSX file. The controls are a little coarse not simply from a UI perspective but a point accuracy perspective as well. The curve has 5 X,Y coordinates that can be adjusted. Since much of the time the black and white levels are not adjusted, that leaves a three point curve with a very coarse [100 points] scale.

So, if we were to interpret the scale in linear values, it would mean that every point we move on the curve in one direction [say, lifting the mids one point on the Y axis] will illicit a ~40 value adjustment. That's a big jump!

Looking at the color side of things, we are afforded 25 points of tint in either direction [in the camera, admittedly it can be more specific in R:A]. Since tinting away from a green spike [the most common adjustment] tends to bring out noise fairly aggressively, a lot of people will jump to the RGB sliders that behave similarly to print lights. But with these sliders, adjusting a few clicks in either direction will make dramatic shifts in color. I often find myself stuck between two points when using either of these adjustments. It's either too green or too magenta.

Now, again, I'm specifically pointing out limitations of the system to help make my point here. From a RAW development standpoint, the options the camera gives us are fantastic! But from a creative LUT creation standpoint [we are talking about LUTs, after all!], all of the color and exposure tools are very coarse.

So when I say I want more flexibility, what I'm looking for is more control. More variables, finer adjustments, etcetera.


Man, this is going to be long.


Ok, so far I've talked about two things: some pros of the GDP system and lack of accuracy of control in the RED "RSX" or look system. There is another very important point to discuss about the RED "RSX"/look system: consistency.
 
This problem stems from two places. The first is technical. I'll start with generating a look in camera with the joystick.

In my experience with literally hundreds of RED bodies, I have found that there is some significant inconsistency in the in the HD-SDI output color accuracy. Even if sensors match perfectly, if you point 10 cameras [from different "batches"] at the same subject and plug them into 10 calibrated monitors, you'll see 10 slightly different pictures. Some will be a little green, some will be a little magenta, some will be contrasty, some will be flat, some will be brighter and some will be darker. It's minor variation, but it is noticeable.

I want to restate that what I'm doing here is comparing the consistency of a $18,000 camera to a bunch of $200,000 cameras. I understand it's ridiculous and I do not intend to fault RED for these shortcomings. It's all just a part of a discussion.

So let's say that a DIT on a shoot and our cameras have very well matched sensors [read: the RAW footage is very neutral and matches between cameras] but inconsistent video boards. The DP and DIT decide to create a "LUT" in camera.

The DIT twists the knob on each camera and selects identical values. The looks are matched. But when he returns to the monitors, one is a little green and flat and the other is nice and contrasty.

No matter, it's time to match the cameras! He goes back to the green/flat camera and dials in more contrast and less green. Let's even say he does it in the video menu instead of the tint, for example.

The result are two cameras that match perfectly on the monitors with one major problem: the footage will not match between the cameras when it is developed. Since the DIT dialed in more contrast and less green in one camera to overcome the video board's inconsistency, one camera's footage will be too contrasty and magenta.

Going a step further, let's say that the footage went off to a post house where they simply cranked out ProRes files from the RED footage using the camera's look and the RAW files will never be referenced again. Now the production company has baked in files that do not match and it will cost them time and money to undo it.

That is one point of technical difficulty with the system: inconsistency at the camera.

Another point to discuss: consistency between programs. If you open up a piece of footage in every RED app that can read it and acknowledge a camera look and generate an uncompressed TIFF from a specific frame, you'll find that there will be slight color variations [just tested it to make sure I'm not insane]. These differences are minor, but they grow as you reach further. Different versions of the SDK, QT gamma issues, etcetera... it can get pretty bad pretty quickly.

Now this is largely at no fault to RED. How can they control what version of software that people use? This will be an issue no matter what camera system you use, assuming that camera system has a flexible back-end. This is more a life reality than a system reality.

But whatever the cause [discussed more below], I have found more color inconsistency in patched-up RED post workflows than with any other camera I work with. There are some reasons I can point to, but I'll get to that in a bit.

So that's another point of technical difficulty: inconsistency within a system.

Another very important point to discuss: the accuracy with which we can view the data we're modifying. When we change settings in R:A, these data are presented to us in a limited way. We can view it in an application window on a computer with only a histogram available. As such, there is no proper way to view it with a waveform monitor or vectorscope [hardware or otherwise]. There is also no proper way to get it onto a broadcast monitor. The application window is also tied into the inherent inaccuracies and inconsistencies of a computer's DVI output and associated gamma [not a paralyzing issue, but present nonetheless]. Setting aside monitor accuracy [one must assume a person has a worthy computer monitor], there are a lot of limitations to the ways in which we can view image data and trust what we are seeing when we make critical image decisions.

So that's a third point of technical difficulty: inconsistency and inaccuracy of viewing.


Whew.


So with the camera look or RSX system, there is no monitor I can point to [be it out of the camera or on a computer] and tell the DP "this is what it will look like in post". That is a HUGE deal.

So to sum this chunk up, when I say I'm hoping for the system to take on some of the finer characteristics of the Genesis system [and other similar systems], I'm not saying that I want to plug a RED into a GDP - far from it.

What I'm interested in is an increased level of accuracy of control of the RED system as well as an increase in consistency and color accuracy from camera to pre-post to post.


So the flexibility of the RED system as it currently exists, but with finer controls, more consistent color reproduction and more accurate and trustworthy monitoring. In other words, I want the best of all worlds. :)


Now how can something as wonderful as what I'm asking for be free? I'm not looking for that - I'm mostly just explaining what the top of the mountain looks like today. The great news is that I have managed to build an on-set workflow that already does all of this [with the exception of in-camera looks]. Using 3cp, I've been able to do this. I can take RED footage, read it from the original R3D files, create a LUT using very fine controls and parameters, view it and check it on a calibrated HD-SDI monitor and send it to post as a LUT in whatever format the post house needs to work with their DI system. And when that information gets sent, I can trust that they will see in post what I saw on set. I love it.

But it's not free. I bring a lot of hardware and software to set and it costs money. This is a heavy-hardware approach to DITing and post production. This brings me back around to pro number "Three and a half" of the Genesis system.

It costs money. Lots.

Now why the hell is that a pro? Before people run with that and infer that I like things when they cost more money or that I'm some Hollywood union type, bear with me.

When things cost money, people respect them. They understand large metal boxes with blinky lights and scary knobs are not to be messed with. They know that the monitor they are looking at is way better than what they have at home because they're paying for it. When they look at the equipment, it is both unrecognizable and intimidating.

As such, it commands respect. People take it more seriously, they ask more questions, they trust those with knowledge and they want to learn more. On a big-budget movie or TV show, this is a good thing. It's comforting to see something big and scary running the ship, because it feels like it's all built on something a little more solid.

The opposite end of that spectrum is a MacBook Pro with a Digital Blasphemy desktop picture sitting on a pop-up table running a copy of R:A. :)

It is impossible to take that seriously.

Now I don't want to suggest that everybody needs heavy hardware or that the only solution is more cowbell. I also don't want to suggest that all of the work RED did to cram this technology down so that it'll run on a laptop is somehow to be forgotten. Not at all!

The second issue with the RSX/look system is not an issue of technology. In this case, it's an issue of perception.

I don't want to talk about the rental issues, integrity of the position or anything like that here. With a laptop and a piece of desktop software, people accept things they would never accept with a larger system. No shade? It's cool, I can still see the screen. Green lights overhead? Don't worry about it, it won't make a difference. No scopes? Don't really need those anyways, right? Inconsistent? Come on, so many people have told me it works perfectly this way. You don't think it's a good system? Well that's not your problem, is it!

Then the phone rings a week later when their LUT is the wrong color, wrong contrast and wrong curve. It looked fine on set!!! What went wrong?

In all honesty, it is so freaking cool that all of this can be done on a laptop. But sometimes people slather more and more variables on a device that's basically a giant variable to begin with and are surprised when it comes falling down.

If you tell a producer "we can do it cheaper and faster," all they hear is "free" and "instant". So they keep their expectations the same [say, the accuracy and consistency of an HD show with all the hardware and support trimmings], lower their budgets and tell you that you don't really need that gear because they heard it works fine without it and "it's just not that show". :)

It's not about somehow calibrating the world, it's about minimizing variables.

So no, I'm not expecting RED to somehow manage to do all of the things I'm discussing here and do it for free. Not at all! Systems already exist out there [systems that are not free] to bring this workflow to anybody... right now. It's tested, it works and it's way cool.

What I'm either assuming, asking for or recommending is that some of the cons of the RSX/camera look workflow improve. It seems like a given, of course RED is going to give us something better this time around! Look at all they accomplished the first time around with no camera-building experience!

So that's what I meant. Probably.


Moving on!
 
Last edited:
Carsten Fenstermacher said:
I'd add that 3D Luts can also be pretty innaccurate and more destructive in grading because when they get processed with those luts in rendering, you're baking in before you get to the grade. Then you throw away flexibility from r3d to get the best image.
Click to expand...

Very good observations. I'll give a few examples to clarify my position.

Regarding baking in looks, I agree that it is a bad idea. The good news is that the days of "baking in" looks on set are largely gone. With most of the modern HD cameras and "next generaton" RAW cameras like the RED ONE and D-21, many people choose to shoot in raw [log] or true RAW. I'm opposed to "baked in" looks completely. I think it's a terrible idea. In my mind, the days of painting on set and baking it in are gone.

So one resultant argument is that we don't need fine accuracy because it's a soft LUT that can be modified later. This is a good argument in many ways, but it overlooks the primary way that people work with the RED on smaller or shorter shows. The vast majority of commercials shoot RED, send the R3Ds to a post house and ask for ProRes or Avid files in return [with REDspace, a camera look or RSX applied]. They will edit with these files, do a final color pass on these baked in video files and deliver from there.

So even though the LUT that was created on set [RSX, camera or otherwise] is technically a soft, nondestructive LUT, it has the exact same effect as a baked-in look on the day. It's fixable if it's broken, but it costs time and money.

So generating an inaccurate, inconsistent or coarse LUT on the day often as detrimental as messing up a live painting job on set.

Now here's another workflow example: a show that will transcode to ProRes, edit ProRes but conform back to the RAW files for final color. That fixes the problem, right?

It fixes one of the problems, but not all. Yes, if the look is broken, it's fixable... and fixable for free because a final RAW pass was already in the books. But it causes a lot of problems in the mean time.

Dailies will have the ugly look on it. Executives and producers will be upset. The editor will have to stare at the ugly look for months, either falling in love with it and recommending it to the director or hating it and trying to change it. Either way, it can quickly become one of those "stepping on the DP's toes" sorts of things. These sorts of things can snowball on larger shoots and they can get people fired.

Now none of this is really appropriate for this thread. It's not a request for RED, it's not a prediction, etcetera. It's more of a state of the union glance at how things can go wrong when they go wrong. I think my point is in there somewhere though.


One final point I want to make is regarding what all this accuracy is for. I'm not talking about making final decisions on set. I'm not talking about spending the day in front of zillion dollar monitors tweaking knobs and breaking the filmmaking process. I'm certainly not talking about shot-to-shot color correction or the things that are best suited for dark, calibrated post houses and DI suites. I'm talking about fast, low-maintenance LUT creation here. I don't want to go back to the tent, I don't want to go back to the things that RED has helped us escape.

What I'm talking about here is increasing the accuracy and consistency of what RED has done for us. Finding a way to convey intent accurately. Finding a way to make a bleach bypassed digital movie actually look like it should. Having the control to push the highlights to 98% and not clip them. Things like that.

On Southland, we made some very interesting looks with such a system on set. As the DIT, I was able to manipulate the image with a lot of control and push it to its limits with confidence. And I did it while looking at a relatively inexpensive [<$10,000] field monitor under mixed lighting without a tent. The point is that I was getting enough accuracy to create the look and feel we were interested in. What I lacked in monitoring accuracy, I could always regain in scope accuracy.

Sometimes the LUTs we created were baked into entire scenes but most of the time we still had a dailies colorist double-checking our work in case we missed something because of the schedule [3-5 locations a day, 8+ pages, 2-5 cameras]. Nothing about this forces the DIT and DP back to the tent... I want to make that super clear. It's a wonderful new blend of both worlds. The flexibility and freedom of the RED RAW workflow combined with the detail, accuracy and consistency of a high-resolution LUT akin to the days of HD.


So to tie it into the previous points, increasing accuracy, control and consistency of LUTs on set [whether FLUT does it or otherwise] results in a few things. It means that the DP can truly communicate his intent to post, dailies and editorial. It means that the DIT has the control he needs to do his job and assure some level of consistency. It means that dailies will look right [within reason]. It means that editorial will look right [within reason]. It means that when the director is staring at the same editing system for 6 months, he falls in love with what the DP intended, not what the DIT screwed up. And at the end of the day, it means that it's all soft, undoable and nonpermanent.

But what if somebody from post goes in there and deletes or modifies the LUTs? Well, hell, there's nothing we can do about that. With great power comes great responsibility... or something like that. If we can't trust people to respect what we create, what are we doing here anyways?
 
I understand I've gone miles off topic. I'm excited for whatever FLUT brings and downright chuffed at what it might mean. That's just a glimpse into the state of things. :)

Best to all.
 
After reading some of you guys post leaves me a little loopey, DIT, LUT, FLUT etc. What do these acronyms stand for? Please elaborate in a short concise manner? hopefully I can get a clearer understanding. I will probably have to wait to get the Scarlet in my hands to fully understand.
 
You must log in or register to reply here.
Back
Top