Best LUTs for Weapon?

[Hrvoje Simic]

I can't argue with that at all. If you've come up with a specific customized LUT intended precisely for the way you shoot and the material you've shot, that's perfectly fine.

What I'm referring to is not designed for the way of shooting but to improve the camera's initial results regardless of it, treating a consistent part of its recorded light response. LUTs for specific way of shooting are much more limited in usability and designed differently. These are designed to work in general and the process of designing them is much more tasking.

 

[Björn Benckert]

I can't argue with that at all. If you've come up with a specific customized LUT intended precisely for the way you shoot and the material you've shot, that's perfectly fine. And if a DP has shot with a LUT in mind for a project, I always have the ability to pull up that look. From there, we can choose to either throw out that LUT and create the same look a different way, or we can just use the custom LUT and move on.



I'd argue that the image isn't optimum because it needs a window. The wall behind the person is too bright in relation to the person, and we need to direct the audience's eye away from the wall and towards the guy. The LUT won't give you a window. I also think the blacks are borderling muddy, but that's a subjective call.

Remember Brian's original question:

Hey guys, I am curious as to what everyone's preferred LUTs are for Weapon in camera and post?

I would have three answers to be reasonable:

1) just use DragonColor and DragonGamma4 for dailies, which is very simple, one-click and done.

2) or start with DragonColor or RedLogFilm in the color-correction program of your choice and create your own look.

3) or invest in Resolve and try one of the 50 or so LUTs BMD gives you for free with the program.

I personally think there are better ways to make good pictures than using a LUT, but if you like the picture, that's all that matters. Just bear in mind you will run into situations where the LUT looks bad, and sometimes it'll look bad in a way that can't be easily undone. You'll have more control over the look if you do it all yourself.

The image above... read my post again.. its a VFX plate. Im standing there with a smoke charge in my hand and move it around to match the movement of a girls hand in a shot from the edit which we do VFX for. That shot is played out on a iPad above my camera screen So I do a little hand driven match move with my hand to match another plate. The chimney the roof and the rest has nothing to do with anything. Im filming a layer of smoke agains black. Or the most black I had at the time which was they sky above my summer house at night, as that was the only place I could fire a few smoke charges without getting in to much trouble. and did not have to much stuff on the horizon behind. Midnight sun is a bitch here it needs to be a cloudy day and then we get about one our of half darkness before the sun goes up again making the skye all pink.

So in this case there is nothing to be done in terms of grading. Just cut out the smoke and match the smoke to the backplate and use a additive "Punchback key" and match with the smoke of the back plate that was shot down in paris on a roof where we could not have to much smoke charges due to the hunt for terrorists.

So again having luts in camera has nothing to do with what you talk about, windows, hold out mattes etc. yes those things I think most people understand is something that is far easier to deal with in post instead of gaffer tejp a computer with a live feed to the camera and try to live track those things in. Thats not what this thread is about so just leave it for some thread that actually discuss post production grading. This is not such thread.


So shooting with luts in camera is really cool everyone should try it. Then when you get to your grading session its even cooler as then the colourist has a lut in the clip folder and he can choose to interpret the footage as the DP made it with luts and everything or adjust iso, kelvin etc, and ofcourse also toss everything away and start from cameraRGB and build his own grade. Luts in camera does not limit anything its a cream on top of the cake that you can choose to take off at any time from applying it to mastering your grade in the grading session.

Further more if you are a DP and want some more stylized look on your ref monitor you can hire a on set colorist that can do live grading for you or, you can get pretty freaking far with some cool stylised luts and then use fool control or the in camera controls for looks and stylize the look of the footage. This look is if wanted applied on your progres files so as A DP you now got far better options than before to control both your client screens and your proxies.

Actually you got far more control than the alexa guys they always had luts but not like this not in a stack of red color science, red look control (curves etc and the rest of the stuff that is both in camera and RCXP) and then now with weapon also you can cake on a lut on top of it.... So for the first time it´s not a single layer of things.

Much like when filming film, you pick your negative, you lab it and you strike your positive... see same amount of steps there, and with red / weapon you can be a dumb ass not knowing anything and still manage quite well as you see all three steps at all times if wanted. Where as in with celluloid.... you pick your film, did not see jack shit while exposing, developed your neg without knowing if it was better to push it or not... got your neg and could do the R,G,B color timing and choose your positive stock and then see the result... it took quite a bit of experience to do the right decisions in all those steps as it was done blind folded. So having this shit popping out of the touch screen controlled via an iPhone is quite ground breaking and make things instant which it has not really been before.


Again, we are derailing the thread.

 

[Hrvoje Simic]

Björn, the opening line is unnecessary and the opposite route from the suggested one. I propose re-considering that one, as well as the input LUT for Marc's words. Otherwise there is a good chance for clamping, clipping, posterization and noise.

 

[Björn Benckert]

Björn, the opening line is unnecessary and the opposite route from the suggested one. I propose re-considering that one, as well as the input LUT for Marc's words. Otherwise there is a good chance for clamping, clipping, posterization and noise.

Sure, took that first line out.

But Hrvoje, explain how would it be a good chance of clamping, clipping and posteraziation to have a lut in camera. Thats like saying changing the raw settings is risking to ruin the image.. its not they are all metadata, just as the lut. Sure the recorded proxies are getting that lut burned in but thats the same with iso curves kelvins and the rest. And sure a small lut with heavy alterations will cause banding and clipping and crunching long before the source image break. But I think it´s not to far fetched for DP´s to understand how it works as most DP´s used luts for a long time when shooting ARRI cameras. The RED system now works perfectly the same, with the difference that its better, got more options and more control for handling the metadata and it´s not like arri that still have cameras that can not record raw.

So again it´s metadata, and there is no risk of harming the material. And as always check exposure, goalposts, and stoplights and you be fine. Luts will not break the image they will just end up in the sidecar.

 

[Hrvoje Simic]

Sure, took that first line out.

Thanks, cheers.

But Hrvoje, explain how would it be a good chance of clamping, clipping and posteraziation to have a lut in camera.

That was an analogy in the context of suggestion.

 

[Björn Benckert]

Thanks, cheers.



That was an analogy in the context of suggestion.

To cryptic, Think my english is failing me or atleast I don't understand.

 

[Hrvoje Simic]

To cryptic, Think my english is failing me or atleast I don't understand.

Input LUT = interpretation,
Image chain = discussion

 

[Marc Wielage]

As far as I understand it, he created a technical LUT that generally improves camera response - similar to what LI is currently doing at PV. This brings the footage into a better starting place, it's not specific to the way he shoots and/or specific shots - and it's not a look, it improves the captured data. Everything else comes afterwards.

That can work in theory, but I'd have to see the execution. The Light Iron people are pretty bright and do great work.

Björn, the opening line is unnecessary and the opposite route from the suggested one. I propose re-considering that one, as well as the input LUT for Marc's words. Otherwise there is a good chance for clamping, clipping, posterization and noise.

Very well-said. The issues of clamping, clipping, posterization, and noise are huge ones when you throw a random LUT on an image. You might have to sift through 20 or 30 different LUTs in order to find one that works in all ways. In that amount of time, I could dial in a custom setting without any problem in 1 minute or less.

 

[Björn Benckert]

That can work in theory, but I'd have to see the execution. The Light Iron people are pretty bright and do great work.


Very well-said. The issues of clamping, clipping, posterization, and noise are huge ones when you throw a random LUT on an image. You might have to sift through 20 or 30 different LUTs in order to find one that works in all ways. In that amount of time, I could dial in a custom setting without any problem in 1 minute or less.

I dont think people shooting arri cameras agree as they all use luts. And normally they dont pick them randomly.

Just as having luts and shooting raw on the alexa we can now shoot raw on red get the used lut in the side car and have proxies with applied lut for offline edit. And no just as you really dont need 20 versions of red gamma to get different looks you can use one and adjust your curve iso and the rest underneat and you need to push things quite far to make a subtile 33x33x33 lut break the image... and even when thats done just as on the alexa you still got the raw file so no harms done. The whole operation can be done to get closer to what the dp envisioned on the client screen.

Here in sweden there is a trend of using on set grading. Basically an advanced video assist that grades the videotap of the camera and toss the same grade onto the dit proxies. To me thats a bit to far from camera and grading the 422 2k proxies and video tap is far worse than doing it in camera on 16 bit with a 33x33x33 lut. And further more dialing it in in camera gives instant proxies and no need for the rig that on set grading demands. All you need is your favorite luts and an iphone and then you can do it all with foolcontrol. No lut boxes needed for client screens and no realtime colorcorection needed.

 

[Mike Nagel]

I dont think people shooting arri cameras agree as they all use luts. And normally they dont pick them randomly.

Just as having luts and shooting raw on the alexa we can now shoot raw on red get the used lut in the side car and have proxies with applied lut for offline edit. And no just as you really dont need 20 versions of red gamma to get different looks you can use one and adjust your curve iso and the rest underneat and you need to push things quite far to make a subtile 33x33x33 lut break the image... .

first off, the reason Arri shooters use LUTs with the footage a lot is because Arri readily provides any kind of transforms that are needed on their website - something that has been requested from RED for a long time... and with Arri footage, after these simple transforms the footage is often good enough - because it is very neutral after debayer - that a lot of folks don't do much more... fast workflow for good looking footage, hence why a lot of productions use the cam... especially in TV productions where time is of the essence

second, u can very, very easily break an image with a 3D LUT... and by "break" we mean introduce noise, banding, posterization etc...

but, if u do what u seem to refer to most of the time, which is only adjustments to the 1D LUT - ergo greyscale only, the heart of the 3D cube - then it is much, much harder to break the image... but that is a 1D transform (luma, gamma, tonal curve adjustment) within a 3D LUT...

if u actually use the 3D LUT for color transforms (--> any cube point with a saturation value > 0) in 3D space, u can break the image very easily...

if u have broad color (--> 3D) adjustments, it's usually not a problem... but on pin point isolated color transforms of a small range of colors mapped quite a bit across the gamut, u can see some funny stuff very quickly...

some apps have better math than others, so when u make ur color adjustments by creating a 3D LUT these apps try to avoid noise and posterization etc as much as possible... but no matter how good the math is, u still can break the image...

and especially - as I stated before - on video footage that plays back at 24 fps... a mere still image is way less of a problem...

u can visually check the quality of ur LUT with gamut maps, but the real test is to play back the footage

another "problem" of these 3D LUT apps is that (in a best case scenario) u obviously should use these apps on the calibrated grading display in ur grading suite (--> same signal path as ur grading app... same output card etc) ... but the apps don't support output cards and/or can target ur ref screen... so u end up creating LUTs and make color adjustments on a laptop or computer screen, which is obviously idiotic for color critical decisions...

 

[Björn Benckert]

first off, the reason Arri shooters use LUTs with the footage a lot is because Arri readily provides any kind of transforms that are needed on their website - something that has been requested from RED for a long time... and with Arri footage, after these simple transforms the footage is often good enough - because it is very neutral after debayer - that a lot of folks don't do much more... fast workflow for good looking footage, hence why a lot of productions use the cam... especially in TV productions where time is of the essence

second, u can very, very easily break an image with a 3D LUT... and by "break" we mean introduce noise, banding, posterization etc...

but, if u do what u seem to refer to most of the time, which is only adjustments to the 1D LUT - ergo greyscale only, the heart of the 3D cube - then it is much, much harder to break the image... but that is a 1D transform (luma, gamma, tonal curve adjustment) within a 3D LUT...

if u actually use the 3D LUT for color transforms (--> any cube point with a saturation value > 0) in 3D space, u can break the image very easily...

if u have broad color (--> 3D) adjustments, it's usually not a problem... but on pin point isolated color transforms of a small range of colors mapped quite a bit across the gamut, u can see some funny stuff very quickly...

some apps have better math than others, so when u make ur color adjustments by creating a 3D LUT these apps try to avoid noise and posterization etc as much as possible... but no matter how good the math is, u still can break the image...

and especially - as I stated before - on video footage that plays back at 24 fps... a mere still image is way less of a problem...

u can visually check the quality of ur LUT with gamut maps, but the real test is to play back the footage

another "problem" of these 3D LUT apps is that (in a best case scenario) u obviously should use these apps on the calibrated grading display in ur grading suite (--> same signal path as ur grading app... same output card etc) ... but the apps don't support output cards and/or can target ur ref screen... so u end up creating LUTs and make color adjustments on a laptop or computer screen, which is obviously idiotic for color critical decisions...

No image is not broken as we are talking about meta data here. And all such can be used later or tossed away and replaced with better data at a later stage like in a grading room.

Im not using 1d luts but 3d luts and underneat that lut i got the in camera 1d transformation tools like saturation gain iso color temp etc. And making proxies in camera has little to do with an exact color grade still out of all the dp's i work with most want best possible screens on set and black tents for them to be viewed inside. And further more alot of them are frustrated to show end clients an image that is to far from where they want it. For long most of them use dit's and on set colorists that brings lut boxes that is applied to the videotap. That is quite a problematic setup compared to what we now have with lut option in camera where the lut is applied on 16 bit signal and not 8bit video tap.

So what you are saying is that somthing that has been standard procidure for many, for a long time and now is improved / built into camera and can be applied on full bit depth is a bad thing and should not be used...

Sorry but i do not see the logic in that reasoning.

And regardig the apps used. on set colorists usually use resolve and got pretty decent screens and its not the difference between two graphic cards that the dps are normally stressing about they simply want an appealing image on the screen. It does not have to be uber accurate they know it can all be changed later they just want to show the director, the agency guys and the clients what their aim is. Then later its graded and very rarly its even landing close to what was showing on the screens at the studio. Still this process is very much in demand and the better it gets and the easier it can be done the better, no?

 

[Mike Nagel]

No image is not broken as we are talking about meta data here. And all such can be used later or tossed away and replaced with better data at a later stage like in a grading room.

Im not using 1d luts but 3d luts and underneat that lut i got the in camera 1d transformation tools like saturation gain iso color temp etc. And making proxies in camera has little to do with an exact color grade still out of all the dp's i work with most want best possible screens on set and black tents for them to be viewed inside. And further more alot of them are frustrated to show end clients an image that is to far from where they want it. For long most of them use dit's and on set colorists that brings lut boxes that is applied to the videotap. That is quite a problematic setup compared to what we now have with lut option in camera where the lut is applied on 16 bit signal and not 8bit video tap.

So what you are saying is that somthing that has been standard procidure for many, for a long time and now is improved / built into camera and can be applied on full bit depth is a bad thing and should not be used...

Sorry but i do not see the logic in that reasoning.

And regardig the apps used. on set colorists usually use resolve and got pretty decent screens and its not the difference between two graphic cards that the dps are normally stressing about they simply want an appealing image on the screen. It does not have to be uber accurate they know it can all be changed later they just want to show the director, the agency guys and the clients what their aim is. Then later its graded and very rarly its even landing close to what was showing on the screens at the studio. Still this process is very much in demand and the better it gets and the easier it can be done the better, no?

fully agree, 3D LUTs for on-set preview are mega useful. For technical application (calibration etc) and/or color space transforms they are also as good as it gets.

For look creation in post (--> not replication for VFX, but actual creation) - which I think is what Marc implied - they are very powerful but limited and I would rather encourage to explore the actual potential in the raw file first for a given shot, but if it works for u and u slap on a LUT look at the end of the chain, so on top of the base grade - as per your own example of ur colorist - then sure, if you as the client like it, u're done ;-)

I just think that some peeps in here are looking for a magic LUT for RED footage and this is not how it works... unless u have lower standards or no clue what the actual potential of the shot is.

 

[Marc Wielage]

I dont think people shooting arri cameras agree as they all use luts.

Naaa, I've done 50 or 60 Alexa projects and we used no LUT at all -- just Log-C and turning the knobs. Very simple, very straightforward. There was a look​ there, but it was one that I created with the input of the DP. Their temp look for dailies was just the normal Arri Log-C -> Rec709 LUT, and you can duplicate this easily in any good color correction program.

I don't have a problem with people using the Arri LUT Utility to create custom looks, but generally that winds up just as a temp for dailies and editing. And it has no bearing on the Red Weapon, which was the original question.

I just think that some peeps in here are looking for a magic LUT for RED footage and this is not how it works... unless u have lower standards or no clue what the actual potential of the shot is.

That is exactly my opinion.

 

[Björn Benckert]

Their temp look for dailies was just the normal Arri Log-C -> Rec709 LUT, and you can duplicate this easily in any good color correction program.
.

Main point here, the trick is not to have to use the color correction program to render proxies. But instead have dailies baked straight onto the SSD of the camera, with no additional rendering needed. And I´m quite positive that nobody here thinks that we can create a look in camera that is better or impossible to recreate in post.

 

[Marc Wielage]

Main point here, the trick is not to have to use the color correction program to render proxies. But instead have dailies baked straight onto the SSD of the camera, with no additional rendering needed. And I'm quite positive that nobody here thinks that we can create a look in camera that is better or impossible to recreate in post.

I think it depends on the creative intent. If the DP wants color-corrected dailies, and if the production is willing to pay for it, it's not a big deal to do that immediately following (or during) the shoot. Otherwise, I think the look out of the camera with just DragonColor/DragonGamma looks perfectly acceptable a lot of the time, certainly enough as a "ballpark" look to get you through. The moment you start coming up with complex looks and throwing those at the picture, to my way of thinking all this is going to do is just add more time to the project. If there's enough time for that, just have a colorist sitting off to the side or in a truck or down the hall from the shoot, and then instruct them on how the shots should look. At the worst, they can make sure every shot in the same scene will match, and that helps a lot during the edit.

As I've often told DPs, "this is the worst your image will ever look," and they understand it will get better in post -- with their input. I have seen several DPs relieved to see that, unlike the dailies, we were able to squeeze more detail and more dynamic range out of the picture than the LUT they had been using on set. One reason is that the colorist can dynamically react to sunlight changes, a light getting moved, one character who needs more light than the others, an unexpected lens flare, and so on. A LUT cannot do that, because it can't anticipate changes on the set or a specific creative mood that the filmmakers want to create that they didn't think of on the set or hadn't considered prior to the session.