Welcome to our community

Be a part of something great, join today!

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 :)

Black Level off??

Eric Haase said:
Why not just have the in camera look properly represented when loaded in to redcinex? You say it's to leave room for grading- isn't that why we're shooting RAW.it's just metadata anyway so if i then want to leave room for grading i can loosen the contrast, etc. Why is the default that it doesn't match? And why bother with all those settings and look files in camera- set it up and preview it on set but then load it in to rcx and have to do it all over again. Sigh.
Click to expand...

Thing is that if you make DPX's with this look baked, it is no more metadata. The same would be the case if you started off with a really punchy preset look in a RAW grading setting. The choices have been done, and all you can do is to further degrade the colometry, precission and latitude.

Yup. That's why we shoot RAW, but for that to be an argument, one have to have a pipeline from set to DI that takes care of all that data. If not... Well, it's gone. And that is NOT best done by dialling in a "finnished" look on a rather poor preview on set.

SO question is: What are your priorities on set and what are your priorities in post?

A firstlight isn't intended to make the image look "good" but to make it "gradeable". And you see a lot more issues once the images are recorded (which makes perfect sense, BTW), than you see on set, so I cannot really see how it would be wise to make final decissions on set the blatantly counteracts the flexibility of the RAW workflow.

Again - this is for for productions that will be going through a DI.

I think composition and latitude-handling maybe are the most important parameters on set.

far too often DPs doen't understand anything of the technical side of the signalchain for online. And why should they? It's NOT their craft. BUT then again, for that very reason - it is a bad choice to let the DP make final decissions on the online signalchain on set.

A good thing might be to make some stills with the right/intended looks off set, and let the online people handle how they get there?

It's anyway much easier to dial in a look in RCX or whatever on a well exposed image in post, than to make those looks in-camera and possibly end up with bad exposure which is really hard to get through post - as I have seen a lot of examples on.

Respect each profession, sort of...

But again:
This depends fully on what kind of production you are currently involved with...

(Gee I wish more people were involved in this debate!)
 
Gunleik Groven said:
Thing is that if you make DPX's with this look baked, it is no more metadata. The same would be the case if you started off with a really punchy preset look in a RAW grading setting. The choices have been done, and all you can do is to further degrade the colometry and latitude.

Yup. That's why we shoot RAW, but for that to be an argument, one have to have a pipeline from set to DI that takes care of all that data. If not... Well, it's gone. And that is NOT best done by dialling in a "finnished" look on a rather poor preview on set.

SO question is: What are your priorities on set and what are your priorities in post?

A firstlight isn't intended to make the image look "good" but to make it "gradeable". And you see a lot more issues once the images are recorded, than you see on set, so I cannot really see how it would be wise to make final decissions on set the counteracts the flexibility of the RAW workflow.

Again - for a certain kind of productions that will be going through a DI.

I think composition and latitude maybe are the most important parameters on set.

far too often DPs doen't understand anything of the technical side of the signalchain for online. And why should they? It's NOT their craft. BUT then again it is a bad choice to let the DP make final decissions on the online signalchain.

A good thing might be to make some stills with the right/intended looks off set, and let the online people handle how they get there?

It's anyway much easier to dial in a look in RCX or whatever on a well exposed image in post, than to make those looks in-camera and possibly end up with bad exposure - as I have seen a lot of examples on.

Respect each profession, sort of...

But again:
This depends fully on what kind of production you are currently involved with...

(Gee I wish more people were involved in this debate!)
Click to expand...

I want to dial in a look, hand it off and have it transcoded the way i made it look on set. If they allow me a Dit with a rocket setup on set i will do a best light dailies grade with a tangent wave. If not, i will dial a look in on set. But i want the dailies to look like what i want them to look like. I don't want them milky, desaturated, or anything else. I don't want an editor coloring the dailies. If you want to make dpxs for final grading then at that time you can make a flatter look and render dpxs. If you want to grade on top on the dailies let me know and ill leave room for it. Best to put color decisions in the hands of a cinematographer than an assistant editor or editor. I don't have enough fingers to count the times I've discussed color management with post houses involving red footage that they have had little or no clue how to properly represent what we are doing on set.

Example, i shot a recent commercial for a very well known (legendary) feature director. Production did not want to have a rocket on set to dial in grades. Talked to the post house and they were going to transcode the footage in redrushes with camera metadata, but go back to the RAW for final grading. No colorist would be grading the dailies. Where does this leave the look of the dailies? If I dial in a look in camera it will still look washed out in the dailies. Luckliy i have a great support team and my DIT brought his rocket setup free of charge. We dialed in some heavy looks and made some rmd files on set. The post house used the metadata we created with the rocket and rcx to transcode. I got a call from the director who left a message raving about the look of the footage. I got lucky on this one and am grateful to my DIT who has helped up the odds ill be hired by this director again. Flat washed out dailies leave a bad impression.

Often times, post houses will do whats convenient for them and not what is going to get the best look out of the footage or represent your intentions.
 
I see your points.

My counterpoint is simply:
The onset preview is and will be wrong as long as it is half debayer 8 bit 4:2:2 to a 720-ish monitor.

It's not really compareable to what you see once you fire it up in post.

OK. THAT is an issue, but it won't be resolved on the R1 anytime soon...

better then to make people remember how the dailies looks off a filmshoot. Not allways very consistent or good... AND make the preview the best tool possible for the DP.

Untill quite recently, even whiteballancing a shot with the kelvin slider meant a huge degradation of the image. And you wouldn't notice untill you fired it up in post because of the abovementioned factors.

But I do se your point. Far from all posthouses have much knowledge about the RED signalchain...
 
Gunleik Groven said:
I see your points.

My counterpoint is simply:
The onset preview is and will be wrong as long as it is half debayer 8 bit 4:2:2 to a 720-ish monitor.

It's not really compareable to what you see once you fire it up in post.

OK. THAT is an issue, but it won't be resolved on the R1 anytime soon...

better then to make people how the dailies looks off a filmshoot. Not allways very consistent or good... AND make the preview the best tool possible for the DP.

Untill quite recently, even whiteballancing a shot with the kelvin slider meant a huge degradation of the image. And you wouldn't notice untill you fired it up in post bhecause of the abovementioned factors.

But I do se your point. Far from all posthouses have much knowledge about the RED signalchain...
Click to expand...

I can produce a better looking image dialing in a look in the camera off 8bit 422 720P monitoring than 99% of editors or assistant editors messing with RCX in a 10 bit 444 1080 environment. To me, in the absence of a rocket setup on set, which does happen more often than not, making decisions off the 422 720p would still be far superior to taking the chance that an unqualified person will be making color decisions, even for dailies.

This system could potentially give the cinematographer more accuracy and control over the look, but instead because of these inconsistencies does the very opposite and causes frustration, extra work, and extra cost.
 
I take your point.
I just doubt that it is possible on the R1...
The processing is different, thus the result will differ.

This may very well change on the Epic, though!

I would very much like to see the option of a totally different approach, too.

As it is today, neither you or I are very happy... -:)

I think both our approaches have merit, though.
 
1. RedRocket is NOT essential on set even for creating detailed looks.

2. If you are shooting Red you should be professional enough to understand and have a workflow in place to deal with post-pro colour correction. If the post house/editor/whoever isn't taking the RAW footage and grading it properly, then I would say your chain is not really worthy of the "professional" name. If you want an 8bit 4:2:2 image pre-graded, shoot Alexa, or any number of other, inferior cameras, and don't complain about lack of dynamic range etc. Shooting Red is a responsibility which too many people are trying to shortcut, then complaining about how it's less than brilliant in its results. Guess what? It's designed as a cinema camera first and foremost, with a cinema-like workflow. Yes, it can and has been used in all sorts of other workflows but like any professional, it is up to YOU to control your image pipeline and impress upon the post house etc that the workflow needs to match the camera, otherwise you're wasting everyone's time and money....

Just my 2c.

Cheers,

Paul
 
Eric Haase said:
This system could potentially give the cinematographer more accuracy and control over the look, but instead because of these inconsistencies does the very opposite and causes frustration, extra work, and extra cost.
Click to expand...

Amen. If it can be done, there's no valid argument as to why it shouldn't.

I really don't care about the base of users or the history of RED, which I'm pretty familiar with. That's not the issue and it's no excuse. This is about monitoring our images properly.

If a DP is using a Sony L320 on set to light, then a DIT should be able offload and send the same image back to that monitor and have it be consistent. There's no reason for the image to change unless it's a technical hurdle that the RED simply cannot do.

All it really takes is looking at a finished project you worked on and go "what the hell happened in post?"
 
Paul Leeming said:
1. RedRocket is NOT essential on set even for creating detailed looks.
Click to expand...

For monitoring out, it is.

Paul Leeming said:
2. If you are shooting Red you should be professional enough to understand and have a workflow in place to deal with post-pro colour correction. If the post house/editor/whoever isn't taking the RAW footage and grading it properly, then I would say your chain is not really worthy of the "professional" name. If you want an 8bit 4:2:2 image pre-graded, shoot Alexa, or any number of other, inferior cameras, and don't complain about lack of dynamic range etc. Shooting Red is a responsibility which too many people are trying to shortcut, then complaining about how it's less than brilliant in its results. Guess what? It's designed as a cinema camera first and foremost, with a cinema-like workflow. Yes, it can and has been used in all sorts of other workflows but like any professional, it is up to YOU to control your image pipeline and impress upon the post house etc that the workflow needs to match the camera, otherwise you're wasting everyone's time and money....
Click to expand...

Wow. I couldn't disagree with you more. And you just eliminated most of Hollywood from "professional" status... which may actually be correct. ;)

With as many post houses as there are, there seem to be that many workflows. There's existing infrastructure, trained personnel, incompatible hardware/software etc.

Your solution for most post workflows that I've seen you suggest to people on this forum is "Go buy Adobe" or something to that extent. That really just isn't going to fly with a lot of Avid & FCP editors and houses. If that is the actual demands from a camera company, then it really is best to use the other cameras... some of which do have more DR.

These one-man shops which tend to be the majority can adapt very quickly with ever-changing software & color sciences, in terms of workflow, but you guys really have a hard time working with others existing flows. Which is absolutely essential to large scale production.
 
Gunleik Groven said:
Thing is that we disagree about what is "monitoring your image properly".

But that is only fair.
Click to expand...

Right. But even that's not the root disagreement, which is fine.

I'm very happy that the discussion/debate is taking place. We find ourselves on different sides, is all.
 
And I learn a lot from that... Really!

Actually that's the main point for me to involve in a discussion like this. This is an issue that have been bugging me for a while after seeing bad results of in-cam pregrade.

But as I said, I'll be using the sextape in my writeup, which is a loud argument for the opposite view... -:)
 
Paul Leeming said:
1. RedRocket is NOT essential on set even for creating detailed looks.

2. If you are shooting Red you should be professional enough to understand and have a workflow in place to deal with post-pro colour correction. If the post house/editor/whoever isn't taking the RAW footage and grading it properly, then I would say your chain is not really worthy of the "professional" name. If you want an 8bit 4:2:2 image pre-graded, shoot Alexa, or any number of other, inferior cameras, and don't complain about lack of dynamic range etc. Shooting Red is a responsibility which too many people are trying to shortcut, then complaining about how it's less than brilliant in its results. Guess what? It's designed as a cinema camera first and foremost, with a cinema-like workflow. Yes, it can and has been used in all sorts of other workflows but like any professional, it is up to YOU to control your image pipeline and impress upon the post house etc that the workflow needs to match the camera, otherwise you're wasting everyone's time and money....

Just my 2c.

Cheers,

Paul
Click to expand...

Paul, you'd be surprised at how many well known post houses here in Hollywood have workflows in place which do not fully consider color management from on set creation to final delivery. As a DP, I have my IDEAL workflow and I fully understand how to get the most out of RED and get what I want. Shoot RAW, have a colorist one light, offline cut, conform, grade from RAW. Very easy and simple. But in reality, I don't dictate how things are posted- almost every job is a different post house not chosen by me and they all have different workflows, different budgets, etc. So it really isn't up to ME. It's up to producers, or a post producer, or some weird existing relationship the network has with a post house. It's a constant battle and struggle to convince people who have their predetermined workflow that has worked for them before to adapt to anything new. But I do know this: when I dial a look in on camera, it would be nice if it showed up just that way when someone pulls it up on any system/software that can read RED RAW. It would make the struggle a lot easier.
 
Eric Haase said:
But I do know this: when I dial a look in on camera, it would be nice if it showed up just that way when someone pulls it up on any system/software that can read RED RAW. It would make the struggle a lot easier.
Click to expand...

You do it the other way. You match it on the software side and put the look into the camera. That's why that's there because the interpretation is done at the destination because destination is often different than the source monitoring.
 
Mike Gross said:
You do it the other way. You match it on the software side and put the look into the camera. That's why that's there because the interpretation is done at the destination because destination is often different than the source monitoring.
Click to expand...

Sorry, but that wouldn't help.

The image would still look differently when they come back, just in an other way...
 
Gunleik Groven said:
Sorry, but that wouldn't help.

The image would still look differently when they come back, just in an other way...
Click to expand...

Nope it looks the same perfectly..
Because the metadata is being set in redcinex for the output monitor, when you put that look back into the camera the metadata then comes from the camera to redcinex the same as it would be set in redcinex.
 
Good discussion everyone and my apologies if I came across a bit undiplomatically.

I agree with Mike Gross - if you set up a look in RedCine-X and load it into the camera (provided your monitor is properly calibrated to Rec709 or your chosen workflow colour space) then the resulting shots will indeed come back to your workstation with the same RedCine-X preset already applied. So if you are an FCP person that look will be there for you to transcode to ProRes and hand off to clients (if that's what they request).

I also agree that the monitors on Red do not match what you see on your PC, but this is primarily because Red never intended for them to be perfectly calibrated to Rec709 or whatever. You've got the original 5.6", the 7" Pro, the old EVF and new Bomb EVF to contend with, along with the fact that each differs in its colour response, brightness, contrast etc. Not to mention lighting conditions on set such as sunlight falling on the monitor, shooting in the dark etc etc, that all change how the human eye perceives the monitor image.

But Red negates these apparent disadvantages by giving you all the tools needed to properly expose the image to maximise dynamic range and not blow highlights by accident. All metering except for the histogram is now based on RAW. So you can create a look in your computer, upload it to the camera to live within each R3D as metadata, and shoot with confidence knowing that you aren't clipping highlights by judging the traffic lights and the channel clipping bar to the right of the histogram. You can also jump between RAW and RedSpace monitoring modes for additional help if you wish, as well as use False Colour etc.

What you must remember to do if you choose to use a LUT from RedCine-X in camera, is to judge the histogram with the view that if it's clipped either side, then your existing LUT will also clip the image once you reach post, unless you subsequently alter it by changing the metadata again in colour correction (assuming your RAW values weren't clipped to begin with).

As Red have said many times, they built Red One with a clear view to cinematographic applications. Of course it can and is being used in every other field (we've shot every genre as well over the last three years), but first and foremost they built it to mimic 35mm film cameras. When you shoot film you know you have to take the result through a DI process. Red is no different. We have the luxury of using digital, so we can apply metadata to pre-process the image, but really that's just another way of saying you are doing a DI, but putting that step ahead of the actual acquisition. All the curves in the camera (RedColor etc) and those set pre- or post-shooting are a form of DI/colour correction. It's just that our perceptions are changed by it being digital and an expectation (right or wrong) that WYSIWYG should give us a broadcast worthy image with no further work needed.

Thanks for all the feedback, it can only help us all create better images.

Cheers from Tokyo,

Paul
 
@ paul
I agree that the best way to use the RED, or rather: the way you can get the most out of your RED - is shooting for DI.
But I also guess that statistically most users are not.
That is reflected in the very wide userbase, which I think is cool.

My dislike for in-cam grading comes from seeing projects where the DP have dialled in a LUT, loaded it, and then have started to adjust lighting and exposure to the LUT rather than the camera native, and thus messing up shots for cinematic/high-end TV use (despite having a good time on set) and actually making it less likely to achieve that look through tweeking in-cam metadata rather than lightning and exposure on set.

AND
My beloved waveform is also not RAW, and is a bit crippled with the current color-scheme.

'nuff said.

This thread is actually two discussions, and I admit to being the threadjacker, because I finally found a thread where people were willing to discuss the RED 1 on-set preview at all.
Sorry. I should be banned, or at least warned!

Because I see why some users want consistency from set to post.
AND an easy workflow to achieve that, without having to think to much about a number of deliveryformats, colorspaces, gammas and gammut of different formats, media and screening situations. It is all quite a mess, and in the end, it is all images, isn't it?

Why is it that one have to learn to know about all these variables, when it simply should look as I made it on set?

This thread is actually a great argument for an OVF on the Epic! -:)

BUT.
I also see another approach to on-set monitoring being viable: where you get a representation of available options for post, clearly see where the cam goes black and clearly see where it goes white.

So that you get full latitude info on set, and can project a bit more as to how the post will have to be done to achieve your intended image.

This is where I start to be a hijacker and maybe even troll... -:)

Because this is a totally different request and image-approach than what the OP was asking for. It is my interest, and I haven't managed to make anyone bite on this discussion, untill now.

And I have forced myself on a different discussion.

But if it is interesting for more than me, I would like this discussion to go on for just a bit longer....

By, f.ex. starting to post examples.
• On-set.
• Edit
• For grade
• Graded

and/or:
• On set (with in-cam grade)
• delivery.

To me it would be intersting to see examples on how the image develops through the chain, and discuss what we would like to see on set on basis of the resulting and initial image...

If this is too far off, I'll accept it, and maybe try to open a new thread on this subject.

Sorry for threadjacking!

Cheers!

G
 
You must log in or register to reply here.
Back
Top