Black Level off??

[CJ Roy]

This would differ a lot as to what kind of production it is.

Sure, perhaps. I really don't think asking for accurate monitoring is being unfair, no matter the budget. We used to expect such things.

Also, I haven't used redlogfilm yet. It's only a week old or so, right? But that's neither here nor there, as the camera's release build has only RedColor as an option in terms of colorspaces to view on set.

The userbase for RED ONE is very diverse, and to make solutions for people who want to "see what they saw on set" on one hand, and people who "want to protect every bit of info for post/DI" on the other hand is quite a wide bill to cover.

I think this is where we may be misunderstanding each other. I'm not suggesting that anything is less protected. What I am asking for is either accurate monitoring that reflects what the defaults are in Red's raw viewer of choice (redcinex) or that that same raw viewer defaults to what you saw on set.

As Deanan stated, they're intentionally lifting the image for you. I don't think that's necessary since you could so easily do it yourself in camera/cart, if needed.

RED does a good job at it for post. When it comes to onset preview I would have preferred an even more milky look, while others want it to be more punchy... There you go -

And others want the "on set lut". Again, that should be our defaults. How we protect ourselves or adjust the look should be up to the artist hired in the first place.

I am actually not at all a big fan of the "judging the post process from a half-debayer 8 bit 4:2:2 image on a less than 720p monitor with a pretty rough S curve applied" approach, but I can see how that can be a good thing for some work....

I'm not sure I understand this but I think the majority of production work is done in less than ideal standards because of confusion, time and budget... maybe add laziness to that as well.

BTW- Where are your Alexa/MX comparisons? I've loved what I saw thus far with the Alexa.

 

[CJ Roy]

I don't agree. You should dial in the look in redcinex with the video output and then put that back to the camera. The camera is the video tap not a video camera. The output of importance is from redcinex/rocket side so you want that to be your reference, not the camera tap. If you're dialing in the look for full or restricted range, it should be done at the delivery point.

I agree that redcinex/rocket monitoring output should be accurate. All I'm asking for the on board monitoring to be accurate as well, as it's intentionally not, as far as I understand it. There's no reason the two defaults couldn't match.

I find the film DPs have no problems with this.
The guys who are video dps who are used to having to have to dial in a look exactly because you had to do that on the camera before it went to tape. Film workflow is you dial it at the printing side, not on the video tap.

I've worked for every type of DP under the sun. What I've learned is this, if it's greater than a b&w SD video tap, then it should represent the image as accurately as possible. In its current configuration, it's intentionally not.

Again, asking for accurate monitoring is not an unreasonable request, unless they simply can't do it. I have no idea.

 

[Gunleik Groven]

As Deanan stated, they're intentionally lifting the image for you. I don't think that's necessary since you could so easily do it yourself in camera/cart, if needed.


BTW- Where are your Alexa/MX comparisons? I've loved what I saw thus far with the Alexa.

@1
Nope. If it is recorded "down there" you cannot get it back, it would be lost.

You can push things to black, but not get them out of there.

@2
Shot last batch of tests the day before yesterday.

Schedule is to have it ready one week from monday.

Funny you should mention it in this context - though, as Alexa is a very strong exponent of exactly the points I am making in this thread... -

You basically shoot for DI (if you don't record, REC 709 to prores, that is, where the cams look more similar and both act like cool videocams.)

As I am still discussing and experementing with what I see AND will be running some of the images through DI before rekleasing anything, I won't comment directly on the comparison yet. I think it will be counterproductive for the test and both cams.

 

[CJ Roy]

Here's the issue - You want it exactly the same, then you do testing and provide a processing methodology that is repeatable - just like a telecine/film processing house would have done with film. Alternatively, if they want an absolutely WYSIWYG post set up then I suggest you just record a feed off the SDI out of the camera... Because downscaling to a broadcast format from the 4K images, and also putting those images into an SMPTE broadcast colour scheme is likely to change how the colours appear to some extent anyway (just as going from Rec.709 HD to Rec/ITU601 for SD will cause some amount of shift because of a different gamut.) And then there's the H.264 o Mpeg-2 compression for broadcast...

With that rationale, that images just sort of change on their own due to different formats, you're suggesting we shouldn't ask for better?

From monitor to rocket, the image and format technically isn't changing. r3d is r3d, right? The colorspace is supposed to be the same, yet it reacts like 2.2 gamma vs 1.8.

Maybe it's just because I work in post, but I'm confused as to what sort of DOP would EVER think that what they took on the day wouldn't immediately be effected by the post chain in some way, and thus the post guys may have to do some work to retain the look that they were going for...

I think most of the DP's I've met would embrace consistency if they could get it. It's their image, their artistry, their talents, etc.

 

[CJ Roy]

@1
Nope. If it is recorded "down there" you cannot get it back, it would be lost.

You can push things to black, but not get them out of there.

I understand that. Why doesn't on set monitoring represent that?

@2
Shot last batch of tests the day before yesterday.

Schedule is to have it ready one week from monday.

Funny you should mention it in this context - though, as Alexa is a very strong exponent of exactly the points I am making in this thread... -

You basically shoot for DI (if you don't record, REC 709 to prores, that is, where the cams look more similar and both act like cool videocams.)

As I am still discussing and experementing with what I see AND will be running some of the images through DI before rekleasing anything, I won't comment directly on the comparison yet. I think it will be counterproductive for the test and both cams.

Understood. Looking forward to it.

 

[Gunleik Groven]

I think most of the DP's I've met would embrace consistency if they could get it. It's their image, their artistry, their talents, etc.

YUP, but you forget "knowledge about how the signal chain works"

When shooting film, the DP's are aware of the process.
When shooting video, DPs are used to "see the result" sort of.

Actually I think it might be better in some contexts if the tap on the RED was forced to a low-contrast B&W image, to let the DP do what he/she does best: gauge light, compose with it and see the result out of post.

OK. That's extreme. And it is just making a point. But the current situation is creating unwanted confusion for a lot of scenarios... -

 

[CJ Roy]

(deleted)

I think we're just gonna keep going in circles here. I'm sure I'm in the minority on this board, but not on a lot of sets.

 

[Mike Gross]

I agree that redcinex/rocket monitoring output should be accurate. All I'm asking for the on board monitoring to be accurate as well, as it's intentionally not, as far as I understand it. There's no reason the two defaults couldn't match.

What do you set as the defaults? Camera puts out a couple of signals. Ok, choose hdsdi which is restricted range if you put it on a scope. Go to redcinex, now what should the default match? 4:2:0 h.264?, 422 prores?, dnxhd 8 bit?, dnxhd 10bit?, dpx?, rocket sdi 444? rocket sdi 422? rocket dvi? rocket hdmi?
Pick any one and the others will mismatch. The proper way to do it is to pick one, set your look on that output and feed that back to the camera. I know tons of DITs that do this also. It really works.

 

[Gunleik Groven]

I understand that. Why doesn't on set monitoring represent that?

Couldn't agree more. I really wish that was an option, too.

 

[Gunleik Groven]

You're against in camera luts?

Partly. Because a punchy image is hardly a good start for post. And before B30, in-cam whiteballancing on the M could be quite destructive for the post-signal-chain.

It's less destructive on the MX than the M, though.
But for drama/film, it has no particular values other than as a pointer for post as to where the DP wants to go (which is nice)

As LUTS affects most metering and how you expose, though - they can make a lot of harm still, even when it's only metadata. (which is bad, and more destructive for the post chain than NOT having the good part).

As I say, it all depends on the kind of production. RED gives you options, which is good, as long as you know when to use which....

Have to say, though - that I'll be using the Sextape video as an example of a successfull and smart use of the"other" approach... -

http://reduser.net/forum/showthread.php?t=49470

I like both, as long as you know what you do, how it affects your result and the signal.

For filmcomparisons, in camera LUTS are often pretty destructive...

BTW, this may/should change on the Epic, though, when the DP can see "what the cam really records" while the director can have another LUT to see a proxy of the finnished image.


Then in-cam LUTs will be a fantastic tool!

 

[Gunleik Groven]

If you look at these very compressed jpeg's, you can get a vague idea about how much of the precission and latitude is pushed out by the default preview LUT on the R1 MX (redgamma vs redlogFILM and some quick curve adjustment)

And DPs of course adjust to what they see. It's logical. They're artists, not geeks... And I def know which I would have preffered to see on set....

There's just soooo much you cannot see or judge on set with the current preview, and you don't really get happy by getting that back in post... -

"Ouch, what is that weird looking nude doll in the darks?????"

I think it's more important to enjoy the full latitude on set, than to get crushed blacks.

But that's just me, probably...

Cheers!

 

[CJ Roy]

Gunleik-

I'm sure I'm not explaining myself properly, but it does seem like we're agreeing on a couple points. My main point being this; if the lut in camera looks a certain way, it should show up that way in redcine-x. Or vice versa.

So if your DP chooses to shoot in camera on RedColor, RedSpace, Rec709, LogC, blah blah blah... that's how the footage should look in Redcine-X when first pulled up. It really is just a lut, and should have nothing to do with losing information. I'm not saying it should be locked away and never corrected again, I'm just saying it should use what the DP saw as a meta default.

Of course those luts affect the way a DP shoots, but that's their decision. Uninformed or not, they are in charge.

Fair?

 

[Gunleik Groven]

Not really, as that is not really possible without setting the preview to 1/4 debayer 8 bit 4:2:2.

That's like a 16th ish of the pixel, luma and chroma data in the image, and as long as there is NO WAY to show that on set, I think the LUT approach in 9 out of ten times makes more harm than good, and I know it can really mess up post.

But that's my 2c...

 

[CJ Roy]

So unless you have a full debayer playing through a 4k projector on set, the DP should expect pretty large discrepancies that may completely misrepresent his image, no matter how he views it.

 

[Craig Parkes]

Large discrepancies is something you can argue (i.e how large).

But no one has ever had perfect viewing conditions before on set, and the larger the difference between the recorded medium and the viewing technology the greater the difference is going to be.

Video tap looked nothing like film. Film DOP's just dealt with it.

Heck, prior to video taps there were no solution.

What you are asking for is an answer that has never been around... Ok, MAYBE for a very short time when SD acquisition as the only option and people were still lugging around heavy CRT's for on set monitoring.... Then they might have had a pretty damn good WYSIWYG on set monitor.

Other than that its always been an issue.

The issue here is not that the RED camera doesn't do this in my view, the issue here is the lack of ability to explain to the DOP the very reasonable reasons why it doesn't look identical AND what you can do in the monitoring chain to make it as close as possible.

Yes a compromise has to be made somewhere, but these guys are adults, as long as you understand WHY the compromise has to be made, and are confident about it, and don't just blame the gear, they will trust you and they will trust the gear (once they see the finished product.)

Which is why you do camera tests all the way through to distribution format when you want to convince someone. Because that's the only way.

 

[Tom Gough]

The way that I explain it is that the RED ONE is NOT a video camera, it shoots RAW. It sits awkwardly and beautifully between film and video. It does not give you a WYSIWYG monitoring image it gives you a great looking video tap! The RAW footage still needs to be developed, enter Redcine-X(and soon Storm) your virtual telecine wonder tool! Redcine-X is defaulted to give you the best compromise between decent image to edit and to grade in post whether offline or online(when using ProRess/DNxHD/etc...). It is the DITs job to inform every DP new to shooting on RED of these nuances.

I'm not an advocate of applying a final grade on set, i look at Redcine-X as the distributor to the many different post pipelines that exist.

I do think that CJ's argument is valid and can see his POV. It would be great if there was an option in RCX that gives a What You See Is What They Saw (WYSIWTS ) view/look. That way as a DT/DIT/DM you can render out, without any adjustment, exactly what the DP viewed on camera - as long as they are informed of the compromise they are making further down the line - limiting DR, baking in a harsh gamma curve which is extremely difficult to change without image deterioration etc...!

It must be a nightmare for RED to keep everyone happy - the more complaints the more options they introduce, the more options the more complexity, the more complexity the more problems, the more problems the more complaints - it's never ending! The problem with being all to all people is that it's not possible so they have to make it easy for people to adjust to them hence the wonder tool that is REDcine-X. But still some people struggle to adapt to a different way of working(not CJ, I'm talking about the RED blinker brigade that can't accept that times are a changing). Graeme should ditch the Magic Pixies and just drop a Genie into the Epic

Not trying to be a smart ass or teacher of the sucking of eggs, just expressing how I understand it for the benefit of those new to RED (not describing the Pros that have already posted.

All the best,

 

[CJ Roy]

The issue here is not that the RED camera doesn't do this in my view, the issue here is the lack of ability to explain to the DOP the very reasonable reasons why it doesn't look identical AND what you can do in the monitoring chain to make it as close as possible.

There are many things that can be done monitoring-wise, and also lut-wise or colorspace-wise to make this a non-issue. Don't mistake someone voicing an opinion that represents dozens of DP's inquiries, requests and criticisms as a communication problem.

Which is why you do camera tests all the way through to distribution format when you want to convince someone. Because that's the only way.

I agree in part that I wish this is what should be done and redone with every new color science and sensor update from Red, but it's probably something that is mostly done in motion pictures and television where a more definite pipeline is setup. Not so much the case with commercials or music videos.

Just because we've never had accurate monitoring before, doesn't mean we shouldn't. RED is about breaking new ground, right?

I talked to Graeme about this 6 months ago and as he said, it was a known issue in regards to monitoring rocket vs. camera. Apparently the issue will be gone in the Epic.

 

[CJ Roy]

The way that I explain it is that the RED ONE is NOT a video camera, it shoots RAW. It sits awkwardly and beautifully between film and video. It does not give you a WYSIWYG monitoring image it gives you a great looking video tap! The RAW footage still needs to be developed, enter Redcine-X(and soon Storm) your virtual telecine wonder tool! Redcine-X is defaulted to give you the best compromise between decent image to edit and to grade in post whether offline or online(when using ProRess/DNxHD/etc...). It is the DITs job to inform every DP new to shooting on RED of these nuances.

Exactly. The raw is a digital negative, so to speak, or one of the closer things to it. It's not a miracle worker and doesn't mean you can fix anything, but it's very very good. This is one of my favorite discussions when I work with someone new to RED. This is what gets me jazzed up about it in the first place.

I'm not an advocate of applying a final grade on set, i look at Redcine-X as the distributor to the many different post pipelines that exist.

I wouldn't think anyone is a fan or would recommend it. However, the position I find myself in with large scale productions, more than I'd care to admit, is that they're not ever going back to the raw. It's been a pain in the ass for them for 2 years, and it still is. Raw still isn't an option for everyone yet. Obviously it's getting easier, but less than stable as of this moment depending on the workflow.

It would be great if there was an option in RCX that gives a What You See Is What They Saw (WYSIWTS ) view/look. That way as a DT/DIT/DM you can render out, without any adjustment, exactly what the DP viewed on camera - as long as they are informed of the compromise they are making further down the line - limiting DR, baking in a harsh gamma curve which is extremely difficult to change without image deterioration etc...!

Right, if it's just a view/look, then you're not baking it in. If you're rendering out for final, then yes... it is baked in. Or even just for dailies, it's baked into those. That's easily explainable and I haven't met someone who hasn't been able to understand it.

But ultimately, regardless if it's dailies or final, or reference grade... the DP really should have more of a say over his/her image than the inaccurate monitoring offers. Everything should match what they set, as much as possible. Otherwise why even calibrate monitors on set, use waveforms, histograms or anything? Why offer a color video tap? If you're going to do it, might as well do it right.

Anything else is almost undermining the position and the art of it.

 

[Eric Haase]

Frequently, productions will take the footage and have it transcoded without a colorist grading the dailies. There is a misunderstanding amongst producers that you can just take what the DP shot and transocode it without making adjustments and use this as a final. So now an editor or assistant editor or some tech guy at a post facility is going to take the footage and run it through their pipeline. I dial in a look on set, knowing if I've gone too far they could go back to the RaW. But this look isnt what we get after they transocde it at the post facility because what i was looking at and dialing in on camera doesnt match what they output. A lot of times i don't have a Dit with a rocket and proper monitor on set, to set the same look (again) in the file that i set in the camera.

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.

All this means for me as a DP is more frustration in trying to preserve the look I am creating. It means i have to have a Dit with a rocket on set which is an added expense. It means someone else (likely not a colorist) has to make a creative choice as to the color and contrast of the footage just to get a nice looking image. Sometimes i feel like this camera and software is designed for people who shoot, cut and grade their own material and not with cinematographers in mind. I don't see here any GOOD reason why you wouldn't want the monitor output on set to match the look of the file when you load it in to RCX.

 

[Gunleik Groven]

Agreed CJ, but let's go back to the start: Why is this an issue?

Generally (and that's VERY generally) you have two kinds of RED users:

1. The one that wants to see everything until delivery on set
2. The one who knows that you shoot first and post after (for lack of better wording)

BOTH these uses are very legitimate and I am NOT trying to pit them. The fun part is that RED1 can work for both.

Initially the monitoring path was tilted towards user 2 - as far as memory serves me and most people I have discussed this with lately. This created a lot of misunderstanding amongst the first type of users. RED was initially markeded through DVXuser and bought in high numbers by "dvxusers", for whom feature 1 is the most important, and eventually the monitoringpath have been adjusted to suit user 1 more than user 2.

Which is nice and good and fine. Actually, to some extent: I like it.

BUT

On features and TV-series I have seen a lot of downsides to this.
The preview is much "better" than a videotap, how can you expect DPs to NOT relate to what he/she sees, even though it would be a much better way to keep the DR intact if one disregarded the preview and in-camera meters.

But they are there. And so they often throw away 2-ish stops on set.

Then comes post. Many productions only go through a batch transcode to dpx with some kind of preset without a firstlight step.

Boom goes another stop of DR.

And I can see why, but let's leave that out for now...

But the result is that the effective DR and colormetry of the R1 is highly underrated by a lot of DPs.

So, I wish one could choose on-set wether you are shooting a "filmlike" or "video-like" production, and get a corresponding preview option.

Either:
Max DR and low contrast, preferably on both RAW view and ISO/Kelvin adjusted view

Or:
Shoot for finnish, grade-in-cam-deliver-tomorrow view.

I think that would be a good way to get rid of a lot of latitude confusion on the M and MX REDs, and if you don't like the signal as it looks in one of these presets modes... well... choose the other.

None of these approaches are omni-good or omni-bad. But RED covers both kind of users and - heck: I am involved in both of these projects a lot... -

But with such an option, the DP would actually know what he or she shoots for.

As it is, it is at its best a not very nice compromise both ways.

Cheers!

G