RED Build 30 is a step backwards

[Nick Shaw]

And Nick, I'm not sure which image is which, but your first image still shows information above 100....

The first image is the camera's live output, and the super-white pixels you can see on the scope are the graphic overlays. No image pixels go above the white level.

No offence mate, but you're sitting with a post perspective. We NEED that kind of info whilst shooting even if it's not useable later on. It's useable because it gives a sense of *degrees* of highlight clipping. I'm not expecting you to be able to recover any detail out of there.

I admit I am a post guy, so it is not for me to tell a DP how to do their job. But I am just trying to help when I see people making assumptions which I believe are incorrect.

I don't understand what you mean by "*degrees* of clipping". A channel is either clipped or it isn't. The exposure tools are derived from what is 'seen' by the sensor, so cannot give information about a signal which cannot be recovered in post. All that was happening previously was that the frame buffer which drove the HD-SDI output and also the exposure tools was mapping the sensor range to 0-108%, when the post tools mapped it to 0-100%. This was just inconsistent. There is however a valid case for not allowing values over 100% to be used for zebra thresholds.

I am speaking from theory here, as my Red is currently in transit to RED for its MX upgrade, so I can't test right now.

 

[John Brawley]

The first image is the camera's live output, and the super-white pixels you can see on the scope are the graphic overlays. No image pixels go above the white level.

Yes but the information is represented (as red colour) in both over and under. This is still useful information to me on the set, even though i know I can only access 1-100 in post.

I admit I am a post guy, so it is not for me to tell a DP how to do their job. But I am just trying to help when I see people making assumptions which I believe are incorrect.

No disparagement intended. Just an observation that you are looking at it from a post perspective, not an acquisition perspective.

I don't understand what you mean by "*degrees* of clipping". A channel is either clipped or it isn't.

Agree.

The exposure tools are derived from what is 'seen' by the sensor, so cannot give information about a signal which cannot be recovered in post.

Agree.

All that was happening previously was that the frame buffer which drove the HD-SDI output and also the exposure tools was mapping the sensor range to 0-108%, when the post tools mapped it to 0-100%. This was just inconsistent. There is however a valid case for not allowing values over 100% to be used for zebra thresholds.

I'm not only using zebra to see where the sensor clips (at 100) I'm also using it sometimes to judge in a precise way how slightly above clipping something is. Sometimes you want something to clip, sometimes you want to have it just before clip. I can see something that's at 104 (cause that's what I've set my Zebra at) and then close down enough so it's back just before clip. How would i know that it was just above clip if i don't have a zebra that tells me it's at 104 ?

It's like knowing whats just outside of clip and I can choose to *recover* it or not. Of course I can do the same at 100, but it's not possible to do it during a take where the iris move would be seen on camera.

jb

EDIT.

I think of superwhite zebra as an exposure *lookaround* something you can use to spot before it's too late to do something about it....

 

[Stuart English]

Zebra and sensor clip

Zebra and sensor clip

I apologize if I am not being clear in what I am trying to communicate.

Let's recap:

- Zebra by definition is an RGB level measuring tool, and as it is adjustable, you can use it to indicate a variety of conditions.

- There is no such thing as degrees of sensor clip, but there can be degrees of not quite clipping.

- In Build 21 the REDspace LUT often pushed the RGB values beyond 100 IRE, but with REDcolor's FLUT the camera will no longer do that.
(You can still gain up the output via R Gain, G Gain, B Gain and create levels above 100 IRE, so the yellow and red color bars do still exist)

- The only change in Zebra 1 operation is LO now has to reflect what the FLUT is doing to the RAW to RGB conversion.
On M-X at ISO 800 it's LO=99. This is why the default reflects that value.

- If by experiment you find that LO=98 is more appropriate to indicate clip on M at ISO 320, then O.K.

- If you want to have Zebra come on 1/3 stop from your full clip iris value, close the Iris 1/3 stop and lower the LO value to indicate that point instead.

This is what I mean by my comment that Zebra is adjusted only a little from Build 21. Although the specific LO value has changed (to reflect FLUT not LUT) the Zebra tool works just as before, and frankly it seems to me to it operate no differently than any other professional camera on the market in this regard.


What has changed between B21 and B30 is the role of the multi-false color overlay.

 

[sander kamp]

the Zebra tool works just as before

From a technical perspective that might be correct because it is the LUT that has changed.

But from a user perspective it doesn't work like that. When I put the zebra to 104, like I used to do for the past 2 years, it doesn't work anymore. I have to set the zebra to a different value than before to make it work so the zebra tool has changed.

Anyway, I am happy with the discussion so at least I understand what to do.

 

[John Brawley]

I apologize if I am not being clear in what I am trying to communicate.

Let's recap:

- Zebra by definition is an RGB level measuring tool, and as it is adjustable, you can use it to indicate a variety of conditions.

OK lets start with this. And I'll relate it to false colour because maybe that will make sense and you tell me what's now different.

The old false colour went red at 109 IRE, orange between 105-107, yellow from 101-104 and straw from 96-98 IRE.

Now that you map 109 down to 98, is this also why these colours have *disappeared* in false colour in redcolour ?

jb

 

[Nick Shaw]

Now that you map 109 down to 98, is this also why these colours have *disappeared* in false colour in redcolour ?

Yes

John, I sent you a pm where I tried to explain, to avoid repeating myself on the thread.

 

[Eric Haase]

"Why not just make eleven, ten?"

Remap false color's colors so that red is 99+ in the Redcolor RGB space and orange is 97-98 or what not.

 

[CJ Roy]

Let me echo John's frustration with this. The DP's I've been working for in the past few months, and myself, with B30 are equally frustrated with the change. And now they're starting to ignore false color entirely.

I've quoted you, Stuart. I've told them that they're not supposed to use video False Color any longer for exposure and that's just for... well, I don't know what it's for.

If RedColor can't "officially" be overexposed because RED simply eliminated 100-108 IRE, then the values need to be relatively remapped to reflect what is actually lost due to clipping. Right now, if anyone has used False Color before, they can and do think that the yellow/straw band of the False Color is somewhat recoverable. It's not, dependent on the ISO.

I strongly urge you to reconsider and work on different and accurate exposure tools that will help DP's instead of confusing them. Adjusting zebras for every ISO just isn't going to work. Pushing the gain isn't going to fly either.

The request I most often hear of is an accurate false color for the LUT (RedColor) and separate one for RAW, with similar values. You're able to do this with a histogram for RedColor & RAW, why can't you with an overlay?

This camera needs to get easier to use, not more difficult.

Thanks.

 

[John Brawley]

A simple solution

A simple solution

Why not simply re-instate 709 as a monitoring option for those that want it and to use the associated monitoring that goes with it.

jb

 

[Matthew Rogers]

John, Matthew. We have two different arguments going on here so lets separate these out.

Matthew, we'll put your discussion on Video false color on hold for now. O.K?

Hey Stuart, can we talk about FC now?

Matthew

 

[Stuart English]

Zebra and sensor clip

Zebra and sensor clip

Hey Stuart, can we talk about FC now?

Matthew

Sure, but just this final comment on Zebras. The need to adjust Zebra 1 for each ISO rating is a B21 characteristic.

If you you need to adjust Zebra 1 at all under B30, it's a much more subtle change.

 

[Blair S. Paulsen]

IMHO False Color is most useful when it tells me what I am putting on my digital negative which would be RAW. Not an RGB derived value, no log-ish shoulder, just RAW. I would like it to be divided into 10 regions roughly analogous to Ansel Adams Zone system so I can make choices about densities on the digital negative. This concept of FC is intended to provide absolute reference values for technical evaluation by the Cinematographer, DIT, on set post supe, VFX person, etc.

Last comment on Zebras - would just changing the default LO setting to roughly mimic common expectations be OK with everyone. It's not like you couldn't still set your own preference, but at least in rental world there wouldn't be an educational hurdle. Just a thought.

Looking forward to EPIC when we can assign every monitoring path its own specific display state. Anybody like to send RedColor with no look-around, timecode and record indicator to scripty/video village :emote_happyhappy:

Cheers - #19

 

[Matthew Rogers]

IMHO False Color is most useful when it tells me what I am putting on my digital negative which would be RAW. Not an RGB derived value, no log-ish shoulder, just RAW. I would like it to be divided into 10 regions roughly analogous to Ansel Adams Zone system so I can make choices about densities on the digital negative. This concept of FC is intended to provide absolute reference values for technical evaluation by the Cinematographer, DIT, on set post supe, VFX person, etc.

This is my exact point. FC (old style), when it represents RAW, tells me FAR more information than the new exposure tool. In fact, it helps me expose far better than "Exposure" does. Personally, there are times that I would like to see FC in regards to an RGB output (IE: FC monitored in REDColor without FLUT pushing the clipped down to just within the limits.) So it would be nice to have REDColor actually show me when information is clipped, and not just have FLUT crush the highlights down to under 100IRE with no detail (again, please inform me if I am not understanding how FLUT works...)

Matthew

 

[Brian D. Goff]

This is my exact point. FC (old style), when it represents RAW, tells me FAR more information than the new exposure tool. In fact, it helps me expose far better than "Exposure" does. Personally, there are times that I would like to see FC in regards to an RGB output (IE: FC monitored in REDColor without FLUT pushing the clipped down to just within the limits.) So it would be nice to have REDColor actually show me when information is clipped, and not just have FLUT crush the highlights down to under 100IRE with no detail (again, please inform me if I am not understanding how FLUT works...)

Matthew

I just got back from a two day shoot on B30. I really used exposure a lot, and can now say for sure that I MISS THE OLD FC soooooo much! I did get FC adjusted in a way that it would show me burned out highlights relative to RAW "exposure", but it just is not as good as the old FC was, were you would see all the relative colors in RAW view. The new "exposure" B/W just is not telling me enough, because the grey values in the EVF are not good enough. Color overlays in raw from clipping to under exposure like in pre 30 builds was great and makes fast judgments on the over all exposure easy and fast - the new FC and EXPOSURE are time consuming and if you are not careful and view in the wrong mode, FLUT will make you think nothing is overexposed!
Now RED, just take the best of both worlds and make us conservativ FC users happy again:thumbsup:

Starting the poll in a new tread.

 

[Raul Gonzo]

+1 for a hybrid of both or the option of having the old FC.

 

[Eric Haase]

Just finishing up 2 shoots with the mx and build 30. I too miss the old false colour for all the reasons stated in this thread. Fc is not nearly as useful now. We really need a combination of the raw exposure check and new false colour.

 

[Matthew Rogers]

Sure, but just this final comment on Zebras. The need to adjust Zebra 1 for each ISO rating is a B21 characteristic.

If you you need to adjust Zebra 1 at all under B30, it's a much more subtle change.

???? Any more comments about the old FC Stuart? Really would like to see something come back that gives me more exposure info.

Matthew

 

[Michael Epple]

I agree with Brook's statement about zones being more useful than just selecting certain areas of IRE to map (the way it is now). I believe though, that FC should work just like the old way. It maps out values of whatever view space it's in. If you're in RAW, it should be raw, redcolor, then values for that. I understand that RAW view mode is not exactly RAW, but it can be mapped out for sensor clip. This way you always have a consistent base to work from. If you're changing the look in the menu, you always have RAW viewspace to go back to for a consistent reading. If you want to monitor where you're values lie after the in-camera look is applied, then FC works for that as well by using that viewspace.