False Color & Stop's Question-

[Ryan E. Walters]

So I've recently begun to play with False Color in determining exposure. I think that it could be a powerful and great tool once it is calibrated to the RAW view. But for now, it is what it is... With that said I have to following question in regards to the colors in False Color mode-

According to the manual-
Green = 18% grey, or base exposure.
Pink = about 0.3+ stops, or where Caucasian skin should fall.
Orange = 1+ stop over
Red = Clipped

According to the Zone system-
Zone V = 18% grey or base exposure
Zone VI = 1+ stop or where Caucasian skin should fall

As I'm trying to integrate these two systems, I can't seem to make them line up or make sense of each other. Shouldn't skin tone be listed as 1+ stop and not 0.3+ Stop? Is the 1+ stop really equal to Zone VI? Now I'm really confused ... (Maybe my problem is that I'm trying to make them both work together...)

[Irri]

Are you using it in raw view? I think the colour space you are viewing in will affect not only where you are placing 18% and skin tone, but also the the difference between them. I'm not sure, but I think the gamma curve in raw is similar to the one in REC709. I wouldn't place skin and 18% when viewing raw as they will change when transferred. Having said that, I agree that placing Caucasian skin 0.3 over 18% does seem a little low.

[Ryan E. Walters]

Are you using it in raw view? I think the colour space you are viewing in will affect not only where you are placing 18% and skin tone, but also the the difference between them. I'm not sure, but I think the gamma curve in raw is similar to the one in REC709. I wouldn't place skin and 18% when viewing raw as they will change when transferred. Having said that, I agree that placing Caucasian skin 0.3 over 18% does seem a little low.

Unfortunately you cannot use False Color in RAW view as it is not reading the information correctly. You can only use False Color in Rec709 and RedSpace.

But regardless of the color space used, it does seem odd to me that the False Color is broken up into 10 stops below base exposure and only 1 stop above base exposure ...

[Irri]

That seems like an odd distribution. I think with REC709 you get 2 stops over and 7 stops under. A lot of people had an issue when rating the camera with 18%. Redspace goes some way towards rectifying that by lifting the lower levels so that 18% is closer to the middle of the scale. Sorry if that sounds a little vague, I've had a few glasses of wine.

[Ryan E. Walters]

Yep, it is an odd distribution indeed. I get what you are saying - regardless of the glasses of wine. :)

I do not get the distribution either- it makes no sense to me. Especially since the distribution is the same regardless of the color space used. (I understand that the different color spaces give different results, but False color still has the same scale to work with - 10 stops under and 1 stop over exposure...)

False Color seems like it could be a very handy tool- but since it does not work with RAW view, and the distribution seems to be askew - it seems that I'll have to not bother using it, and just go back to using the histogram and the color bar...

I really wish False color would work... :(

[Blair S. Paulsen]

Allow me to join the chorus of those who look forward to a tweaking of the false color tool. BTW - I would love to see it propagated into RA, RC and Scratch as well.

FWIW my dream set up would be 10 colors that directly correlate to the Ansel Adams Zone System. From what I understand, the key is to intelligently manage the gamma issues inherent in re-mapping a linear system that starts with most of the code values all crunched together into a logical visual representation.

I know the Red Team has more than enough on their plate but, IMHO, False Color 2.0 has the potential be the gold standard for exposure tools in the entire imaging industry. Beyond that, in my experience, nailing exposure on the RedOne is the key to realizing the potential of the camera system. The better the exposure tools the more likely that is. Just sayin'

Cheers - #19

[David Mullen ASC]

So I've recently begun to play with False Color in determining exposure. I think that it could be a powerful and great tool once it is calibrated to the RAW view. But for now, it is what it is... With that said I have to following question in regards to the colors in False Color mode-

According to the manual-
Green = 18% grey, or base exposure.
Pink = about 0.3+ stops, or where Caucasian skin should fall.
Orange = 1+ stop over
Red = Clipped



Yes, that seems odd, skin tone being only 1/3-stop over 18% grey. Of course it all depends on whose skin tone you are talking about... And if Orange (which would be Zone VI) is one-stop over, then what is Red? Two-stops over? Three?

Perhaps, like Zebras, the user should be able to have four false color zones and then assign them to whatever exposure levels they feel would be informative. Or maybe it should be set for three-stops under 18% (for near black), 18%, one-stop over (for skin), and clip point, in Rec 709.

[Ryan E. Walters]

FWIW my dream set up would be 10 colors that directly correlate to the Ansel Adams Zone System. From what I understand, the key is to intelligently manage the gamma issues inherent in re-mapping a linear system that starts with most of the code values all crunched together into a logical visual representation.


Yes, that seems odd, skin tone being only 1/3-stop over 18% grey. Of course it all depends on whose skin tone you are talking about... And if Orange (which would be Zone VI) is one-stop over, then what is Red? Two-stops over? Three?

Perhaps, like Zebras, the user should be able to have four false color zones and then assign them to whatever exposure levels they feel would be informative. Or maybe it should be set for three-stops under 18% (for near black), 18%, one-stop over (for skin), and clip point, in Rec 709.

Great ideas- I'm really hoping that Red does one of these or both of them- right now it seems to skewed to be of any real use. Especially since it cannot be used reliably in Raw view ... False color is a great idea, now if it could just be finished out / completed...

[rotten]

can somone tell me if the little column on the red lcd shows the same information of the flase color?
and, sorry about that, what does the other little column shows?

[Ryan E. Walters]

can somone tell me if the little column on the red lcd shows the same information of the flase color?
and, sorry about that, what does the other little column shows?

If you are referring to the column in-between the histogram and the traffic lights: It is a meter for the RAW sensor / view. When it hits red there is a part of the raw image that has been clipped. No Red = no clipping. It does not show the same information as false color, since false color is dependent on the view mode, and false color is not accurate when viewing RAW.

The traffic lights (Next column over) go off when 2% (or is it 1%) of the image is clipped. The corresponding color light shows up for the channel that is clipped. This is also based on the viewing space.

[anzyyy]

:emote_head_explode:hey i still cant figure out.........

[Ryan E. Walters]

:emote_head_explode:hey i still cant figure out.........

Figure out what? You'll need to post a better response then that to get a specific answer ...

[Casey Green]

I had an idea and wanted to run it by others interested...

Would it valuable to have the ability to generate custom False Color meters?
(Perhaps they could be called something else to keep confusion at a minimum)

Here's what I have in mind:

On-set, record a camera test clip (or with DSMC, a Still) and off load it to the Data Capture station. Then, in REDCINE-X (or RA!, etc.) have the ability to assign specific False Color hues to custom values pertaining to that clip or image's unique exposure. Also, you would be able to have control of which colors and values even get a color representation. For example, you could leave the entire image grey scale except for +1 stop, clipping, and perhaps -8 stops.

Then, be able to Save that custom False Color file, and upload it into the Camera.

It seems this could allow for more precise exposure monitoring. Would this be of value?

[JonathanF]

Perhaps, like Zebras, the user should be able to have four false color zones and then assign them to whatever exposure levels they feel would be informative. Or maybe it should be set for three-stops under 18% (for near black), 18%, one-stop over (for skin), and clip point, in Rec 709.

I'd second that.. it seems a reasonable simplification.

You generally have an idea of what range is most important so being able to user define that seems a great idea. You can make the response suitable for those areas of importance. In any case being able to map whatever user color you like to any exposure seems like just the ticket.

But this all begs the question, when is RAW actually going to get fixed so it works? Some of us aren't working in 709 or redspace and would love to meter in raw.

Cheers,

:J

[Ryan E. Walters]

I had an idea and wanted to run it by others interested...

Would it valuable to have the ability to generate custom False Color meters?
(Perhaps they could be called something else to keep confusion at a minimum)

Here's what I have in mind:

On-set, record a camera test clip (or with DSMC, a Still) and off load it to the Data Capture station. Then, in REDCINE-X (or RA!, etc.) have the ability to assign specific False Color hues to custom values pertaining to that clip or image's unique exposure. Also, you would be able to have control of which colors and values even get a color representation. For example, you could leave the entire image grey scale except for +1 stop, clipping, and perhaps -8 stops.

Then, be able to Save that custom False Color file, and upload it into the Camera.

It seems this could allow for more precise exposure monitoring. Would this be of value?

I think that is an interesting idea, and it might work for some people. For me, however, I'd just like a False color that actually worked in RAW view, and was evenly distributed across the zone system. Even better would be one that was user definable like what has been suggested in other posts.

With a system that I would have to "print" out and then reload a specific profile for it to work would not be ideal for a lot of the smaller projects that I work on as it would be an additional step. For the larger projects, with the crew to help, it could come in handy.

[David Mullen ASC]

I'd find, though, ten false colors representing Zones to be rather a mess to look at on a screen image. If you wanted that sort of even distribution of false colors, I'd skip every other zone or something to make it less confusing to look at.

Otherwise, probably what you really want to see is Zone 1, Zone 5 (18% grey), Zone 6 (caucasian skin tone), and Zone 10.

[Ryan E. Walters]

I'd find, though, ten false colors representing Zones to be rather a mess to look at on a screen image. If you wanted that sort of even distribution of false colors, I'd skip every other zone or something to make it less confusing to look at.

Otherwise, probably what you really want to see is Zone 1, Zone 5 (18% grey), Zone 6 (caucasian skin tone), and Zone 10.


That is a good point and that might cause some issues. Right now with False Color there are about 14 or 15 colors, and with the limited experience I've had using the False Color mode it doesn't seem to be a huge issue- but then again the scale / distribution is way off so many times not all of the colors are displayed at the same time ...

[Kim Frank]

Otherwise, probably what you really want to see is Zone 1, Zone 5 (18% grey), Zone 6 (caucasian skin tone), and Zone 10.

I'm always using Raw view, where False Colour isn't reading appropriate at the Moment, right? That's why I'm wondering how I would have to adjust the Zebras setting, to get the same readings David suggested for false Colour.
What IRE would be perfekt for reading out the noise floor in the Blacks (where should I set it with the new colour scince) and the clipping in the highlights? I don't own a red that's why I have to ask, am I able to set midtone values for 18%grey and skin tone (zone6)?

[Casey Green]

I'd find, though, ten false colors representing Zones to be rather a mess to look at on a screen image. If you wanted that sort of even distribution of false colors, I'd skip every other zone or something to make it less confusing to look at.

Otherwise, probably what you really want to see is Zone 1, Zone 5 (18% grey), Zone 6 (caucasian skin tone), and Zone 10.

I agree. While very useful, False Color can be confusing to look at. I proposed something similar to this idea a while back for a "False Color 2" mode... (slightly updated version below)



I love the False Color mode, and think it is one of the best tools available on the RED, HOWEVER, it would be great to add a False Color 2 mode which only had the colors represent these modes (or something similar):

PINK = Clipped Hilights
RED = Almost clipped (set to a specific Zone or Zebra)

Grey Scale = all shades in between
Dark Orange = certain skin tones or Zone

Dark Blue = "Somewhat Under Exposed" (set to a specific Zone or Zebra)
Dark Purple = "Fully Under Exposed" (set to a specific Zone or Zebra)

Some of the other colors in the False Color mode are hard to remember and also make viewing the image confusing.

I would love a False Color 2 mode as an option.


...But customizable uploadable False Color modes for certain projects would be very helpful.

[Tai Wah Lim]

Agree with Casey on his points on using False Color. Anyone tried the Leader False color -http://www.leaderusa.com/whatis/cinezone.htm. Any difference? Lim

[Blair S. Paulsen]

If you've followed the thread this far you probably love the concept but have issues with the implementation. Like a lot of features on modern electronics there is a sweet spot of customizability vs ease of use that the Red Team will have to ultimately determine.

Our job on this forum is to propose the most elegant solutions we can and subject them to the marketplace of ideas. That said...

The impact of the gamma encoding of a view space has the potential to mis-lead so I think limiting false color to RAW only makes the most sense. BUT - (and it might be the reason that hasn't already been done) the RAW linear light image is very dark and the distribution of code values might be problematic. Perhaps a transform (LUT) specifically built for false color is the answer?

In terms of the actual colors how about: zebra hatch 100 IRE + (clipped), red 90-100 IRE (Zone X), orange 80-90 IRE (Zone IX), yellow 70-80 IRE (Zone VIII), flat grey 50-70 IRE (Zone VI & VII), primary green 40-50 IRE (Zone V), blue 20-40 IRE (Zones III & IV), purple for under 20 IRE (Zones I & II). That funky pink for skin tone would be a good candidate for a settable value with a default for quick set ups.

I agree with David that 10 colors might be too much. I would prefer a system that would be consistent from rig to rig for multi-cam shows so I do not support any protocol that would be unique to an individual camera unless it could be easily toggled off.

Cheers-

Blair S. Paulsen
4K Ninja, RedOne #19
San Diego, CA

[Ryan E. Walters]

In terms of the actual colors how about: zebra hatch 100 IRE + (clipped), red 90-100 IRE (Zone X), orange 80-90 IRE (Zone IX), yellow 70-80 IRE (Zone VIII), flat grey 50-70 IRE (Zone VI & VII), primary green 40-50 IRE (Zone V), blue 20-40 IRE (Zones III & IV), purple for under 20 IRE (Zones I & II). That funky pink for skin tone would be a good candidate for a settable value with a default for quick set ups.

Now that would be a GREAT system / setup- I'd be all for that. :)

[Ido Karilla]

Originally Posted by Blair S. Paulsen
In terms of the actual colors how about: zebra hatch 100 IRE + (clipped), red 90-100 IRE (Zone X), orange 80-90 IRE (Zone IX), yellow 70-80 IRE (Zone VIII), flat grey 50-70 IRE (Zone VI & VII), primary green 40-50 IRE (Zone V), blue 20-40 IRE (Zones III & IV), purple for under 20 IRE (Zones I & II). That funky pink for skin tone would be a good candidate for a settable value with a default for quick set ups.

Very well organized but to reduce further the amount of Colors to follow I do not see a reason to keep zones IX and VIII separate in False color mode, or maybe split zones X, IX and VIII to 2 colors instead of 3.

[JonathanF]

I think user definable colors are the best bet. You're never going to know if your idea works till you really start using it. Best laid plans and all...

How about this one, let's get false color working with RAW first and then worry about colors.

:J

[Dave Edsall]

What are the problems associated with using false color while monitoring RAW?

Dave

[Stuart English]

What are the problems associated with using false color while monitoring RAW?

Only that there is a bug when you switch out of REDspace to RAW via a USER KEY, FALSE COLOR doesn't read back correctly.

However FC is correct when you switch out of REC709 to RAW via a USER KEY.

And FC is correct when you select RAW in the menus VIEW = RAW.


This bug is being addressed.

[David Wyatt]

Only that there is a bug when you switch out of REDspace to RAW via a USER KEY, FALSE COLOR doesn't read back correctly.

However FC is correct when you switch out of REC709 to RAW via a USER KEY.

And FC is correct when you select RAW in the menus VIEW = RAW.


This bug is being addressed.

Hi Stuart,

My original understanding was that False Colour in general wasn't accurate in RAW mode, but this appears to have changed (not sure when that announcement was made but I must have missed it):

http://205.234.135.241/forum/showthread.php?p=318287#post318287

Just to be clear about this are you saying false colour basically doesn't read back correctly in RAW...BUT only if you've used a User Key to flick from REDspace to RAW? This seems to suggest that if you flick from REDspace to RAW via the menus rather than the User Keys then False Colour will be accurate/calibrated (which your third point also suggests?). Likewise if you're using Rec709 as your View and you use a User Key to flick to RAW then False Colour IS accurate/calibrated? Is this true for Release Build 17, Beta 18 or Beta 20 or all three?

Can you just double-confirm these things because nailing exposure can be pretty important if you're exposing to the right and trying to get as much out of the RAW image as possible. I have to admit, and I hate to sound like a whiner, but with all these provisos & bugs about False Colour that unfortunately most of us seem to accidentally find buried in other threads, I'm now thinking of avoiding False Colour as a critical exposure guide...too many bugs...too little information trickling out...which is a shame because on paper False Colour is a revolutionary idea. Maybe it's time for a sticky about this sticky situation? Just to clear up the confusion and/or lack of clear information?

[JonathanF]

Maybe it's time for a sticky about this sticky situation? Just to clear up the confusion and/or lack of clear information?

I'll second that sticky.

:J

[Stuart English]

Just to be clear about this are you saying false colour basically doesn't read back correctly in RAW...BUT only if you've used a User Key to flick from REDspace to RAW? This seems to suggest that if you flick from REDspace to RAW via the menus rather than the User Keys then False Colour will be accurate/calibrated (which your third point also suggests?). Likewise if you're using Rec709 as your View and you use a User Key to flick to RAW then False Colour IS accurate/calibrated? Is this true for Release Build 17, Beta 18 or Beta 20 or all three?


That's correct - all three builds.

Give it a try yourself - Enable FALSE COLOR and shoot an 18" Grey card at a slight angle so that you can see green and pink color bands.

Now switch the VIEW menu between REC 709 and RAW, and then REDpace and RAW. The color bands in RAW are at the same position.

Now set a USER key to VIEW RAW. Go back to the VIEW menu. Select REC 709 and then press the USER key. Observe the band position.

Press the USER key again and change to REDspace. Press the USER key one more time and observe the band position.

You will see that the "VIEW RAW" following that key sequence isn't the same as the "VIEW RAW" following the earlier key sequences..

.. hence my "it's a bug, we will fix that" comment.

[Harry Clark]

This is a good thread. I, too, love false color as a tool but wish for a more logical implementation, one that mirrors the zone system and has some customization.
Thanks for the tip on the redspace user button bug. I was unaware of that.
Cheers,
Harry

[David Wyatt]

That's correct - all three builds.

Give it a try yourself - Enable FALSE COLOR and shoot an 18" Grey card at a slight angle so that you can see green and pink color bands.

Now switch the VIEW menu between REC 709 and RAW, and then REDpace and RAW. The color bands in RAW are at the same position.

Now set a USER key to VIEW RAW. Go back to the VIEW menu. Select REC 709 and then press the USER key. Observe the band position.

Press the USER key again and change to REDspace. Press the USER key one more time and observe the band position.

You will see that the "VIEW RAW" following that key sequence isn't the same as the "VIEW RAW" following the earlier key sequences..

.. hence my "it's a bug, we will fix that" comment.

The thing that confuses me is that now you're saying False Colour IS calibrated/accurate in RAW mode (as long as you avoid the User Key bug), whereas by earlier posts I was lead to believe False Colour was NOT accurate/calibrated in RAW view (however you accessed it):


Well to avoid going round in circles on this, lets just say this then - NO, its not calibrated, and YES I agree it should be.

Meanwhile its accurate enough that you probably haven't had any footage come back from post because of clipping.

[Stuart English]

Well, it is calibrated now.... i.e has been since Build 18. Can't recall offhand for 17 or 16...

[David Wyatt]

Great, thanks for clearing that up.

[Ryan E. Walters]

Hmmm, interesting- I'll have to check it out. It is good to know that it is calibrated when in RAW now. That is VERY helpful.

Now if it was just possible to have it reflect the zone system ...

[JonathanF]

Well, it is calibrated now.... i.e has been since Build 18. Can't recall offhand for 17 or 16...

So Stuart, is this fixed in 20 now?

:J

[Stuart English]

The latest version of Build 20 includes new False Color overlay - see

http://reduser.net/forum/showthread.php?t=32891

The modified False Color overlay provides the following information:

Purple
Minimum Exposure

Teal
3 stops under reference

Green
Reference Exposure for 18% Grey Card

Pink
Reference Exposure for Caucasian Skin

Straw
2 stops over reference

Yellow
2/3 stop under maximum exposure

Orange
1/3 stop under maximum exposure

Red Maximum Exposure

[Ryan E. Walters]

Thanks- this is a big help. :)

[Blair S. Paulsen]

I deeply appreciate the efforts of Stuart, Rob and everyone else on the Red Team who found the time to improve false color. In some ways its a small thing but in my day to day work with the camera it is such a tremendous assist. I often have a stressed DP, director or producer looking at me to confirm that the exposure is right and by checking false color I can tell in less than 5 seconds. Gold.

Tip for those of you who have not done so, the false color mapping is a fabulous way to check how evenly a green screen is lit. I often repo the monitor for the gaffer or CLT during set up so they can massage the lighting. Makes keying so much easier & faster.

Cheers- #19

[Stuart English]

Thanks for that Blair.

A note on Zebra operation in this version 20.1.2 of Build 20.

As you probably recall Zebra settings can be adjusted by the user and are remembered over power cycles.

Zebra settings are also stored as part of a USER PREFERENCES file.


In this Build the defaults were adjusted so that -

a) They are both enabled - so that's what the diagonal lines in the image are.

b) Shadow Zebra falls on the Purple false color bar

c) Highlight Zebra falls on the Yellow / Orange / Red false color bars.


Again, these are just default values, these settings can be adjusted to taste.

And you can turn them off if you wish - your choice.

[Paul Leeming]

I posted this in the main Release thread but in case anyone missed it there, here is a useful guide for use on your iPhone to show DPs the new False Color steppings:

http://www.visceralpsyche.com/misc/web_images/RedFalseColor_build20.1.2_iPhone_Reference_Chart.p ng

HTH

Paul

[Ryan E. Walters]

Thanks. :)

And here is one that will fit on the 5.6" LCD:

http://reduser.net/forum/showthread.php?t=33085

[nitin sagar]

I'm new to Red. Yesterday was my first day of Shoot[ though I had done some testing earlier] my worst fears of testing came true yesterday. I like what I see in the EVF but one the rushes are transcoded the mov file looks a stop under...since i've entered a long schedule with no real solution in sight..i over exposing by a stop to feel better! any comments please ...any. I want to crack RED.Period!