[Glenn Chan]

The following article has information from Apple on how FCP handles gamma:
https://docs.info.apple.com/article.html?artnum=93794

According to it, FCP tries to make the computer monitor gamma match that of a broadcast monitor.

I think the colors are different between paused and playing.

[Glenn Chan]

Use the 3-way color corrector. Expand out the limit color section.

The concept is definitely the same. Holding shift while you have the eyedropper will add the clicked color to your selection.

[Glenn Chan]

Only Sony has made a claim that the Luma series emulates SMPTE-C. But neither SMPTE nor NAB has come out to say “This will be the new broadcast standard for LCD & Plasma displays.”

I think you may be misinterpreting how the standards… my understanding is that their standards are defined with an ideal monitor in mind. Right now, even the most expensive Sony CRT available (the 32″ Sony BVM-A, $37k list at B&H) meets that criteria for Rec. 709 HD. It doesn’t use the Rec. 709 primaries (it uses SMPTE C; I haven’t checked if it lies within the tolerance) and it isn’t capable of hitting 35fL without problems.

In this case, ideal would mean the monitor that a manufacturer *should* build. The LCD and plasma displays that are produced should be designed to be as close to the ideal monitor as possible. In practice, this doesn’t happen and there are some areas where displays don’t fall around the standards. Among consumer CRTs, you have some systematic differences like higher color temperature, and image cheats (i.e. flesh tone correction, scanning velocity modulation, excessive sharpening, etc.).

If you wanted to, you could grade with the audience in mind. i.e. knowing that most HD viewers are watching on LCDs, you make the assumption of a raised black level. Knowing that most monitors have that flaw, you try to correct/adjust for it. This approach has a disadvantage in that monitor trends shift over time. Now that LCDs and plasmas are more common, the average color temperature is going to get much closer to D65 (the standard color temperature / white point for everywhere except for Japan).

The strategy (of grading for your audience) runs counter to engineering standards… since you are no longer grading towards the engineering standard (i.e. not following it). The problem with this strategy is that you aren’t grading for your audience.

2- In an ideal world, the monitor manufacturers would follow engineering standards.

Because technologies change/improve (i.e. emergence of LCD, and other display technologies) and because of economics/politics (brighter CRTs supposedly sell better), you have manufacturers moving away from engineering standards. Among consumer CRTs, the color temperature is higher because it makes the CRT brighter.
In some cases, things are wrong because the engineers screwed up (i.e. chroma bug in DVD players) and/or because it’s cheaper not to fix it (i.e. may be why most DV equipment doesn’t put analog black level at 7.5 IRE).

3- For some information on consumer device characteristics, see
this article comparing display technologies
*There may be some technical inaccuracies, but I believe most of the information is good. The inaccuracies that I know of are: 6500K and D65 are different colors / white points; CRT gamma is not quite as simple as 2.2; suggestion of Lu’v’ processing is dubious

[Glenn Chan]

The black level is definitely not great.

The colors look a little weird… Sony has some information on how (not) close the monitor is to Rec. 709 colorimetry… at least they had such a chart at NAB. The monitor is switchable between Rec. 709, SMPTE C, and some other colorimetries/primaries.

The CRTs are a much better buy IMO.

[Glenn Chan]

I guess, with computer monitors so easy to calibrate with affordable probes and software, I think we should be able to view the image in FCP on a calibrated and profiled computer display as a standardized monitor.

If photoshop can do it, so should FCP.

A good computer monitor plus a computer should easily be able to equal a 20,000 dollar calibrated LCD production monitor.
Photoshop does it because (for print work) the only way to get decent color is to profile all your devices and apply color management in the software.

2- The ‘affordable’/lower-end probes and software don’t do that great a job. see
https://www.drycreekphoto.com/Learn/monitor_calibration_tools.htm

3- If the monitor’s color gamut doesn’t encompass the color gamut of the output device, it will never do a perfect job. The software has to apply cheats to squeeze one gamut to fit into the other.

A computer monitor probably won’t perform as well as a reference-grade monitor. Regardless, FCP doesn’t implement any form of color management currently (i.e. it doesn’t do ICC, it doesn’t do 3D LUTs).

What’s the best way to set up my mac/FCP today so that the canvas image is as accurate as I can get it?
As I stated in my previous message, hopefully the iOne changes your video card’s LUT (look-up table) to get the desired gamma, grayscale tracking, white point (D65), white level luminance (35fL), etc.
White point: In print work, D50 is the standard for low-cost monitoring and D65 for higher-quality monitoring. For video work, D65 is the standard. 6500K is not the standard… it is a different color than D65.
Luminance: This doesn’t matter too much because almost nobody follows the standards here. In high-end facilities, you have facilities using a wide range of figures (i.e. 20fL, 30fL, 35fL). Part of this is because the largest Sony BVM-A can’t do 35fL without problems, so it’s best run at 20fL. Most consumer monitors are significantly brighter than the 35fL standard.
If you run a LCD, you should probably run it as bright as possible so you have less black level error.
If you run a CRT, white level should not be set via the video card’s LUT. Do it via the monitor controls.

Gamma: See
https://www.apple.com/support/shake/lbn/40/#Gamma
for information about how FCP handles gamma. Some of the information I stated previously about 0-255 versus 16-235 is incorrect.

The practical thing to do is still to get a broadcast monitor and/or check your work on your target displays.

[Glenn Chan]

To get accurate color, you’re looking at the following:

A- Calibrating the interface. With broadcast monitors, you do this via color bars. For computer monitors hooked up via DVI, it’s digital so the interface doesn’t go out of calibration.

B- Calibrating the monitor for grayscale tracking (i.e. are black and white images perfectly black and white). The lower-end broadcast monitors don’t let you do this. On the mid-high end ones, you can manually tweak the monitor controls to get close. Probes by Sony, Ikegami, and Minolta can automate the process and make it more objective.

For home theatre, there’s some device for doing this. Because the controls are rather crude, they can’t fix serious grayscale problems (i.e. from a CRT aging). For critical color, CRTs should eventually be replaced.

For print work, various monitor profiling devices like the iOne handle this. If you want to adapt this for video use, hopefully the iOne messes around with the video card’s LUT to load the calibration there. This will carry over into FCP.

You may also want to calibrate the monitor for gamma, white point, white level.

C- Some monitors like the aluminium apple cinema display have color problems which can’t be calibrated away. There may be some other monitor defects that you can’t calibrate off (i.e. LCD black level glow)

For video work (especially SD), not showing interlacing is something you can’t fix.

D- Differences in color gamut between the monitoring display and the output device.

In your situation, your computer monitor’s color gamut is likely different from the reference standard (SMPTE C, EBU, or Rec. 709). I don’t know of any cheap tools to fix this. For TV work, this is not a problem when you monitor with a reference-grade monitor (i.e. a CRT with SMPTE C phosphors).

For print work: The output device’s color will vary depending on ink and paper. To handle this, you can use color management systems like ICC. The iOne probably is designed to generate ICC profiles.

ICC was never designed to work with video. Real-time performance is not there. Also, you don’t need it for TV work (you just get a reference-grade monitor). For film output work (i.e. DI), there are 3D LUT solutions available (i.e. Truelight).

—-
The simple solution for SD work is to get a broadcast monitor / reference-grade monitor. The entry level ones are around $500 (not the greatest color, but better than other options at the same price). Use a DV camera to go from firewire to monitor.

For HD work, I think the cheapest CRTs are around $5k.

—-
To answer your question:
Hopefully the iOne can tweak your video card’s LUT to achieve the desired gamma.

I’m actually not sure what the correct gamma is…

—-
I don’t believe FCP has any color management per se. However, there are some behaviors to note:

A- When the timeline is not playing, I believe FCP will decode the Y’CbCr video signal to 0-255 RGB (RGB color, with black at 0 and white at 255). When the timeline is playing, it will decode to ~16-~235 (presumably because this is faster). The color when playing isn’t exactly right.
This is based off my memory…

FCP calls Y’CbCr “YUV”. YUV refers to an analog encoding scheme and is not the same thing as Y’CbCr, but (almost) everyone in the computer world calls Y’CbCr YUV. This is not important in your case.

B- When you import stills, FCP applies gamma correction. This is to convert the still’s sRGB gamma to match your video footage presumably.
The superwhite versus non-superwhite presets also affect how stills appear.

—-
what might be the best compromise approach?
The best compromise approach is to output your project and view it on your target audience’s displays (or just a few of them). I’m not kidding… this is advantageous because the real world differs from engineering standards (i.e. even FCP differs from engineering standards because of oversights/bugs).

[Glenn Chan]

The ideal/standard (i.e. SMPTE) color temperature of the light is D65, not 6500K. The D scheme is for daylight illuminant series, the Kelvin scheme is for black body radiators (or things that behave like them). There’s a subtle difference between D65 and 6500K, which is unfortunately confused most of the time (heck, I have a Sony monitor which reads 6500K). There are some sources on the internet which gives the correct chromaticity co-ordinates for D65 and 6500K… one of them:

http://www.etconsult.com/papers/whitepoint.pdf

The practical thing to do is to make sure the white points of lighting sources in your suite matches. It’ll be ok if you’re slightly off from D65 (your eyes will “white balance” itself regardless). It’s only when you’re way off from D65 that it’s a problem (i.e. using consumer CRTs, which oversaturate reds to compensate for their higher/blue-ish color temperature). Do try to match your computer’s white point to your monitor, or vice versa. If the white points are different, your eyes will white balance to one source. When you look at the other, the white balance is off and this affects color appearance.

2- Fluorescent lights (most of em) tend to be a bit green-ish… since they naturally output more light that way. Higher CRI lights tend to be less green. The CRI figure is for how closely the light resembles that of a black body radiator.

Black body radiator: If you heat up an object, it will emit light (i.e. like a stove element). The particular color of light emitted is linked to how hot the object is (which Kelvins is a measure of).

3- SMPTE RP162 defines recommended viewing practices. One of the things it recommends is a particular luminance for the surround… I believe this is 3.5fL (10% of the monitor’s luminance when it displays a white field). Or it’s 15%, I can’t remember.

The thing about RP162 is that many facilities (high-end, mid-end, low-end) do not follow it.
-Monitor brightness: The largest Sony BVM monitor cannot hit 35fL without problems, so you’ll have Sony equipment that runs at 20fL. There is also confusion about whether 30fL or 35fL is the correct brightness. In practice, you also have facilities using a whole bunch of other numbers for monitor brightness.
-Surround size: All over the place.
-Monitor distance: For monitoring, it’s generally useful to put the monitor a lot closer than what the standards call for.
-Black level: Setting this via PLUGE is inexact, subjective, and depends on how much ambient illumination is hitting the monitor. Ambient illumination hitting the monitor should likely be as low as possible… you can achieve this by positioning some lights behind the monitor (to light the surround), and flagging others.

The TIG mailing list archives (tig.colorist.org) has some discussion on these issues.

e) if yes what would be more ideal just for editing? Halogen or fluorescent and at what lux numbers?
It doesn’t matter much. A fluorescent with fairly high CRI or daylight-balanced incandescent should be fine. Getting close to D65 would be good. Matching your computer monitors’ white points to the broadcast monitor’s would also be important.

If you need to do color matching (for print work), then the metamerism index of the lighting will matter. But that’s not the case here.

d) would it be recommended to install and use different light fixtures when editing?
You can edit with the same light. Some people prefer a brighter environment, since too dark causes eye strain.

Unfortunately, there’s not that much information on this subject. I don’t claim to have all the answers. In practice, I have seen many facilities (with expensive equipment) where the viewing environment is poor. Heck… I’ve seen a video from a manufacturer of video test equipment where you can see the light fixtures reflecting off their monitor (obviously these reflections will cause color inaccuracy where the reflections land). It doesn’t seem like people notice… this may be because our color memory is poor, and because what we see is not very color accurate (because of varying lighting conditions and things like metamerism).

From a technical perspective, knowing about the issues and knowing what to look for (i.e. artifacts, what incorrect levels look like) can be a big help in accurate monitoring. Also testing your monitor is important… I don’t think that much attention is paid to this. A simple and good test is to display a black and white image on your monitor (de-saturate it 100% to doublecheck that it’s B&W). Look at your computer monitor fro 10 seconds, then look at your broadcast monitor. Is the image completely black and white? This tests the monitor’s grayscale tracking and the white point differences between your monitors.
You might want to throw the system out of whack so you know what to look for.

Another issue to be aware of is that your eyes’ “white balance” is constantly adapting itself, and that your brain tends to filter out this adaptation so you don’t notice it. i.e. everyday you experience dramatic differences in illumination, but you don’t notice this. If you are looking at a broadcast monitor, you can be susceptible to white balance drift. This effect is stronger with a darker surround (i.e. what RP166 dictates is darker than people watching TV with the lights on, with white walls). This would be something to keep in mind.

f) we already have in stock enough m2 of neutral grey carpet. Could it be used to cover the walls instead of painting them neutral grey, since the carpet is already lying around and free to use?
The walls do not have to be a neutral grey (not according to RP166 anyways).

g) any other tips would be truly appreciated
From a practical viewpoint, you should also consider how the suite looks like to the client. This can be more important than the technical aspects.

[Glenn Chan]

If you’re looking for accurate monitoring, you’re looking at the following things:

A- Calibrating the monitor so it interprets the video signal correctly. On a broadcast monitor, you do this via color bars. With DVI, you don’t really need to do this since the connection is digital.
However, you do need to make sure that your NLE is displaying the colors correctly. With FCP (from what I vaguely remember), you can get different colors (in the viewer) between paused and playing.

B- Calibrating the monitor so that it tracks grayscale well (i.e. black and white images don’t have color casts in them), is of the right gamma, and is of the right color temperature. The gain, bias, gamma controls in most (but not all) broadcast monitors do this. Calibration probes by Sony, Ikegami, Minolta automate this process. These controls aren’t going to fix everything… after a while, a CRT’s colors will get too whacked and should be replaced.

C- Color gamut differences. If no color management is in place (which is generally the case with video), then the color gamut is whatever the display’s natural color gamut is (for CRTs, this is determined by the phosphors). The ideal color gamut for HD is the one that corresponds to Rec. 709 primaries. I don’t believe there are any CRTs that do this… which is unfortunate.

There are LCDs which do Rec. 709 (i.e. there’s a menu setting between it and other color gamuts)… but lower-end ones like the Luma series LCDs only hit that color with a huge margin of error. IMO, the Luma series should be avoided for this an other reasons (i.e. even Sony recommends their CRT lines).

D- Image cheats. With computer monitors, I don’t think you have to worry about this.
With consumer CRTs (TVs), they do things like red push, scanning velocity modulation, flesh tone ‘correction’, and maybe some other things.
Broadcast monitors don’t do this stuff.

[Glenn Chan]

For many fields, people with degrees definitely make more $$$ (medicine, law, engineering especially). However, for film and video the degree may not help you very much. For example: video/picture editing. Your clients will almost never care if you have a relevant degree. The degree might give you a little edge in getting an entry-level job… employers will prefer those with some experience. However, you don’t necessarily need a degree to get an entry-level job and work your way up. At the highest ranks in editing, it doesn’t look like people with degrees dominate.

If you assume that the highest you can go in editing is facility ownership, then I know of at least two owners who didn’t finish their degrees because they failed a course (one for Humber college/technical school, the other for Ryerson University/liberal arts education). A liberal arts education doesn’t really seem to help in this regard (as the first example shows). Even if you look at the Fortune 500, many CEOs don’t have an MBA (i.e. Bill Gates, the richest man in the world, dropped out of undergrad).

If you look at all the greats in particular fields, there’s little correlation between success and higher education. Except in math and science, where you need education to give you a foundation of relevant knowledge.

2- The saying “learning how to learn” doesn’t even make sense. If you didn’t know how to learn in the first place, then how would you learn how to learn? Perhaps this oxymoron is intentional, but the saying doesn’t exactly prove itself.

In my opinion, the best school can do for you is to provide you with base knowledge. And even with this base knowledge, you will need further skills once you get a job. There’s a lot of things that don’t get taught in film/video programs. The biggest skillset they don’t teach you is how to work with people (other than learning by doing, which you could’ve learned yourself) or how to get clients.

If you learn how to work people (i.e. be persuasive, be easy/enjoyable to work with), then you will likely make a lot of money. Editors who are hard to work with will not have clients and soon be unemployed. Or if you want to be a producer, it’s really important that you can convince others to invest in your projects (even in terms of your crew- you want them to ‘invest’ 110% effort into your show).

[Glenn Chan]

Many standards for color reproduction define colors using chromaticity co-ordinates. The SMPTE C, EBU (like the EBU phosphors), and 709 standards all define red, green, and blue as specific chromaticity co-ordinates.

Eventually everyone is supposed to move to 709, which is a compromise between SMPTE C and EBU. I believe very little work is colored with for the 709 chromaticity co-orderinates (i.e. you can’t even buy a CRT monitor with 709 chromaticity coordinates).

Charles Poynton’s book has information on things like this.

2- Of the companies producing ‘broadcast-grade’ LCDs, I’ve heard of:
sony
panasonic
ecinemasys
cinetal

Most or all of these products have been demo’ed at trade shows already.