How FCPX handles color managemet

All,

An excellent article from Philip Hodgetts and worth a read

What is the secret to Final Cut Pro X’s color management?

So, I finally read it. Can someone explain to me why, if it's so "up-to-date" and optimal, FCP X still supports only color profile v2 matrix and not the much more accurate v4 tables that my calibration hard- and software are capable of (I know, it's an Apple OS limitation, but this could/should have been fixed in Lion, no?

I get a bit curmudgeon-y when someone goes on about how straightforward video calibration was in the days of PAL/NTSC. This is simply uninformed. No matter how good the broadcast monitors, you still had to take a "best guess" at how the result was going to look on a (non-calibrated) home TV, especially in terms of black levels and chroma blur. Many of these issues have become less severe under digital, but are still there and sometimes even worse in terms of the number of different technologies you have to bear in mind for end-use.

Or am I wrong? Is there something rally simpleI ought to be doing so everything that comes out of FCP X looks "good" everywhere else?? I haven't been in the industry for a while, and would sure appreciate a guide to digital colour correction for general (DVD, Blu-Ray, Projector, Web ...) use. 'Cause as soon as I push my chroma and luma anywhere near the popular "CSI Miami" (say) look, the results become very unpredictable. And I've looked at the histograms of those files, but am still a bit baffled.

And no, unlike what the writer claims, the Apple monitors are not a useful reference, because they tend to be way too bright & flattering. I've had to calibrate mine way "down" to get anywhere near reality.

Btw., here in Germany the Squid2 and BasiCColor calibration software are quite effective at the low end of colour correction; but as I said, the frustrating thing is that they are able to output better profiles than the Apple OS can actually handle!

Apple's color sync is about making sure all apps handle gamma the same, and we don't get color drifts between them. In no way is it a replacement for a professional broadcast monitor. Period.

As for most FCP users not using broadcast monitors, a- how does he know that to be a fact, b- ALL the film & broadcast facilities I know of DO use broadcast monitors.

A chat with the folks at FSI confirms ColorSync will improve color accuracy for computer monitors, it can not in any way make them replace production monitors.

And today's HD broadcast monitors, such as FSI's, don't lose calibration like they used to, being digital devices now.

Fair enough, but since I'm a bit behind in my knowledge of that world (I quit professional broadcasting before the advent of HD), I'd love to be enlightened further:

What does a "professional broadcast monitor" mean nowadays, especially in/for the "guerilla filmmaking" world. How crucial is it really, and what does it really tell us about what our edit will look like "in the real world" (on the web, on a digital medium/TV /projector-- via dvd/blu-ray or media server -- , and on an analog TV? Do broadcast suites now use both analog and digital monitors side by side, and how do they decide which levels to conform their footage to? or do they create entirely separate media for each possibility? How can a saturated-to-the-hilt digital video still look ok via SCART, and how does a "PAL/NTSC-safe" video not look way too grey on a digital TV? This baffles me. For I have certainly watched digital episodes of "CSI Miami" that would cause an analog TV to implode.

Back when I was creating computer animation for broadcast, one problem was of course that I could come up with brilliant, unnatural colours that were way off the scale. But the bigger problem was that even iron-clad "broadcast safe" animation bled all over the place on a home TV, and you could never get the blacks to match completely, no matter how good your equipment (there are movies made with cutting-edge 80's & 90's gear that drop their "professional" knickers when remastered on DVD).

So back then I'd always end up doing the final "acid test" on the worst composite-input TV I could find, rather than completely trust the beautiful rendition of the component "pro monitor". And let's face it, even at that time, at least in the boutiques I worked at (never was part of a really major outfit, maybe they did things differently), color bars and 'scopes were only useful to make sure your output was "legal", they told you squat about whether it was going to be what you wanted when it ended up at home.

Now that most or all of the pipeline is digital, I'm not sure I buy the "pro monitor a must" argument entirely (though of course it would be nice to have one!!), because:

Even a pro monitor has to be calibrated to some other arbitrarily agreed upon standard beyond the reaches of Colorsync but probably not (yet) universal. Which would seem to mean to me that I COULD get my lowly monitor and software scopes to tell me what I need to know, even if it doesn't actually LOOK the way it will "out there"; but really it only needs to look the way SOMETHING ELSE looks that IS within the proper -- or rather, desired, since the "legal" limits are now dead simple to conform to -- specifications.

Since my output is digital, I should be able to get one or several digital representations of "good" levels for a particular look and conform my footage to that, period. Where are those "digital test pictures"? Anyone?

Obviously I'm not asking, in a few sentences, to be admitted to the hallowed mysteries of Cineon/10-bit conversion/what have you. As an indie, my equipment for the moment hits the I/O ceiling of 4:2:2. But come on, to tell me in this day and age that I could't possibly create halfway decent video without a $5000(?) "broadcast monitor" doesn't ring true, since that's patently not the case in any other area of desktop image creation. Once I know by experience, for example, what a photo should look like on a decently calibrated iMac, I can create repeatably consistent output for print. The same should be true of digital video, no?

Or am I still not seeing the forest for the trees?

You'll never get a program on NPR, Discovery HD, etc, if you're not doing that level of quality control and accurate color correction. Yes, you can guess how something will look on a TV screen, but it depends on the level you're working in. Shape me a national cable broadcaster that doesn't use production monitors.

Computer displays do not work the same as HDTV screens, nor film. Simple fact of life. But if you're not concerned and don't need to be that accurate, you don't need a production monitor. But the high end market does.

I work in broadcast and have been checking spots I color corrected in FCPX with broadcast monitor through FCP7 and have noticed FCPX displays are very close to broadcast monitor. I find FCPX + Colorsync are doing great job showing SD/HD colorspace on computer monitor. For me interlacing is bigger problem. But if you're cutting progressive material and have calibrated your computer monitors I'd say you can quite well trust your FCPX displays. Of course if your destination is broadcast then you should absolutely at some point check with real broadcast monitor.

Best
Tapio Haaja

OK, ColorSync requires you load a specific ICC profile for your specific monitor, and unless you're using a newer Apple Cinema Display, set to it's default ICC, you're not seeing ColorSync do much of anything. And on my HD production monitor, I'm seeing some important differences. I color grade for a living, and would not ever rely on FCP X to show me appropriate colors on a computer monitor. I always go through FCP 7 into Color. ColorSync and FCP X may get there some day, but computers monitors are factually not physically capable of the HSL and contrast accuracy of a real production monitor.

I color graded a promo we recently did as a test in FCP X. Then I took that exported full rez QT file into Color, and via HD-SDI to my FSI monitor, the colors were WAY off from what I was seeing in FCP X on an Apple Cinema Display. So I'm not buying it based on that test.

My understanding of ColorSync, the way I've heard an Apple rep explain it, is quite different from what Phillip claims in his article. Proper calibration tools are very expensive, and a Spyder is only working to match print colors, not video colors. The ICC over rides any calibration you may do, and resets the monitor according to the ICC, so manual calibration doesn't matter anyway.

I'll wait until it matures a bit more, and we'll see if it really holds up. So far, in my tests, and what I know about what goes into an HD production monitor, I'm not seeing it.

I'm going to sit back and wait for some engineering data to show it's matured and actually working as Phillip claims it will when the ColorSync ICC's are made for various makes and models of computer monitors.

There we're at apples and bananas again. It's fine for you to "sit back", while others are creating art right now that wants to be seen. So -- just as was the case in the days of Never The Same Color -- we work with what we've got, and chances are some of that work will end up being a lot more beautiful and memorable that some incredibly well-graded prime time show-of-the-moment. I completely respect and at times aspire to the technical perfection that is available at the top of the professional spectrum, but it's not, unfortunately, necessarily the hotbed of meaningful creative content. And I will not accept the blanket statement that what I (or rather, someone else of greater talent with similar resources) create is just not "good enough" (thank you Tapio for your encouraging assessment). Because that is a value judgement more so than a technical qualification.

For BBC broadcast, the greatest monitor in the world isn't going to get your video broadcast if it wasn't shot on a BBC-approved camera. Which is of course a lorry of bull's locks, but there it is. That's not the point here, I don't think.

Not, will so-and-so network approve it, but, what do I do in practical terms to get the best possible quality I can within the limitations of my equipment in order to create my film. And I strongly believe the answer to be, that it can be a great deal better than the "pros" would like to admit. I've recently seen a lot of (qualitatively interesting) British shows that look a lot like they were filmed on a Canon 5D or "worse" -- poor low-light quality, rolling shutter, over-exposure ... .

I would be very happy to accept solid evidence that footage recorded on a Canon HV40 / d.o.f. adapter / good lens / Atomos Ninja rig and edited in 4:2:2 in FCP X on my home system is somehow inherently inferior to the same thing shot on a Canon XF300 and edited in a big studio, given that I know what specs to aim for and compare to.

But if you're not concerned and don't need to be that accurate, you don't need a production monitor. But the high end market does.

I am concerned about quality, I expressed that. I am not that concerned with whether it measures up pixel by pixel to something shot on a Red. I don't know what exactly you mean by "accurate" -- accurate to what? There is no way I can predict what it will look like on X's TV at home, certainly a dedicated monitor can insure consistency but not that much more -- but I ought to be able to output media that are within the required numerical standards and also really pretty to look at, including broadcast in the context of independent art or documentary film-making. I should theoretically be able to submit it to a prestigious festival and give them a technically acceptable HD screener. Great films have been created on far less "professional" equipment than is available to the lowly "consumer" today.

THAT's the context from which I am asking, are there "standard" samples out there which I might analyze to learn what to aim for.

Oh, and I thought NPR was RADIO. I doubt they'd need much color correction :side: :silly: :lol:

P.S. I managed to get a Benq monitor without hardware calibration, a Macbook Pro and an iMac screen to show me pretty much the same colours with a Squid 2 & BasiCColor software. Of course there are differences, I think the range of error is actually a good thing in practical terms. But as I mentioned, I am frustrated that at the moment Apple only supports the old .icc matrix; I ought to be able to get much better results with the tables. Wouldn't this be a reasonable thing to ask for?