Gee, how dumb, I posted this on the Apple FCP forum as well as the Cow. Too many messages, not much interest I suppose. I completely forgot all the PROS hang out here! Searched and read some old FCP monitor/colorbar/IRE/setup threads, but did not come across what I am specifically looking for. Just bought a new 15" JVC NTSC monitor, and once again started to pay attention to colorbars, etc.

6-23-05

FCP 4.5/5.0 Colorbar BUG - MYSTERY

Can anyone confirm "odd" behavior in FCP 4.5, also seen by a friend in FCP 5.

1) Using a straight forward new DV-NTSC project. In FCP “Viewer”, bring up the “Bars and Tone NTSC” Color Bars using FCP's built-in generator. Open a new window-tab, Waveform monitor only, set the drop-down to “viewer”. You should see the bottom left “purple” square line up at approx. 7%, and third from the bottom left, the “dark blue” square line up at approx. 5%. This seems correct.

2) Insert these colorbars into a DV-NTSC sequence, switch the waveform drop-down to “current frame”. Both of these square bars now drop down to the ZERO percent line on the waveform monitor (all the others are in their proper position). To confirm, just double click this clip in the sequence which brings it back into the viewer window. If you now switch the waveform back to “viewer”, there is NO change, both bars remain at the zero mark. This does NOT seem normal.

ADDITIONAL TESTING:

3) With the colorbar clip from the sequence still in the viewer, export to a STILL image, choosing the “.psd”, photoshop format (you can choose the TIFF format as well, results are the same). Bring this still image back into FCP, and again examine in the waveform monitor. Both color squares now appear back at very near 5% & 7%, their original positions in step one. This closely matches what you see with the original viewer generated “Bars and Tone NTSC”, seen in the viewer only.

I am puzzled, why do ONLY these TWO FCP generated colorbar squares that appear in the waveform monitor, drop down to zero black when placed in a sequence? How much can I trust the software waveform, or is there a bug in the bars themselves? Is this some kind of remapping or RGB issue?

Actually, you hit quite a number of "normal" issues. No bugs that I know of.

First, my favorite problem, When you export or import stills with FCP, it automatically ads a gamma brightness curve on import and subtracts it on export as well as converting to 0-255 colorspace. If you were expecting colorbars to export to a "normal" RGB signal, I'm here's to tell you it doesn't work that way--at least not in Mac land. If you export and then import the colorbar signal, that picture has been through the gamma brightness conversion twice, once up and once down. Damage ensues.

There are some parts of the colorbar signal that aren't natural. You need to know that everything below the split 2/3 of the way down is magic. All but one of those signals are highly illegal for different reasons. The pluge pulse (three gray stripes in the lower right) can never exist in nature. The darker of the three is illegal and only exists for the purpose of setting up a television monitor. If you convert colorbars to an RGB signal, the dark bar will vanish.

The "blue" blocks in the lower left are also highly illegal. They correspond to the two color modulation axis which you totally don't need to know, but you do need to know that they represent the color signal without the naturally occuring luminance signal. They, too, will get distorted if you ever try to convert the bar pattern to RGB color.

Some video systems will not pass colorbars because they naturally assume everything below black is illegal and chop it all off. The upper two thirds of the pattern should make it through just fine because all those colors can be had in nature.

Koz

Thanks for the quick and accurate response. I have gone as far as re-opeing the old Synthetic Aperture PDF that came with the old OS9 Test Pattern Maker, still works under Classic.

Read this:
"The lower left portion of the pattern consists of three bars of a -I, pure
white, and +Q signals. Because -I and +Q cannot be represented in
RGB, these bars are not accurate. However, IQ color encoding has
fallen out of use and these bars are therefore not of much use anyway.
The 100% white bar, however, is accurate."

Some of the "problems" are for someone like me, were all this stuff has always made my head spin, are that FCP and it's relationship with RGB/Photoshop has changed since the early versions, and I go back to the Mac "EditDV". It's much smarter now, making me much DUMBER! I would have hoped Apple could have included some techie explantion in the FCP PDF/manual, maybe they don't think we would notice. So as not to get too deep into it, can you give me your opinion of the accuracy of the FCP colorbars, overall? I read the old thread, and Grame asked this question in 2004 a few times, and I assume there are various opinions. All in all, I have far too many tools to setup my monitor for both DV and analog signals, none are too expensive, the most being the "Calibar", which puts out true SMPTE bars, but if one is shooting for DV, you have to make the "adjustment". Back on the Mac, in FCP 4.5, it's nice and quick to have the built-in bars and they seem to work fine for monitor setup, but once I saw the missing blue bars on the waveform, I became concered all over again about their accuracy.

Thanks again for your help. I should have come here FIRST!

<<< I should have come here FIRST!>>>

Let that be a lesson to you.

The FCP colorbars are first rate as far as I've been able to tell--and I can tell pretty well. They do rather depend on you not trying to convert them to anything else.

For example, if you push this signal off to a digiBeta tape, they will be pretty much perfect when they get to wherever they're going. Same with MiniDV and the other television formats, even over FireWire.

This is one really big reason the televison signal doesn't go from 0 to 255. It goes fom 16 to 235. The pluge signal uses that 16 units below black to put the "illegal" black stripe and the two oddball blue patches. The difference between 235 and 255 allows for superwhite. The system can handle that even though you can't broadcast it in analog.

The instant you export the timeline in any format that doesn't allow for "black and than black", you're dead as far as passing the colorbar signal correctly.

Koz

<<< "black and than black">>>

Make that "Blacker Than Black."

There are quite a number of things that the television signal hides below picture black.

Koz

I noticed that the dark blue looked to be the same luminance value as the purple on the re-imported still frame. That's about right since I know that a luminance shift occurs when exporting still frames from FCP.

As far as the rest goes, I took some looks with both a color bar clip and a regular piece of video. In the video clip, the identical frame was not identical in the eyes of the waveform monitor, but the shift was not nearly as noticable as in the color bars. Still, when I've output from FCP and then digitized in another machine using the FCP bars on tape and calibrating to them, I haven't noticed any specific darkening of the frame. Perhaps with the bars, it's a display issue of some kind.

Calibration and CCing are probably best done using external WF and Vectorscopes, but the internal ones have worked well for me up to this point.

Andy

That's what I get for starting a response, eating dinner and then coming back to finish. Thanks for the great explanation Koz. I feel like I understand color bars more than I have before. And also image export. Which, BTW, is a pain for me occasionally. How do you handle it if you are exporting stills from your timeline to use as menu backgrounds in DVD Studio? Is there a specific amount you should adjust the pic in Photoshop so that the blacks don't look so high on the TV?

Andy

<<<I know that a luminance shift occurs when exporting still frames from FCP.>>>

Not exactly. It's a gamma shift.

[kozco.com]

Note that the white and black portions of the waves are all the same. It's only the middle grays that shift. That's one reason you were never able to totally make up the difference with straight gain and black level.

You can't use the gamma tool in FCP or the middle slider in the Three Way, either. You need to use the gamma control in Video Filters, Image Control, Levels.

Oddball, little known facts.

Koz

There's also a gamma control in Photoshop that can be made to work.

Image, Image Adjust, Levels. The middle number of those three numbers (1.00) is gamma. Pick 1.2 or 0.833 depending on whether you need to go up or down.

Same numbers inside the levels control in FCP.

Those two number seem to be the exact conversion for the damage that FCP does.

I gotta check the 0.833 number. I'm very sure of the 1.2.

Koz

Well I'm glad this colorbar message has inspired some great feedback, cause I feel better it's NOT just "me". I realize the software based "bars", and "waveform" cannot be compared to hardware scopes in terms of total trust and accuracy. I just wanted to report and field the question I did, since visually, you can see the immediate drop off of the two color bars, in the FCP waveform, going from the viewer to the sequence, and this is what surprised me. So one or the other is OFF as compared to what it "should be" based on what all you guys are saying, some of which is still a little deep for me and I'm sure a heck of a lot of casual users, then again, they may not notice. These are times I wish I were BLIND!

<<<These are times I wish I were BLIND! >>>

I probably wouldn't go quite that far.

If you look at the waveform monitors for both services, you may find that they are different. One goes below black and the other doesn't. This doesn't mean the bars are wrong, it just means that particular part of the video service won't show them to you.

If you export them to a tape machine or monitor system, they should be OK. Please note that you can't do this through most camcorders or decks. They all distort blacks pretty seriously.

Hey, if this was easy, everybody would be doing it.

Koz

All the colors are come through to the NTSC monitor, so calibration is not an issue. I do go out firewire from FCP to my old Sony DHR-1000 DV deck, still works great, and then SVHS cable to the new JVC TM-H150 monitor. Anyways, a lot to digest in such a short thread, now you see it now you don't. Appreciate all the feedback and help.

It's a shame many of the old forums have been overun so to speak, even old 2-Pop links still seen on the "uemedia.net" page, still buttons, but dead ends. Never found much support on the Apple discussions, messages scroll away faster then the purple color bar.............

I miss Ralph F., thankfully this place is going strong, and was a bookmark in my Safari browser, just have to delete all the others so as to come here first next time, and not waste time. Thanks again, I'll revisit this issue if need be.

One of the things I haven't seen mentioned is the very big difference between digital video and analog video, which is the only reason you need external scopes.

Koz is basically (well, completely) accurate. Only thing to add is that digital video has black at zero IRE, as does NTSC video outside North America and PAL everywhere in the world (well the equivalent in PAL which does not use IRE).

But NTSC (ie analog) video in North America has 7.5 IRE Setup for historical reasons. This means that the levels you'll see and use in FCP (which are 100% accurate) are not the same as you'd see in an analog environment. (FWIW, only analog, composite video can accurately be called NTSC - we use it as a shorthand for "a digital format that has 525 lines and repeats frames at 59.94 fields per second")

So, you need to be clear as to whether or not you're seeing digital video or analog video as to what to expect. Other than that, Koz is right, Bob, stop worrying and jumping at shadows.

Cheers

philip

All that and in HiDef, the gray is different. In NTSC, they get gray by adding 59% green, 30% red, and 11% blue. In High Def, the numbers are different.

They're better for pictures and they cure some old NTSC ills, but different.

Koz

<<<"a digital format that has 525 lines and repeats frames at 59.94 fields per second">>>

And the shorthand should not be pressed any too firmly, either. PAL-M uses 525,60,30, but wacky color.

Koz

Philip, it's even more complex than that :-) Expect a über-article on black levels at Ken Stone's soon. I'm still finishing off the details, but it's massive and comprehensive. I'm also trying to include details on all the daft stuff FCP does with importing and exprting images.....

Graeme



&lt;[www.nattress.com]; - Film Effects and Standards Conversion for Final Cut Pro -

I'm always "jumping at shadows", it's what I do. Maybe you remember you included my short article on "four channel audio chapture on FCP from the original Canon XL-1", in your first "FCP Intelligent Assistant", 3-4 years a go?

Bringing up "setup" is yet another can of worms. That JVC shockwave video that has made the rounds, has been helpful to many a newcomer. I have battled it for some time, since I have to shoot both DV and analog video, sometimes a mix of both, and setting up a monitor on location depending on the gear is important, same here at home. On occasions, I have a real engineer with real scopes, but most of the time, the colorbars and the zebras are our only friends, or enemies.

Anyway, I see we spurred Graeme on to do some more writing, can't wait to see this one! Like I said before, one can comb through the HUGE FCP PDF manual, and tons of video tutorials and still not get the whole scoop on RGB/luminance/ IRE, etc. So it would be a welcome addition.

I've been working on this article for a while now, and just as I was getting to the bit on still image import / export, this thread comes up, so it's good to know that the issues I'm writing about are certainly not academic, but are vital to practicaly all FCP users. I'm hoping to make it all as comprehensive as possible, and I'll certainly react to any feedback I get once it's published.

Graeme



&lt;[www.nattress.com]; - Film Effects and Standards Conversion for Final Cut Pro -

I can't wait for this Graeme. Particularly the delineation of what Apple are doing on import and export.

Looking forward to the article.

Philip

Thanks Philip! I'm writing it for Ken, and I think he's on the edge of his seat waiting for it too!

But with something like this, I want to get it right, so I'm going to do more FCP experiments to see what's going on!

Graeme



&lt;[www.nattress.com]; - Film Effects and Standards Conversion for Final Cut Pro -