601 vs. RGB Titles for 10 bit

I've received title cards from a graphics house that are either 601 or RGB in TIFF or PSD format. I tried placing the PSD file into FCP, but it wouldn't accept it. My sequence is 10 bit NTSC, so I then placed the 601 TIFF file into my sequence. The titles are yellow text on black background. Problem is I've noticed during fade in/out of the title card, the black background around the titles looks gray. I'm wondering whether they only show as gray on my Macbook Pro LCD, which is currently set to default "Color LCD" which I believe is RGB color space. I assume they will show as black in the video realm. I realize there are different color calibrations, so I tried putting the RGB file into the timeline and the black level matched fine. My timeline setting is "Render all video processing as high precision YUV". Not sure if that's relevant or not. I'm doing an expensive Digibeta output on Monday so I need to get this right. Any info much appreciated!

Thanks,
Alexander

>I'm doing an expensive Digibeta output on Monday so I need to get this right.

That's an oxymoron! Digital Beta is almost obsolete!

You need a broadcast monitor and an IO device to properly monitor your work.

Y'CbCr tiff files? I thought tiff only supports RGB?

Just wondering... Is the black background you mentioned, supposed to be transparent?



www.strypesinpost.com

TIFF is only RGB. Though the terminology is slightly weird here, I think what Alexander is talking about is the difference between 0-255 and 16-235.

If I remember this right, Final Cut (correctly) ignores embedded ICC color profiles. So the color profile assigned to your computer monitor and the color profiles assigned to any imported stills are irrelevant.

When importing RGB images or footage, Final Cut gives you the option of interpreting 255 as either white or superwhite. (The option's in the "video processing" tab of the timeline options window.) I'm not sure of the out-of-the-box defaults, but on my system, RGB 255 is interpreted as 100 IRE in YUV and 0 RGB is interpreted as 0 IRE. If your stills are clamped to 16-235, you can tell Final Cut to interpret RGB white as superwhite, which will put RGB 235 around 102 IRE and RGB 255 around 107 IRE. That's the right choice if you want your stills to match the dynamic range of other footage that extends up into the superwhite, but note that you'll still need to clamp before you deliver your master.

On the other hand, I don't think Final Cut has any facility for automatically interpreting RGB 16 as black and RGB 0 as superblack. If your stills are clamped to RGB 16, then Final Cut will put that at about 10 IRE, and if you want that to be 0 IRE, you'll need to apply a color correction of some variety to drag the blacks down.

The video scopes built in to Final Cut are your friends here; the waveform monitor can show you whether "black" in your stills is really 0 IRE or 10 IRE, and you can decide what to do based on that information.

In general, Final Cut is set up to take full-dynamic-range 8-bit RGB stills and "do the right thing" with them. So your graphics guys should not double-game the system by giving you clamped 16-235 stills. Final Cut will end up double-clamping those to 10 IRE and 93 IRE.

>That's the right choice if you want your stills to match the dynamic range of other footage
>that extends up into the superwhite, but note that you'll still need to clamp before you
>deliver your master.

Hmm... I've never worked with stills clamped to 16-235. I'm wondering whether the gamma curve will match up in FCP.



www.strypesinpost.com

I think Final Cut assumes a gamma of 1.8 for all imported stills. You can change the gamma of any still by changing the "gamma level" property under "item properties."

This is getting a bit more complex than my level of understanding. Perhaps I'll just use the RGB labeled TIFF as the black backgrounds match in FCP within my default color settings, and for me seeing is believing. The background should be black, not sure if that is considered transparent or not. Unfortunately, I don't have an NTSC reference monitor or HDCAM SR deck in my home to do this "right".

No, black is not considered transparent unless you are pulling a luminance key. Are you trying to superimpose titles over a base layer? TIFF images can include an alpha channel, but it's not an automatic thing.

You do need a broadcast monitor, but even without one you should be able to tell if your supers are working properly.

If you give us a lot more specific information, I'm sure we'd be happy to help you out as best we can.

Quote

Unfortunately, I don't have an NTSC reference monitor or HDCAM SR deck in my home to do this "right".

You have an "an expensive Digibeta output on Monday" and you don't even know what your NTSC output looks like? Not good. Not good at all. At the very worst hook up a plain old TV.

'kay...first of all, I can never understand why people don't post up examples on the forum. a still would eliminate a few days worth of "what's your system / we need more information" posts. If I saw the stills, I could tell you in 5 minutes which to use and probably why.

BTW...if your project is not CCIR 601 (which it probably is not in FCP), use the RGB file. If FCP won't accept it, launch it in Photoshop and save as copy in TIFF format.

When life gives you dilemmas...make dilemmanade.

TIFF imports are supposed to be 16-but RGB, not CMYK, not Greyscale, not 1-bit-- other flavors of TIFF that get folks into trouble.

- Loren
Today's FCP keytip:

Toggle Dynamic Trim with Command-Shift-D!

Final Cut Studio 2 KeyGuide? Power Pack.
Now available at KeyGuide Central.
www.neotrondesign.com

Quote

TIFF imports are supposed to be 16-but RGB

Not true...8-Bit / 16-bit RGB both work. I do it all the time (did it just now). It must be RGB. It's the CMYK thing that screws up FCP and other apps (After Effects doesn't like CMYK as well).

When life gives you dilemmas...make dilemmanade.

Sorry, I should have inserted images. Here:

601 Card TIFF

RGB Card TIFF

601 Card PSD

As I've said, the 601 black background looks slightly grey. When I "get info" on the 601 TIFF, it says it's Color Model is RGB. The fade in/out goes over black filler in the intro.

Jeff, I just want yellow titles over black. I previously had yellow FCP title tool over black filler, but during fade in/out, the edges would blur, so that's why I got TIFF's made.

GrafixJoe, my movie has been professional color corrected already and it's 10 bit NTSC CCIR 601. I'm just replacing 4 title cards in the opening credits.

Thanks a lot guys!
Alexander

Apologies for my newbieness, but I don't know why when I use button "Insert Image URL" it doesn't insert the image, so I used links above instead.



Maybe an image can only be Jpeg?

"The edges would blur?" Sounds like you were looking at an unrendered dissolve playing back in real time. Try hitting option-P.

Quote

As I've said, the 601 black background looks slightly grey. When I "get info" on the 601 TIFF, it says it's Color Model is RGB. The fade in/out goes over black filler in the intro.

It depends how the Artist created the card. It probably still is an RGB image, but he just created the black background @ RGB values of 16 rather than 0 and just called it "601" (sneaky Artist )

Quote

GrafixJoe, my movie has been professional color corrected already and it's 10 bit NTSC CCIR 601. I'm just replacing 4 title cards in the opening credits.

...then I don't see a problem = use the 601 TIFF. If your film has been CC'd as 601, then the 601 TIFF should match. But if as you say...the 601 TIFF looks "grayish", then someone is not doing their job for you. Since I am not looking at your film, I can't make that assessment.

When life gives you dilemmas...make dilemmanade.

Just found out on another forum that:

The 601 file has an RGB value of 15/15/15. It is not truly black.
The RGB file has an RGB value of 0/0/0. This is pure black.

So I'm going to go with the "RGB" file. Thank you guys!

One more thing, how do you guys reply so quickly? I mean do you check your e-mails constantly or do you use a special RSS or other software to reply to forum posts? I just use Yahoo mail and Firefox :-p You guys are GREAT!

FYI = "Pure black" is not 601...so if the rest of your project was "professionally CC'd as 601", these cards will not match the rest of your project. If they DO match your project, then you got ripped off on the CC job because it's RGB (0 - 255 RGB Levels) and not 601 (16 - 235 RGB Levels).

Good Luck!

When life gives you dilemmas...make dilemmanade.

I'll relate this info to the graphics guy. I'm gonna go in with the RGB file, and double check with the graphics guy when I go in Monday. I'm doing a 10 bit output overnight so I hope that works. Thanks grafixjoe!

Okay, this is getting really confused. I'm gonna attempt to clear at least a little bit of this up.

Throwing around references to Rec. 601 is not helping you here. Rec. 601 is a big, complicated specification, and part of it defines how YUV color should be represented digitally. The specification says that for luminance, a value corresponding to 0 IRE is represented as the 8-bit value 16, and the value corresponding to 100 IRE is represented by the 8-bit value 235. That's because television is not inherently digital, and values "less than zero" or "greater than 100%" ? superblack and superwhite ? are possible. The Rec. 601 specification maps 0 and 100 IRE to 16 and 235 respectively in order to leave room for superblack and superwhite to be represented without clipping.

But this only applies to luminance. This mapping does not hold for the chroma samples (which use 128 to represent 0), and it certainly doesn't apply to RGB, which is a different way of representing color entirely.

A lot of people new to video learn bits and pieces of this and think "Oh, okay, I'll just clamp my images to 16 and 235 then, and I'm making TV!" This is not correct. In particular, the phrase "Rec. 601" has no meaning at all when you're dealing with RGB image data. So whoever told you they were giving you "Rec. 601 TIFFs" was yanking your chain. Probably inadvertently.

Now, Final Cut assumes that all 8-bit RGB media it's given ? stills, Quicktime movies containing RGB data, whatever ? map a value of 0,0,0 to a digital luminance of 16 (or 0 IRE) and a value of 255,255,255 to a digital luminance of 235 (or 100 IRE). If you give Final Cut something where "pure black" is represented as 8-bit 16,16,16, Final Cut will assume those pixels are really very dark gray, and convert them to accordingly to the digital luminance equivalent of about 10 IRE.

So the rule of thumb ? really more than a rule of thumb; the rule, period ? is not to try to out-guess Final Cut when dealing with RGB media. Give Final Cut full-gamut media, and Final Cut will do the right thing automatically when it converts them to YUV.

This is all in the manual, by the way.

...now I'M confused

Jeez...there are 2 files...one is "grayer" than the other...which to use? Which ever one LOOKS BETTER

I'm not a bookworm...I just make pretty pictures

Good Luck Alexander

When life gives you dilemmas...make dilemmanade.

Well duh, Joey. Isn't that the overriding rule of everything we do? Use what looks right.

I personally have this little fetish, I guess, for knowing what's going on behind the scenes. There's absolutely nothing mandatory about diving that deep into it. I just find it easier to remember which way works best if I look into all the fiddly little details under the hood first. That's just me, though.

I felt like a mortal cringing as Zeus and Poseidon battle it out. I'll just go with my human fear and use the one that looks better :-p LOL

Quote

I personally have this little fetish, I guess, for knowing what's going on behind the scenes.

Some call it a "fetish", my string of ex-lovers driven mad by my inner-geek would say it's an illness.





*Note-to-self* Next one must be an attractive, witty, clever, insomniac geek. Who works in media, skis, scuba-dives and plays computer games.



For instant answers to more than one hundred common FCP questions, check out the LAFCPUG FAQ Wiki here : [www.lafcpug.org]

I'll see if i can add anything to Jeff's detailed explanation...

In a greyscale image, there's white point, and black point. Black point/white point for 601 or 709 is at 16/235 respectively. Above and below these values are what we call sub-blacks and super-whites. I suppose these values are there as headroom, so we can make fine adjustments, but on final delivery, any values in these regions are supposed to be clipped.

Full gamut RGB models defines 0 as absolute black and 255 as absolute white. By default, FCP will map digital black/white point from the digital black value to 16 Y', and 255 to 235 Y'. This is set in FCP's sequence settings "process maximum white as- white". To map 8 bit 255 white point to 255 on the Y' scale, you set "process maximum white as - superwhite". But that apparently doesn't do anything for the blacks, which is mapped to 10 IRE.



www.strypesinpost.com

You 2 flexing your "geek muscles" is giving Alex & I a cerebral hemorrhage. Please make it stop. MAKE IT STOP!!

When life gives you dilemmas...make dilemmanade.

OK so I got the full scoop from my graphics guy today. The 601 file was purposely set at an RGB value of 15 because an RGB value of 0 would be an illegal black. So even though it showed as grey on my LCD, for broadcast purposes it would show as the same black once it's passed through a limiter. I've left the RGB file in my sequence because the grey still bugs me, and I figure any Digibeta outputs will have a limiter anyway. I think Jeff and Strypes mentioned this, but I got a bit lost with all the jargon as I'm not that advanced.

> I figure any Digibeta outputs will have a limiter anyway.

No they don't. There are hardware limiters, but they are not any built in to digital beta decks. FCP automatically adjusts blacks and whites when your source is RGB. You can do that with your usual video footage with the RGB limiter or the Broadcast Safe filter. However, you do have to be careful about excess chroma levels/saturation.



www.strypesinpost.com

Quote

OK so I got the full scoop from my graphics guy today. The 601 file was purposely set at an RGB value of 15 because an RGB value of 0 would be an illegal black.

Which is what I said previously.

DigiBeta machines do not automatically limit illegal levels. You cannot depend on external limiters or "clamps" either. Ultimately it is your responsibility to make sure everything's legal. Back in the day, if you turned in a tape with illegal blacks / whites, the Engineer would throw the tape back at you and say "NO GOOD" and you would be out of luck.

...and I would NOT trust the "Broadcast Safe" filter either...it destroys gradients in your footage.

When life gives you dilemmas...make dilemmanade.

>it destroys gradients in your footage.

It occasionally pixelates your footage, so it's not something I'll advice if you're not doing proper CC. And as Joey mentioned, it clamps your gradients, which is something it is made to do.

I find it handy to guard against passing flashes of light, etc, where using the 3-way will compromise on aesthetics. I mean, occasionally I do want to suppress the blacks, eg. in cases where I have sub-blacks and raising the shadows will introduce noise, or dipping the highlights affect the parts of the image where I still want to preserve that level of highlights. I use the RGB limiter, but I ALWAYS check my footage for pixelation, noise, banding and know when to stop, etc.



www.strypesinpost.com

So then is the black filler between fade in/out illegal then? Because the RGB 0 value seems to match. So would you suggest changing the RGB of the filler to 15 so it's all legal?

Actually...the number is supposed to be 16 (15 is blacker that 16)...and yes, if the rest of your project is this way you don't want a monstrous dip to super-black @ the slugs. It will stand out like a pimple on your forehead.

When life gives you dilemmas...make dilemmanade.