>The bottom line is, both 601 and 709 call for Y' values where the legal values fall into the 16-235 range (for 8-bit).<

Thank you, most helpful, although I could've lived without the confusing sermon about confusion. :-)

how about blu-ray movies?

they use 16-235 Y levels? or some have 0-255?

they use 16-235 Y levels?
Yes.

tnks.

i will change the picture profile of my ex3 to fix this (the default goes to 16-255)

Depending on the gamma curve the EX by design goes to 109%.

Bob.

exactly.

but, if the levels should be at 100% ire, why they go beyond?

would be better a option to set white point (like black adjustment), instead to use a gamma that have "the level i want"

I disagree with cameras going beyond digital white level / ("100 IRE") / Y'=235, but in practice that's what they do. I think it's a bad idea, but that's what they do.

In regards to the EX camera, I don't have it and haven't played around with its menu settings so I'm not 100% sure what they do. But on some cameras, there is a setting to clip highlights and there is a setting to turn the gain down. If you want to get the most dynamic range / minimize noise (they go hand in hand), you don't want to clip any digital values as that would be throwing away dynamic range. So you might want to play around with your camera to figure out what the menu settings do.

2- Another approach is this:
http://www.glennchan.info/articles/vegas/color-correction/tutorial.htm

That's one way to deal with values above white level, and it will work for typical/"mainstream" situations where footage comes in with superwhites / values above white level.

From Sony themselves:




Serena has spent considerable time carefully plotting the EX gamma curves and has posted them on DVInfo. These hypergamma curves are non linear, that's on top of the usual 2,2 gamma.
My understand is that when used correctly white = 100% however there's things brighter than white in the real world e.g. specular highlights. The hypergamma curves attempt to preserve these. What you do with that 'data' depends, there's no simple rule.
If you're simply wanting to put that out into a traditional video 'system' then clipping them off would seem the right thing to do although that does beg the question of why you used a hypergamma in the first place. Where they can be usefull is in compositing to give a more optically correct result.
From what I've seen and understand you also need to be very careful using these gamma curves. Incorrect exposure can give some really ugly results, putting anything other than white highlights into that region of the curve could produce issues of color banding that you'd definately not want.

Bob.

tnks for the info farss.

i also participate the dvinfo forum, and looks like only cine2 limit to 100%.

maybe using other gamma with -3 gain (limit to 104%) would be the "half-fix" for limiting to a better levels. ;)

we will be starting to produce blu-rays in about a month or so, and im worried about every aspect ;)

I defer to Glenn if I'm wrong on this, but my understanding is that though going to negative gains will reduce noise, it will also reduce the dynamic range as well. Basically, it's that reducing gain below 0dB means you are no longer using the full dynamic range of your sensor chips.

I have a Sony XDCAM HD camcorder, and I now almost always use the CineGamma 4 curve on it. It has a nice gentle roll off of the highlights, and it takes a really, really bright highlight to push me over 100 IRE. This happens very rarely. Note that I can't say what curve this would compare to in other cameras, they're all rather different.

Other gamma curves that intentionally go over 100 IRE are assuming that the shooter is going to be doing more sophisticated color grading, where some of the illegal areas may be recovered in the final product. At least that's my understanding of why going over 100 IRE isn't unusual in-camera.