GY for broadcast TV?

[Stephan Ahonen]

Originally Posted by Steve Oakley

there are ways of getting some control, and if you can't at all, maybe you shouldn't be shooting it, at least not significant amounts material.

So giving your editor imperfect footage is worse than giving him no footage at all?

if you can't have some control over what you are shooting, why are you shooting it ?

Because they're paying me to?

you should look at how you can control the light to make it better.

As much as I'd love to drive to every gig with a full lighting truck, I don't think the clients could afford it.

Don't get me wrong, your points are very valid for the sort of stuff you do, but pretty much worthless for the sort of stuff I do.

[Werner Wesp]

I think you're sharing the same pov, just not sharing it...

If you're shooting for a feature film, you woudn't be shooting anything that you can't control. That would make no sense indeed. In this type of situation you have time for organising a set/decor, time for re-shoots, ... Why not use it?

If you're shooting a live event, no-one expects it to be perfect. That's why it's live, it only happens once etc (that doesn't mean your professional honor will get you to try to make the best possible images of it). You can still make the best of it, but it wouldn't yield the same results as the careful lighting and planning for a feature film would enable you to achieve...

[Keith Winstein]

Originally Posted by Steve Oakley

well I don't have a spec for 709 colorspace in front of me... but - first if you transfer film to tape, your in the same color space.

My point was that with film, you have extremely fine color gradations in the source material and can adjust colors in post and do all your color adjustments _before_ having to put it in the ~2.7 million color colorspace of BT.709. (E.g., with a DaVinci telecine.)

Whereas with HD100 (recording to tape), you're stuck with BT.709 colors at the time of recording. You can adjust the colors, sure, but your source material is stuck in that 2.7-million-color space already. Then at the output, your transformed colors have to _again_ be snapped to a BT.709 color.

So there's at least two generations of BT.709 color quantization when you acquire in 8-bit BT.709 (like the HD100 and DVDs and HDTV use), whereas film capture -> HDTV or DVD means you only need one generation of BT.709 color quantization.

if 100IRE=235, then 108 IRE=253RGB, therefore 253*253*253=16.2 million colors so I'm not sure about that statement of 2.8m colors. to get so little color you'd have to be limited to around 145 values per channel from back to 100 white.

The problem with this calculation is that most of the combinations of Y, Cb, and Cr do not yield legal colors. For example, if Y is 16, essentially the ONLY legal values of Cb and Cr are 128 and 128. All other combinations (e.g., {Y,Cb,Cr} = {16,16,16}) correspond to out-of-gamut colors with R, G, or B less than zero.

If you go through all the possible combinations, it turns out there are only about 2.7 million legal 8-bit BT.709 colors.