Better Color/Film Looks by Under-Exposing

[Leonard Levy]

The cine gammas are downright strange and they are different from one another. All seem to have an extreme amount of knee compression that sets in very early, so if you were just looking at 100% zebras ( at least with some of them they might not show you much until a good part of your image was in the compressed knee.
Be careful with them. BTW it seems if you decide to alter the gammas even a little bit the curves start altering radically in the knee area.
They are particularly dangerous around flesh tones ( esp cine 3 ) and can easily give you compressed areas in faces.

[Sebastien Thomas]

Originally Posted by Bill Ravens

Guys...

Take a look at this, if you haven't already seen it...
http://bssc.sel.sony.com/Broadcastan...Final_1-08.pdf
On page 8, there are plots of the cine gamma curves. There's no scales, so everything is really only relative. It's a shame they don't also show the STD curves becaue they do so for the XDCAM HD. Wish I could find the URL, but, I can't. (edit, I found it....here, page 9...
http://bssc.sel.sony.com/Broadcastan...mhd_family.pdf

The XDCAM HD curves show is the std gamma in comparison. What you see is a curve that rises steeply all the way up, without curving over like the cine gamma curves do. In fact, the top of the STD curve(the hilites) ends about twice as far up the ordinate scale as the cine gamma curves. OK, that's the XDCAM HD.

When I look at the luma values of the EX1 on a scope, I see evidence of the same thing. The STD curves have about twice the overall range(is that latitude) as the cine gamma curves. So, what it appears to me is that Sony has really compressed 100% IRE in the cine curves, in order to open up the shadows.

Not saying that I KNOW this is why you're blowing out hilites, but, it seems like a reasonable assumption on my part.

Cine curves goes a little above the 100IRE, but in a controlled way.
Standard gamma curves do not controle that and rely on the knee function to handle high IRE, leading to color aberation around highlights.

Cine curves are spreading the sensibility of the CMOS where it is needed, sometines more in the blacks, the middle or the brights.

I don't understand what you mean when talking of "usual bright picture" opposed to film look.
Use STD gamma for a video look (high saturated color, crushed blacks and burnt hightlight).
Use the CINE gamma to get the whole sensibility, without blowing out the hightlights and beeing able to edit in post.

I haven't seen the curve of the EX1 STD gamma, but I bet it is like the one on the XDCAM-HD, which is almost the same as other Cine-alta from sony (like F900).

[Michael H. Stevens]

Originally Posted by Sebastien Thomas

Cine curves goes a little above the 100IRE, but in a controlled way.
Standard gamma curves do not controle that and rely on the knee function to handle high IRE, leading to color aberation around highlights.

Cine curves are spreading the sensibility of the CMOS where it is needed, sometines more in the blacks, the middle or the brights.

I don't understand what you mean when talking of "usual bright picture" opposed to film look.
Use STD gamma for a video look (high saturated color, crushed blacks and burnt hightlight).
Use the CINE gamma to get the whole sensibility, without blowing out the hightlights and beeing able to edit in post.

I haven't seen the curve of the EX1 STD gamma, but I bet it is like the one on the XDCAM-HD, which is almost the same as other Cine-alta from sony (like F900).

So why is Bill saying the STD gamma has more latitude, and I think he said they might even be better for taking into post. I'm still experimenting - I have not even shot a standard yet.

[Bill Ravens]

Michael...

I'm not gonna retract what I'm suggesting, however, I will add the same thing that Leonard Levy said. The knee function is strange. It's understandable that you're dubious. I'm basing my opinion on the waveform displays in HDRack, which kinda knocked me out of my chair when I saw them.

In shooting tests, I'm finding that CINE gamma is a lot more forgiving than the STD curves, Mike. There are situations where that isn't true, tho'. The STD curve images are real "punchy", not terribly pleasant without some CCing...;o)

[Craig Seeman]

Ultimately I wish someone would shoot some test charts and scopes and post the video or shots online because I think the terminology and explanations are confusing some.

[Paul Kellett]

Originally Posted by Craig Seeman

Ultimately I wish someone would shoot some test charts and scopes and post the video or shots online because I think the terminology and explanations are confusing some.

Yep ! .....

[Bill Ravens]

Here's some outdoor shots with my TC2 at 4 gamma preset settings, for comparison. I'll try to get some screenshots of the wfm for a DSC testchart, later, when I have some time.
TC2S1=STD1, TC2C1=Cine1, TC2C3=Cine3, TC2C4=C4

[Randy Strome]

Originally Posted by Bill Ravens

In ANY high contrast scene, you'll blow something with the cine gamma settings, usually it's the hilites. Only the STD gammas allow the full dynamic range of the EX1. Cine gammas are really designed for indoor shooting with carefully controlled lighting.

Hi Bill,
You know way more about this than I, so I hesitate to comment, but this seems counter indicated by your examples above. You are very close to blowing both blacks and whites on your STD example, but have a bunch of headroom on your Cine examples. Did you reverse your initial comment? If not, I am lost, so set me straight.

Also, it seems like in your Cine Examples, a little push to the right might nail the whites and set the darks in great position for a little level/curves tweak in post without introducing a lick of noise.

[Bill Ravens]

Randy..

Can't argue with you. And...not so sure I know more than you. You're quite right.

You say I'm very close to blowing whites on STD...yes, but they're not, are they? I expose so 100% zebra is just beginning to show...as you just confirmed.And it recovers quite nicely with a little levels adjust with the Color Curves FX. The cine curves need that push to the right, indeed! The histogram showed that STD1 fills the scale, while the CINE presets pushed the histogram to the left. In my mind, that's wasted bandwidth. Playing with color grading in post reveals that STD1 can be tweaked with more detail than the CINE curves. And that "punchy" look in STD can be moderated, again, with a little reverse-"s" curve adjustment.Try it.

By the way....the auto expose selected by the EX1 was really effected by where I took the exposure...sky or foreground. It seems to work just like a spot meter rather than an averaging meter. Either way, I didn't like the camera selected autoexposure. I used zebra, as I explained.

[Phil Bloom]

Bill, when I come to your lovely home town of Santa Fe on Saturday I would love you to give me some complete PP settings for shooting outside.

Cheers!

Phil

[Piotr Wozniacki]

Originally Posted by Bill Ravens

In my mind, that's wasted bandwidth. Playing with color grading in post reveals that STD1 can be tweaked with more detail than the CINE curves. Try it.

Seems like a consensus is being reached on this between you and me, Bill :)

[Bill Ravens]

Piotr...

good. ;o)

Phil...

would love to. write me offline.

[Michael H. Stevens]

I agree with Randy that maybe the CINEs could be pushed more, but to my eye I see more highlight detail to work with in the STD. Of course, there may be detail in the CINEs that might come out in post not obvious here. Would be interesting Bill if you took those four shots and gave them some post to get each one the best you could and then post those four.

Typical New Mexico that. I lived in Taos once.

Mike

[Randy Strome]

Originally Posted by Bill Ravens

You say I'm very close to blowing whites on STD...yes, but they're not, are they? I expose so 100% zebra is just beginning to show...as you just confirmed.And it recovers quite nicely with a little levels adjust with the Color Curves FX. The cine curves need that push to the right, indeed! The histogram showed that STD1 fills the scale, while the CINE presets pushed the histogram to the left. In my mind, that's wasted bandwidth. Playing with color grading in post reveals that STD1 can be tweaked with more detail than the CINE curves. Try it.

Yes, you nailed the exposure on the STD shot. In a relatively static shot that may be the way to go indeed because you get that nicely filled histogram right out of the cam.

What had me confused was the part about being more likey to blow something with Cine than STD. I have been considering Cine as "safety settings" for high contrast or variable contrast scenes. In surf shots, I am dealing with random occasional whitewater exposions on relatively dark water. I am just not ggod enough to be that accurate in selecting the perfect setting and need to "compress the histogram" (whites in particular) a bit.

I have been using your PP's (dig them) with Cine's pushed to the 100 zebra setting (waiting to judge based on the worst whitewater explosion) and then backed off a bit. Leaves a bit of room under the darks, but a curves/levels adjust fills the histogram back out.

Definitely some waste as you stated, but less missed shots due to blocked up whites as well.

Learning...learning

[Piotr Wozniacki]

Originally Posted by Bill Ravens

By the way....the auto expose selected by the EX1 was really effected by where I took the exposure...sky or foreground. It seems to work just like a spot meter rather than an averaging meter. Either way, I didn't like the camera selected autoexposure. I used zebra, as I explained.

Very good point, Bill - the auto iris on my old, good V1E has been a marvel tool compared to how it works on the EX1 (both hardware implementation of it, and the measurement logic behind it).