peaking & zebra together: blessing, or...?
 

Just wanted to share more observations from using the V1E, which - unlike the first Sony HDV machines, the FX1/Z1 and the competitor from Canon, the A1 - can display both zebra nad peaking simultaneously. When I had my first V1 for a couple of days, my reaction was wow, that's cool! I missed (or so I thought) this feature for the couple of weeks with Canon; its display being smaller, I thought not having those 2 aids available all the time made it even more difficult to use. I must have not been alone in this, as in the firmware upgrade wishlist on the XH forum this has been the second most important must-have for people.

Now that I'm using the V1 again, I'm beginning to seriously doubt that it is always better to have both zebra and peaking displayed all the time. The thing is that the peaking (even when set to medium or low) can obscure the zebra, and you don't notice it where it appears first, like on those thin, white birch branches against the sky (BTW, it beats me why the zebra doesn't appear on the bright sky first, but on pure white objects - after all, it's the sky that usually blows your picture, not those tiny white objects. Or is it just my unit? Please confirm)

Anyway, it is not so bad to have the zebra always on, and activate peaking only when you need it to focus precisely - just like with the Canon A1, by a single button push. After having tested both solutions, I am quite unexpectedly suspecting this design was not an omission by Canon! I guess I will use one assign button on the V1 for peaking, rather than have it always on! Or at least create a camera profile with such arrangement, and use it whenever exposure is more critical (and difficult) than focus.

I wonder what other V1 users think of it.

Piotr, I don't have a V1 to confirm with, but, as I understand it, Sony's approach to zebras is to span a 5 unit range, that is, with Zebras set to 100, you would see them on anything between 100 and 105 units. That would suggest that you may be looking at sky in excess of 105 units, you'll have some recordable detail up to 110 units.

It would seem to me that you could quickly confirm this on your V1E by playing with exposure while looking at the histogram and comparing it to zebras.

At least that's the way things work with NTSC and IRE units on Sony camcorders here in the US...

With the histogram ON you can forget zebra -- just keep the signal from getting to close to the right.

I find zebra disturbing as I use peeking for finding focus - zebra would mess up the image entirely if I were to use it. Even with peeking only on sometimes its difficult to see delicate parts of the image on the LCD.

However I like histogram feature, by using that I have never failed any bright spot yet. All of them within the acceptable range.
I rather go for slight underexposure and correct darker areas in the post.

I'm not sure I got you right, but do you mean that zebras covering a span (limited at both sides), a situation is possible that one is not seeing zebra on an object with brightness on the right of the span limit, ie. way over 105? That sounds incorrect to me - when I deliberately overexpose, all sky is covered with zebra which denies this explanation.

However, today I lost confidence in zebra (I've been using it since my very beginnings) after I overexposed a shot because I didn't see the zebra; when I re-created it in exactly the same way, the zebra was there, but only on a some of tiny, pure-white areas - and not the sky, which of course was the brightest!

Still waiting for clarification from somebody actually using zebra set to 100% on the V1.

EDIT: Being too impatient to wait for tomorrow's sunshine, I experimented a bit with zebra at 70% (frankly, I hardly ever used it before). Seth - of course you're right: it's not that everything brighter than the treshhold level gets covered with zebra; clearly it's only some *range* of brightness that the zebra is generated over (perhaps 5 units span, as you're saying). But should this hold true also with the 100% setting, then opening iris I could observe the sky being gradually covered with stripes that would than vanish with further opening (going outside the 5 units scope), with the picture still containing some information of the sky (up to 110 IRE, as you say). However, this is *not* what happens when I overexpose with full sunshine; the sky is *not* striped at all untill it is completely blown out - only then the zebra appears! It *does* appear much earlier (ie. before overexposing the whole picture completely) on *white* objects; when they are only minute it's very easy to get unnoticed - and that's what happened to me today. And I must add it happened for the first time, hence my loosing confidence in a tool I've been successfully using for so long... Please forgive me this (perhaps naive) digressing; I'd like to make sure everything is OK with the zebra generation in my unit. Once I'm reassured, it'll be easy to adjust my habits to the way it is implemented in the V1, and perhaps start using the histogram more.

Thanks in advance for any advise!

Steve, even with a perfectly exposed picture, the histogram does show *some* signal to the right of the yellow line (present when zebra is set to 100%). So what really is "too close"? So far, I used the histogram to check the overall exposure balance, and zebra to control extremities. It worked fine, but today I lost the confidence in zebra!

I guess the Canon's setting at 90% would be more useful (Sony allows only 70 or 100%).

Piotr

In certain situations it is perfectly reasonable to over expose parts of the image. You determine what is important to the viewer and expose for that not what the histogram or zebras are showing. They are guides only.

A perfectly exposed shot for the audience may look under or over exposed to the metering circuits of the camera. This is _the_ skill that everyone needs to learn to be able to interpret what zebras or other forms of metering displays are showing and how that affects the image being recorded.

E.g exposing for the face of an interviewee and having to blow out the background.

TT

Tony, I absolutely agree and have been doing that for years intentionally. It's just that today I observed a kind of the camera's behaviour that is different from everything I was used to - please see my edited comment 3 posts above (I must have typed it while you were writing your answer). Thanks!

Sony's 70 zebra _is_ a range 70-80 and is designed for exposing for caucasian skin tones. This allows you to expose for the skin without having zebras over the whole image. I wish the Canon employed this feature.

The 100 zebra is a limit line and anything on or above 100 IRE is zebra-ed.
From you description your Zebras are working fine.

On the sky blowing out before the zebras appear point. Have you got a cine gamma selected?

TT

Thanks Tony - so I was right the 70% ("skin") setting is a range, and the 100% setting is a treshold (or a range, infinite or open to the right). BUT, is it normal than I can see zebras on those thin, white birch branches against the sky *BEFORE* they start to appear on the sky itself, and I'm getting an overexposed picture?!!

That would seem quite normal...

Are you using cine gamma settings?

TT

You are reading in my thoughts. Yes, I use different settings (not just gamma, but also knee), and am planning to methodically check this behaviour tomorrow with various settings combinations. What is your suspected influence of cine gamma on this?

Just something I noticed on the Canon and I wondered if the same issue was seen on the Sony. Cine Gamma 2 (Canon) blows out detail well before (probably at 90%) the 100% Zebras appear.

I never use these settings as they reduce the dynamic range far too much. You remarked some time ago that Canon footage lacked the Sony's "zing." I am pretty much sure this is down to people's over use of the cine gamma settings. Because they blow out so quickly it necessitates under exposure hence a lot of bland footage.

TT

OK - I'll report my findings tomorrow; good to hear this behaviour of zebra is not a malfunction. Cheers!

Cinegamma curves as pioneered by Panny on the DVX generally increase dynamic range, not reduce it. I have not seen this Sony series tested, but the DVX and HVX attain maximum latitude in Cinegamma mode.

Panny does cinegamma by keep curve nearly straight but changing normal video bias (check http://www.adamwilt.com/24p/index.html#GammaSettings for curvecharts)

I've seen low end consumer cams "cine mode" that crushes blacks - that would reduce latitude. I'm assuming Sony would not reduce latitude in cine modes in a pro cam but worth testing to see for sure.