From my tests, zebras at 100% are showing exactly when your recording has clipping in all three of the R, G, and B channels. (which is equivalent to clipping in the Y channel) You can still get clipping in any one of the R, G, or B channels without zebras showing.

That is the main issue with 100 zebras, it's the same on the 2.5K.

The green channel seems to go the fastest (but it may be my IR NDs, I can't tell really.)

I've actually noticed that even 90 percent isn't protecting you in all cases, and will sometimes drive you to underexpose too far.

mico on another forum said he had been using his waveform successfully and at first I couldn't get that method to work, now I'm using it on my TVLogic, combined with 75% Zebras and pretty much nail exposure consistently. The larger issue comes when using IR NDs, as I have an insane amount of trouble with the Tiffen 1.2 WW IR ND.

Recently tried Schneiders and am almost positive they were better on the camera up to 1.8.

Yep, 100% Zebra scares the crap out of me on both cameras.

/\/\/\ Read above. Pretty self explanatory.

Sorry I am bit confused about the concern over using 100% zebras and overexposing shots?...surely setting your zebras at 100% means you will only be losing the most blown out of highlights and the topend and therefore be closer to overexposing than underexposing?

I thought that being conservative at around 90% zebras would mean the opposite...that you are more likely to make sure nothing gets blown out and be in position where you tend to under rather than overexpose?

My suggestion would be to approach exposure artistically, and try different things. Don't be afraid to make a mistake along the way and blow a highlight here or there, you'll eventually develop a third eye for exposure the more you tip toe that line with experimentation.

Zebras are relative to the recorded video. Changing the display between film and video mode doesn't change what the zebras do.

There's no single correct way to expose with this camera. Every choice is a trade-off, and you have a lot of latitude. But here's my approach: I set the recording mode to film and the display mode to video, zebras to 90 or 100%, and then I just try to get the image to look as good as I can on the screen, erring on the side of overexposure, but of course looking out for zebras on anything with saturation that I want to preserve. I use ISO 800, or ISO 1600 if things are too dark on the screen at 800. With your camera set this way, you'll be recording three or four stops above what appears as white on the screen, and you can dig pretty deep into the shadows as shown on the screen - though of course they are noisy if you dig very deep. Blackmagic's given us very little in the way of tools to consistently get exposure, so it becomes a matter of getting used to the way the screen responds and how that correlates with the recorded video, and adjusting levels shot-by-shot in post.

There's definitely a wrong way to expose with any camera. :)

@spacewig

Yes, I would agree that a meter is the most reliable way to expose in a controlled environment.

Shooting LOG or something linear uncompressed RAW especially, if you do not have displays with accurate and project specific LUTs loaded onto them.

If you're running around just grabbing footage it's not as easy, so you have to rely on other tools, experience being one of them of course.

What balazer says about display video is a good practice for the 2.5K as well, and has been how most if us used that camera. Same applies to the pocket camera with an adjusted application, because it seems as though the video LUT and screen together actually get you close to what a rec709 curve looks like, with about half a stop over something that appears clipped around 80 percent if you're using zebras.

I do think you can harm yourself by feeding the camera too much light, or too little. This is the same for the 2.5k raw or Prores. But, you do have a range that you can rely on:

In my experience, adhering to a bracket of -1.0 - +1.0 exposure yields the most consistent and pleasing image, with either side of the bracket trading noise for highlights or highlights for noise.

Stepping out of that bracket in either direction seems to put you in a place where you've got excessive noise (which actually looks okay on either cAmera if you have gone beyond -2.0) or individual clipped channels which leave you with strangely colored highs that you must balance -- soft clip.

The cool thing is that the camera still produces a very good image even in the event that you've starved it two stops or so, mostly because of the color that it retains: try pulling almost any other format out of the grave and see what happens to skin tones and greens/cyans.

RAW will open up a new image as well, or more data, but IMO will still want to be exposed within a similar bracket.

I actually do not have a light meter, as mentioned before I'm using a combination of zebras and a waveform, and like the 2.5K in a controlled environment, an external studio monitor

All of that helps to nail exposure, leaving the issue of appropriate amounts of IR filtration to be solved.

@kholi solid advice, I was hoping the Rec709 video setting would be more usable on the pocket cam. I occasionally accept contracts for some ultra low budget web videos, often for not-for-profit organizations or the like (I'll go solo or with a +1 crew) and any way I can save time grading in post on tiny budgets means I have a better margin and more time to play afterwards! I'd love to shoot in film mode all the time, but realistically that could be a black hole on some smaller projects. I could always fall back on ol' standby: trusty MoonT7 hacked GH2, but I would prefer to take advantage of the beautiful ProRes 422 codec and the extra DR every chance I can get.

Few points:

Everything above 75% would read, but you don't necessarily have to say that it's "blown out". That's just a readjustment of how you're using the tool. It's been a while since I've touched an HVX, actually last time I had one in my hands I couldn't figure out how to turn it on... but the HVX has two zebra levels that you can set as triggers, independent of each other. W were using those as skin (I believe 70%) and then overexposed areas (around 95... it's been a while). You actually wouldn't see the zebras detail anything exposed beyond 70%, just whatever was in the 70% range.

By setting Zebras to 75% and making sure that only the brightest highlights on skin or any object in frame that I want to keep are at 75%, I'm insuring that I never lose a single channel on that surface.

As @balazer explained, 100% doesn't actually tell you when a single channel is clipped, just that more than one has not clipped. That is very dangerous, especially if you clip a value in skin. You cannot recover clipped information, but you can reduce noise, clip levels simultaneously, etc. if you have the information.

Example: Girl has on shiny make-up (hopefully not SPF based) and you expose just below one hundred percent on her face. When you get back to Resolve, you notice that her skin refuses to balance out, there's a strong magenta cast at the highest range and the more accurately exposed she is, the worse the magenta cast. That's because you clipped the green channel.

Second to that, exposure is only scientific to a degree, you need to learn how the image responds to over and underexposure by shooting a ton of real world footage, then apply that while using your tools as guidelines, not as hard rules. A meter is great for ratios, but can also be used as a guideline to keep an image over or underexposed by a certain measure. People are trying to save skies in footage without the proper hardware... this isn't a good idea, IMO. 13 stops? That's a lot, but you still need grads, polarizers, etc.

Lastly, this is why you've seen what you called "bad footage" all over. It takes conversations like these and a lot of mistakes to actually nail the most important portion of getting a great image. No one picks up a digital camera day one and poops gold. =] When you do get your camera, you'll have a lot of knowledge early users did not.

At 100 percent it's supposed to be sensor level clipping.

Setting zebras below 100 percent shows the range from that setting up to 100 percent, so 90-100, 80-100, etc.

If you set the camera to 100, anything with zebras = clipping at the sensor.

HOWEVER, it does not tell you if you've clipped a single channel earlier than the others.

You can set to 100 percent, expose without zebras, and still clip a channel. That is the issue.

I'm using a Waveform now in conjunction with 75% to set exposure at a certain stop (if shooting for a certain stop), so 75% is fine for me. I'm largely ignoring the zebras as long as they aren't crawling all over something that I don't want to lose to highlights.

As for the underexposure, if you rely on 100 percent zebras to tell you everything it's likely that you'll either under or overexpose footage as well.

Just comparing my Ikan waveform to Ultrscope to see how accurate it was, I've noticed that when zebras are set to 75% it seems I just have a stop or 1.5 stops left till I clip the sensor. This is on both waveform monitors too. They are dead even in terms of exposure. I switched over to 100% zebra and it seemed that was about 1 stop+ away from 75% as well. Is this what you're finding @kholi?

Edit: Shit, probably wrong topic. I'm on the 2.5K of course.