Black Level and Zebra Pattern (Please exercise patience)

[Robert Goodrich]

Two questions to help me understand black level:

 

Is the NORMAL black level on the XL-2 0 or 7.5 IRE?

 

Having seen recommendations that if your intended delivery is broadcast to set it to 7.5, but if you intend to deliver it on DVD or for a film-out you should set black at 0. Could someone illuminate me on why this is? And if (okay it will end up being more than two questions) NORMAL is 7.5 and I want to distribute my movie on DVD, what is the setting to get it to 0?

 

Zebra pattern question:

 

I have chosen for caucasian faces to set my Zebra at 90, and then back off the iris until the last of the zebra pattern disappears from the highlights on the face. Is this sensible? Or is there a better method?

 

Extra credit question:

 

Assuming I'm making a DVD, do I need to make adjustments to my field monitor for 0 IRE black? Excuse my terminology, but right now the brightest of the pluge bars is visable, the next one to the left is set so that is just at the point where it can not be seen. And for contrast I have it so that I stop adjusting just when I see the white start to darken. So in short, if I want to shoot at 0 IRE, should I set it so that the brightest of the pluge bars is "just not visable"?

 

Thank you

[Michael Nash]

We've discussed this here before and there was a link that explained it pretty well, although I have no idea what the link was. Try a search of this site.

 

But basically, "0" vs. "7.5" IRE is determined more by the post path than by the destination format. With most edit systems you can lift a 0-black picture up to 7.5 for broadcast, or crush a 7.5-black picture down to 0 for digital delivery.

 

What becomes important is that you calibrate your monitor properly for the level you're shooting, so that you see the intended picture, and then make sure that you maintain the proper level through post. You don't want to add extra "setup" in post to a picture that already has 7.5 IRE black; nor do you want to crush a 0-black picture down any more.

 

Setting up you monitor can be a little tricky, depending on what the hardware does (and I'm not familar with bars on the XL-2).

 

http://www.greatdv.com/video/smptebars2.htm

 

I work in the broadcast world so for standard def everything is 7.5. But most prosumer cameras and edit systems use 0. HD uses 0.

 

As far as zebra, it depends on how you have the contrast of your camera setup. 90% sounds pretty high for a reference of Caucasian faces, unless the camera is really contrasty. In broadcast, I use 70% and get just a "tickle," if any, zebra on faces in mild lighting. If the person is really pale or the lighting is high contrast and I'm trying to spilt sun/shade exposure, I'll get patches of zebra on the forehead and cheekbone.

 

But like I said, it depends on how the contrast is setup. With the SDX-900 filmlike gamma for example, I don't see any zebras on Caucasian skin tones at 70 IRE.

[Alessandro Machi]

I think Zero Black and 7.5 IRE set-up black is the most misunderstood technical issue in the world of video.

 

I further believe that editors and cinematographers don't actually completely understand it either because their worlds conflict over how this issue should be handled.

 

The DP in the field is supposed to make the picture look as close to the final result as is possible, but sometimes this is unnecessary IF the editor and the post production people know what to do. Keeping the set-up level slightly lifted ensures that nothing is unnecessarily crushed at the time of shooting and further ensures that post can than crush as little or as much as they need to for the final overall look.

 

The problem is the dang producers who don't get it and demand the picture look perfect, on set. For high budget productions, this is protocol and is the proper way to proceed. However, as the production budget drops, money can be saved by spending less time trying to make the picture perfect. This is primarly achieved by reducing contrast at the time of the shoot with the knowledge that the black level and contrast will be readjusted in post.

 

Contrast reduction acquisiton however does not free up the DP to be careless in their lighting contrast decisions.

 

The people that need reduced contrast the most are the people who never do it, namely those who shoot with a low cost digital camera that creates zero black when it should be adding 7.5 IRE set up.

 

The following advice is worth a million dollars, seriously. I discovered that some low cost consumer video cameras have a menu setting called "spotlight", which actually creates 7.5 IRE. This is a gold mine for anyone shooting with these low cost digital cameras.

 

The difference in the usefulness of a picture shot in natural conditions at 7.5 IRE versus zero black is practically mind numbing. The spotlight setting (aka 7.5 IRE set-up) kicks ass when shooting in any real life condition when compared to zero black, yet 99% of the camera world thinks that zero black is somehow better.

 

Zero black is only better when the monitor can actually handle all of the different shades of gray and make them all stand out equally. 7.5 IRE simply massages the lower end of the gray scale spectrum and actually expands that lower end I.R.E. spectrum into gradients that can either barely be seen, or pure black. 7.5 IRE set-up is a form of fine tuning that takes into account that most monitors currently on the market do not actually have the dynamic range to play back lower end gray scale accurately, especially color video.

[Robert Goodrich]

Thank you two. It's been an education.

[Simon Wyndham]

Hmm.

 

http://www.kenstone.net/fcp_homepage/video...s_nattress.html

[Alessandro Machi]

Sad to say, but as technically accurate as that article may be, it's not the full story.

 

The issue of video acquistion in real world, real life videotaping situations can be IMPROVED if set-up is correctly integrated at the time of actual camera acquisition. Real World contrast values are usually above and beyond current digital camera contrast specs and means that the slight massaging of the lower end I.R.E. scale which will actually give one a more USEABLE image to work from.

 

If set-up is correctly applied when shooting with lower end digital camcorders parts of the image that would have been too black to show detail actually comes to life and does show detail. One example I can give was something I shot years ago when the Getty Center first opened in Los Angeles California.

 

The media was told not to bring lights (because of the framed artwork). When I was outside getting ready to shoot an interview, I left the camera running and began playing around with the menu setting as it related to the black level. Keep in mind the camera I was using has automatic 7.5 set-up built in, but then ON TOP OF THAT, one can play with the low end black level without changing the set-up level.

 

I discovered that by just changing the black default setting from zero to plus one, additional detail instantly appeared on the dark suit of the one the people we were to interview. When I later played back the tape and inspected the waveform, I discovered that the low end of the I.R.E. scale had been lifted a very small amount, but in this instance the contrast reduction it created resulted in more detail being actually recorded on tape. Please note that the bright part of the I.R.E. scale did not rise when the lower end I.R.E. is increased. This means that even though LESS of the actual I.R.E. scale of in play, the combination of the 7.5 automatic I.R.E. Set-up PLUS the ability to then massage the lower end black scale actually produced more detail in the image.

 

This is why I say that no matter how well black level is explained on a technical level, that explanation does not necessarily translate over verbatim to the real world of actual camera acquisition.

[Simon Wyndham]

Lifted blacks are okay if you are going through post grading. But I woudn't use them as final release footage. They can simply look too milky. Remember that although being able to see every detail in a persons jacket is the best picture technically. Aesthetically or artistically it may not be the case.

 

A technically 'correct' picture is never normally the best to look at, nor the one many people prefer.

[Alessandro Machi]

Lifted blacks are okay if you are going through post grading. But I woudn't use them as final release footage. They can simply look too milky. Remember that although being able to see every detail in a persons jacket is the best picture technically. Aesthetically or artistically it may not be the case.

 

A technically 'correct' picture is never normally the best to look at, nor the one many people prefer.

 

Definitely agree with everything you are saying. This is why higher budgeted productions shouldn't feel the pressure of competition from lower budgeted productions that routinely make simply errors when it comes to consistency of image and sound.

 

On lower production budget jobs it would be up to the post house to understand how to maximize the footage they are handed. Sadly, NLE means instead of doing the blue collar aspect of editing such as making sure the audio is actually mixed at the proper levels, eq'd correctly, or ensuring that the picture quality is optimized,

NLE becomes about fast editing and creating multiple layered visual effects.

 

I still have a Betacam SP tape editing system and I can pretty much make the picture look really terrific, even when I'm handed older footage that came from VHS, Hi-8 or even Mini-dv. Video sweetening is definitely a lost art that most digital age people don't even know ever existed.

[Michael Nash]

The issue of video acquistion in real world, real life videotaping situations can be IMPROVED if set-up is correctly integrated at the time of actual camera acquisition.

 

...Until some engineer sees a lifted black level on his waveform, and lowers the setup to be "correct" -- at which point your additional shadow detail is lost again.

 

Please understand that black level, shadow detail, and contrast are all different things. They all affect one another, but it's important to learn how they are different.

 

BLACK LEVEL means the brightness or darkness of "black" in the picture. It has nothing to do with the amount of detail that the camera captures in the shadows (unless crushed, which I'll explain), only how bright or dark that black and detail appears when viewed.

 

SHADOW DETAIL means the brightness (and therefore visibility) of dark areas just above black. It has nothing to do with the black level; only how bright that detail appears when viewed.

 

CONTRAST means the difference in brightness between bright and dark areas of the picture. It has nothing to do with white or black levels, only the distribution of values between those limits.

 

If you make the black levels higher you will see more shadow detail because it too is brighter. But if you try that with say, a night exterior, you just end up with foggy looking blacks. You might see more detail in the deep shadows because it is brighter, but the color "black" is no longer black -- it's dark gray.

 

What I'm saying is that you can improve the appearance of shadow detail by raising the black level in camera, but it's at the expense of the normal contrast range of the image. As soon as you include more dark or black area in the frame, the "lifted" black levels become apparent as foggy black instead of true black. So if it looks good for your shot, then go ahead and do it. Just don't expect that the higher black levels will look right the rest of the time. A better approach is to adjust the BLACK STRETCH, which leaves the black level alone but raises shadow detail.

 

If you lower or "crush" the black level too much in camera, there's a limit to how dark the black will appear. Regardless of how you set up a monitor or how you import the footage, the shadow detail simply gets "clipped" or "crushed" -- in other words lost forever. Using the camera's menu to set up a black value of say, -50, will not make the black level any "blacker" than zero -- it only crushes the shadow detail down to zero.

 

BLACK LEVEL and VIDEO LEVEL act independently, and what appears in between can be described as the CONTRAST. So if you raise the black level but leave video level alone, the overall contrast will appear lower. Lower the black level and the contrast appears greater. Video CLIP (white clip) does not lower contrast however; it simply "clips" or cuts off the picture above the specified luminance.

 

Lowering the contrast of the image by raising the black level really doesn't help you preserve any more detail for later color correction. Often times it's just the opposite; the "compressed" contrast range on tape just becomes noisy and loses fine detail (from contrast) when expanded back out to normal levels.

 

In general, you want your black level and video level in camera to be accurate. They are your reference for the range of brightness that falls in between. In the broadcast world you NEVER want to change the absolute black or white level of the signal, because those points are what engineers (and equipment) look at when passing the signal along from deck to deck to satellite and so on. Of course you can set up your camera any way you want to for creative effect, but then you're designing your own custom workflow. Don't expect other people (editors, broadcasters, and so on) to follow what you had in mind. At that point color bars become the only reference for what's "correct," and even then there's no guarantee there won't be confusion when the picture doesn't appear to conform to the signal shown by the bars.

 

So for the sake of simplicity and accuracy, set up the black level in your camera to a value that you can easily maintain throughout post, whether it's analogue or digital. For an XL-2 cut on a desktop NLE, that's likely to be "0."

[Tom Bays]

I have chosen for caucasian faces to set my Zebra at 90, and then back off the iris until the last of the zebra pattern disappears from the highlights on the face. Is this sensible? Or is there a better method?

 

Set your zebras to 70 and have just a tad of zebra showing on the face.

[Robert Goodrich]

I have understood, to the best of my abilities, the issues that have been discussed. Unfortunately, I am still struggling to apply this valuable information to my concerns.

 

So I'll just come out and say, "This is why I'm asking".

 

I make extra cash doing small promotional videos in the 10-15 thousand dollar range. Right now I do about one or two a year. Just me, my XL-2, and my Mac. I also make short films, and am starting to pull things together to shoot a science fiction direct-to-DVD feature. Same deal: Me, my camera, some lights, very low budget.

 

Much of the movie takes place in dark spaces. I've had one experience with trying to correct for underexposed video, and I don't need to tell you how poorly and noisey the image became. I also want to avoid my blacks being, as described above, milky.

 

So I'm trying to establish the proper settings for both my camera and my professional video monitor. I have set my monitor using some on-line guides, and felt confident that it was set correctly until I read about this 0 IRE thing for DVDs and film outs. Even if I didn't feel confident that I fully understood my camera's settings, I felt that if I had the monitor kosher, I could just experiment with the camera settings until I arrived at what I was looking for. But now I'm not sure I even have the monitor set correctly for my needs.

[Alessandro Machi]

Set your zebras to 70 and have just a tad of zebra showing on the face.

 

While this may be completely correct for scenes that have been properly lit, if a scene is E.N.G. style, I like the higher zebra setting of around 90 I.R.E.

 

Another factor also is whether the shot is a wide shot or a close-up. It may be difficult to tell that there is zebra on an actors face in the wide shot, therefore a lower setting may be adviseable.

 

I don't think I've ever been fooled with zebra's set to 90. I know the moment I see even a tickle of zebra's I should probably back off.

 

A better approach is to adjust the BLACK STRETCH, which leaves the black level alone but raises shadow detail.

 

That is the correct technical term of what I was describing. I was able to keep 7.5 exactly where it was while lifting the lower end black detail...

 

Lowering the contrast of the image by raising the black level really doesn't help you preserve any more detail for later color correction.

 

Lowering the contrast of an E.N.G. image or a lower budgeted production by stretching the low level blacks while keeping set-up at 7.5 does preserve detail for later color correction.

 

Sometimes I think the Japanese did the opposite of "black stretch" in the 90's with their cameras. It's like they did "black extend" in which they expanded the lower end of the acquistion spectrum which created greater contrast when it was not necessary. This was done to create a "film look", yet have contrast range of 125-130 I.R.E. was not a wise move for E.N.G. acquisition.

[Simon Wyndham]

So for the sake of simplicity and accuracy, set up the black level in your camera to a value that you can easily maintain throughout post, whether it's analogue or digital. For an XL-2 cut on a desktop NLE, that's likely to be "0."

 

For the larger ENG cameras a lot depends on what gamma setting is chosen. The DSR450 for example has a number of different gammas to choose from. Most of these do not contain blacks at 0 ire. Instead they are slightly lifted.

 

I always tune the picture in post to get the blacks kissing the 0ire line, and the whitest whites just touching the 100 ire line. I have yet to come across footage from any camera that couldn't do with at least some tweaking.

[Michael Nash]

I always tune the picture in post to get the blacks kissing the 0ire line, and the whitest whites just touching the 100 ire line. I have yet to come across footage from any camera that couldn't do with at least some tweaking.

 

Absolutely. One of my clients has a D-35 (BetaSP) where the overall best picture seems to be black at about 9 IRE, and the black stretch up to +5. Normally I'd try to get the black right at "proper" 7.5, but I've tested it through their whole workflow all the way out to cablecast (their own server) and it holds up best this way.

 

It's unfortunate that ENG video rarely gets the color-correction attention that dramatic material gets. That's where I've learned to be more picky about the image in camera. I can't trust that the footage will get the polishing it really could use.

[Alessandro Machi]

For the larger ENG cameras a lot depends on what gamma setting is chosen. The DSR450 for example has a number of different gammas to choose from. Most of these do not contain blacks at 0 ire. Instead they are slightly lifted.

 

I always tune the picture in post to get the blacks kissing the 0ire line, and the whitest whites just touching the 100 ire line. I have yet to come across footage from any camera that couldn't do with at least some tweaking.

 

Definitely agree about the tweaking aspect. In theory, the more time spent on lighting at the time of the shoot, the less tweaking required.

 

Absolutely. One of my clients has a D-35 (BetaSP) where the overall best picture seems to be black at about 9 IRE, and the black stretch up to +5. Normally I'd try to get the black right at "proper" 7.5, but I've tested it through their whole workflow all the way out to cablecast (their own server) and it holds up best this way.

 

It's unfortunate that ENG video rarely gets the color-correction attention that dramatic material gets. That's where I've learned to be more picky about the image in camera. I can't trust that the footage will get the polishing it really could use.

 

I recently edited some B-roll red carpet footage for an awards show that had been shot on BetaCam SP. At first I was just going to do a straight cut component edit, but I ended up tweaking the video and ended up with a more colorful and dimensional looking picture.

 

Video sweetening is definitely a lost art that most digital age people don't even know ever existed.

[Michael Nash]

Lowering the contrast of an E.N.G. image or a lower budgeted production by stretching the low level blacks while keeping set-up at 7.5 does preserve detail for later color correction.

 

True. One thing to watch out for with black stretch is added noise in the shadows. That shadow "boost" is really just gain added to that part of the signal, and as we all know gain can bring noise.

 

What I usually try to do in high contrast situations (like harsh sunlight) is raise the gamma, raise the black stretch, and lower the master black to -1 (for a little "snap" in the shadows), and set the gain to -3db. Then of course expose more for the highlights, and the shadows tend to fall into place. -3db helps minimize the added shadow noise from black stretch. For simplicity I save these settings to a scene file that I can access at the flip of a switch.

 

Today was one of those days that put me, and the camera (D-35), through our paces. Three separate ENG shoots for the same client, but very different subject matter. This morning was interviews and B-roll with a student on a middle school campus. Fluorescent lighting in the classrooms. I used the D-35's "Flourescent" scene file which saturates reds a little more to compensate for the greenish spectrum. White balance as normal; smooths it right out. For interviews under mixed lighting, I used the feature that lets you dial in the color temperature for WB preset; much more controllable than traditional white balancing (RGB painting is even better at this).

 

Next shoot, additonal material for a documentary about trees and arborists. Solid marine layer, overcast sky. Shot starts with me in a tree-trimmer's bucket 60 feet above a Ginko tree, hot sky at the top of frame, and gliding down in a single shot to the shade under the tree at ground level. For this I used the camera's "dynalatitude" feature which works kind of like an auto-contrast, compressing or expanding knee and black stretch in response to the image. It does a fantastic job of keeping the hot sky from clipping under a normal exposure for the tree (better than knee alone). As we descend, the contrast fluidly comes back to "normal" as the tree and shade fill the frame. 3pm the sun breaks, and I use the raised gamma/black stretch I mentioned earlier. 4pm I try to shoot blooming Jacaranda trees against a blue sky. Video cameras don't like purple and wants to make it blue, so I use the "fluorescent" scene file again to pump up the red saturation.

 

Last setup of the day, I shoot a background plate for a blue screen interview I did last week. Because I remembered the programmed color temp (and other parameters) I'm able to make a reasonable match. Back at the shop I wait for the producer to load up the BG in the Avid and do a quick comp, and it doesn't even look like we'll need any color correction.

[Robert Goodrich]

In an earlier post, Simon recommended an article on Ken Stone's FCP site that offered an experiment for determining the camera's handling of black. I recorded a few second of footage with the lens cap on, and captured it into FCP, then viewed the footage on the waveform monitor. Black was coming in well above 0, but not up to 7.5. Somewhere in the middle. I started playing with the setup level, and eventually got it down to 0 (which worked out to be about -3)

 

I then downloaded 0 IRE color bars from Adam Wilt's site, and adjusted my monitor (which ironically brought my setting back to the dead-middle on my monitor).

 

I stored all these settings into the preset I use for things like interviews and short films; things that are carefully lighted, and delivered on DVD.

 

For determining how I set my monitor out in the field, I intend to switch back and forth between my setting using the 0 IRE color bars out of my NLE, and the bars provided by the camera. This should give me a indication of what my pluge bars should look like away from the office.

[Alessandro Machi]

In an earlier post, Simon recommended an article on Ken Stone's FCP site that offered an experiment for determining the camera's handling of black. I recorded a few second of footage with the lens cap on, and captured it into FCP, then viewed the footage on the waveform monitor. Black was coming in well above 0, but not up to 7.5. Somewhere in the middle. I started playing with the setup level, and eventually got it down to 0 (which worked out to be about -3)

 

I then downloaded 0 IRE color bars from Adam Wilt's site, and adjusted my monitor (which ironically brought my setting back to the dead-middle on my monitor).

 

I stored all these settings into the preset I use for things like interviews and short films; things that are carefully lighted, and delivered on DVD.

 

For determining how I set my monitor out in the field, I intend to switch back and forth between my setting using the 0 IRE color bars out of my NLE, and the bars provided by the camera. This should give me a indication of what my pluge bars should look like away from the office.

 

If there is autogain on when you put the lens cap on then you would need to do the reverse of what you did.

[Robert Goodrich]

I had it on 0 gain.

 

Now, since I have confidence that I'm using the full latitude of the camera, and have the setup correct, I still have to admit that I'm not sure about the monitor.

 

I have been using the color bars that come with Final Cut Pro 5. I have always had it set so that the 3.5 and 7.5 pluge bars are the same apparent value, with the 11.5 still visable. But this setting was having me increase the brightness to nearly +15 on my monitor.

 

I downloaded a set of bars listed as 0 IRE, which had two pluge bars. When I set the two bars to equal value, it brought my monitor's brightness setting to about +2, and things looked great, and very near what I was seeing on the computer monitor (not that means much). But when I put FCP's color bars back through the monitor the 11.5 pluge bar was not visable.

 

So, in short, if I set my monitor to 0 IRE which of the pluge bars (if any) should I be seeing?

 

The reason I'm being so persistent, is that I have read on several occasions that I should set my monitor to 0 IRE if I'm working with DV, but no one has bothered to say how that's different from what I have been doing (11.5 visable, 3.5 and 7.5 just touching).

[Michael Nash]

I had it on 0 gain.

 

Now, since I have confidence that I'm using the full latitude of the camera, and have the setup correct, I still have to admit that I'm not sure about the monitor.

 

I have been using the color bars that come with Final Cut Pro 5. I have always had it set so that the 3.5 and 7.5 pluge bars are the same apparent value, with the 11.5 still visable. But this setting was having me increase the brightness to nearly +15 on my monitor.

 

I downloaded a set of bars listed as 0 IRE, which had two pluge bars. When I set the two bars to equal value, it brought my monitor's brightness setting to about +2, and things looked great, and very near what I was seeing on the computer monitor (not that means much). But when I put FCP's color bars back through the monitor the 11.5 pluge bar was not visable.

 

So, in short, if I set my monitor to 0 IRE which of the pluge bars (if any) should I be seeing?

 

The reason I'm being so persistent, is that I have read on several occasions that I should set my monitor to 0 IRE if I'm working with DV, but no one has bothered to say how that's different from what I have been doing (11.5 visable, 3.5 and 7.5 just touching).

 

What kind of monitor are you using? Consumer televisions don't have the circuitry to keep blacks down in the absence of brighter material hitting the tube, nor the ability to keep blacks from getting crushed when there are bright subjects on screen.

[Alessandro Machi]

What kind of monitor are you using? Consumer televisions don't have the circuitry to keep blacks down in the absence of brighter material hitting the tube, nor the ability to keep blacks from getting crushed when there are bright subjects on screen.

 

Which raises the question, does such a monitor exist that does not also have cross pulse and underscan?

[Robert Goodrich]

Sony professional CRT monitor.

[Alessandro Machi]

Sony professional CRT monitor.

 

Do you have a model number?

[Josh Bass]

I'll confess that this stuff still blows my mind. right now I own an XL1s (getting an XL2), and I've always left those settings alone, 1) 'cause you can't adjust the setup between 7.5 and 0 on an XL1s, and I leave things alone that I don't understand and 2) 'cause it looks better to not lower the setup (crushing blacks on the XL1s tends to orange up the skin tones in an unattractive away). I've always left it at 0. So 0 bad, 7.5 good?

[Robert Goodrich]

Okay. Have received via email what I believe to be the definitive answer from a very reliable source.

 

Set the color bars normally, with the brightest pluge bar just visable. Adjust the setup level on the XL-2 so that they register as 0 value on FCP's waveform monitor (which for me was -4 steps).

 

 

Thanks everyone for educating me on this very confusing, yet important, issue.