Jump to content

The Toe and The Shoulder


Matthew Padraic Barr

Recommended Posts

I was recently testing out the Black Magic 4K camera system and I was having a bit of trouble figuring out the actual latitude. I can easily look up the estimated dynamic range provided by the manufacturer, however I like to be exact in my work, and I've found these numbers put out by the manufacturers to be estimations that can be higher or lower than the actual dynamic range. My question is this: since digital systems handle the highlights better, and I'm used to measuring latitude based on the sensitometry of cellulose-acetate film base, how might I adapt to the digital systems which are omnipresent in the industry now? I know this is a complicated question, and given that we have a plethora of gamma curves that vary depending on the camera systems and the settings within those systems, this may yield a complicated answer -- though if anyone would like to take the time to give me the complicated answer I would love to read it.

I really just need to better understand the gamma curves of digital systems, so if any of you know of some recent books devoted to this which I could use as a reference for all the camera systems and Post workflows, I would appreciate the tip.

Also, I realize this may seem excessive -- measuring the latitude when clearly using a monitor would be easier. But I use a few techniques that will not allow me to use monitors on set because they involve post processing similar to the pull and push processing of film.

-mpb

Link to comment
Share on other sites

  • Premium Member

There is the dynamic range that the sensor can deliver in raw form to the recorder, and generally if that raw signal is converted to a log form, all of that dynamic range should still be available for use. And then there is the dynamic range that can be displayed by a device and look "good" in terms of black and white levels, and often that range is less wide than the raw or log form, and is also much more variable due to gamma curves that are applied to the raw or log signal.

 

But generally a digital camera maker would shoot something like a Stouffer step wedge (preferably backlit) to measure dynamic range:

 

http://stouffer.net/TransPage.htm

http://www.dvxuser.com/V6/showthread.php?187624-GH1-meet-the-Stouffer-4110C-dynamic-range-test

 

But sometimes this sort of testing can give you more technical information than is actually practical.

  • Upvote 1
Link to comment
Share on other sites

  • Premium Member

Thanks!

 

As mentioned in the other thread, using an 11-step DSC chart is a simple way of seeing how the gamma is being manipulated. The fact that the two rows of grey steps go in opposite directions means that they form an X shape of steps on a waveform, so you can see where the mid-point is falling and how linear the steps are above and below middle grey. 11-stops isn't enough to show all the dynamic range of many digits cameras in raw or log form, so you will have to imagine that if the white square of the chart hits 70%, then you'll be capturing a couple of stops above that in overexposure detail. This also means that the wider the dynamic range, the more likely the 11 steps will fall within a smaller area of the waveform. LogC out of an Alexa, which has 14.5 stops of DR, for example, causes an 11-step chart to fall approx. between 15% for black and 65% for white on a waveform.

 

Taking this chart shot in Log-C into color-correction to make a "normal" Rec.709 image for a monitor shows you the issues with having 14-stops on a display that is optimized for 11-stops - the image looks very low in contrast, but if you add contrast to get the blacks on the chart down to 0 IRE and the white square to 100 IRE, you're basically just using 11-stops for the image and throwing away the other three stops of information (I'm probably over-simplifying but you get the point, going from Log to Rec.709 is a bit like printing negative to print stock, the contrast has to go up in order to get an image with deep blacks and clean whites.)

 

If HDR for video distribution ever takes off, this will be less of an issue, we won't have to use tricks like knee compression and luminance keys to force overexposure detail down into the current broadcast display range.

 

On the other hand, it's useful sometimes for color-correction to have more information than has to reach the final displayed image, you have some flexibility to play around or correct misexposures.

  • Upvote 1
Link to comment
Share on other sites

Ok, I'm following; for the most part.

Do you think it's really necessary to concern ourselves so much with with these 14 stop DR's, when it is very likely that only cinematographers can perceive the difference?
What's your philosophy about this?

It would be much easier to work under a standardized 10-stop system, and since Rec.709 displays 11 stops, this would make sense even at the DVD distribution level. We could then work on the DI a bit to get a bigger DR for the DCP for theater distribution.

Also, do you know how much DR a BluRay display system can display? -- I've heard BluRay is YCbCr 4:2:0 native, but I've never heard a DR estimation. If BluRay can display closer to 14 stops than all this extra work I do for those last few stops would be worth it in my opinion, since many cinephiles enjoy BluRay.

Edited by Matthew Padraic Barr
Link to comment
Share on other sites

  • Premium Member

When we worked in film, we shot exposure and contrast ratio tests and then printed the results, so basically we were taking the 14 or 15 stops of DR on the negative but looking at the 11 stops or so that the print would show us, hence why most of us learned that if we let an 18% grey card go 5 stops under, it would go black, and if we overexposed the grey card by five stops, it would go white. So we worked under the assumption that about 11 stops would be visible in a print, and the rest of the range on the negative was there is help us print up or down if necessary.

 

So we should "concern" ourselves with the 14 stops of DR in a good digital camera in the sense that this will give us something closer to the experience of shooting on film and allow us to roll off shadows to black and bright areas to white gracefully rather than abruptly, and we should be concerned if our digital camera only had 11 stops of DR or less. But we work practically within a narrower display range, the extreme ends of exposure are either there to save us in post if we have a problem, or to allow that gracefully transition to pure black or white to occur using things like knee compression, luminance keys, etc.

 

Projects released on blu-rays are color-corrected to the same Rec.709 broadcast standards as DVD's, they are just 1080P instead of 480P (for movies). The issue isn't the blu-rays, that's just a storage medium, it's the monitors. CRT technology was limited to some extent and Rec.709 specs reflect that, but now flatscreen monitors have the capability or potential for a wider color gamut than Rec.709 and wider brightness range than just 10 or 11 stops of information, hence the recent interest in HDR displays.

  • Upvote 1
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...