Highlights vs. Significant Highlights

As a television professional who spends most of his time dealing at the very end of an often lengthy, exhausting, all-encompassing process known as documentary film-making I’m often asked, “What could I have done to make my film look more... filmic?”

The First Law of Filmmaking tends to read along the lines of: Know Thy Camera

Eric Escobar’s excellent blog has a recent post that deals with this issue. He shoots the same image with two different cameras. One camera is the HV20 shooting with a 4:2:0 codec the other is the EX1 shooting at it’s highest quality at 4:2:2 with a lens adaptor.

Clearly the EX1 wins this shootout (if you click through, give the image a few moments to download). The HV20 is downright ugly in comparison.

Now - I don’t care at what frame rate you shoot, the EX1 is far more filmic. Yes? Will 24p make the HV20 feel any more cinematic? No way Josť.

Eric is onto something here... Know Thy Camera.

He mentions that he had trouble with the HV20, fighting all the auto controls of this consumer-oriented camera. Whereas on the EX1 he was able to get the exposure he wanted. This, I think, gets to the crux of the problem. And it’s a problem that I was reminded of recently re-reading the terrific book “Professional Photoshop” (the link is in the sidebar on the right). It’s the issue of Highlights vs Significant Highlights.

Go back to Eric’s post and look at those two shots again. To my eyes the biggest difference (besides depth of field) is exposure. On the auto settings the HV20 sees the brake-lights of the cars and the bright patch of light of the sky and thinks, “Gee, those are the highlights. I must protect for those highlights.” The camera ignores that this is a generally low-key image and acts as if the most important part of the image is the sky. Nothing could be further from the truth.

Eric, using his eye and experience, knows better than the auto-iris and sets up the EX1 much differently. Although he says he’s protected his highlights, he’s actually let his highlights blow out in the blur of a short depth-of-field and selective focus. He’s made the choice that the highlights of the sky and car lights are insignificant and instead chosen the significant highlight in the woman’s face. And he exposed accordingly.

The HV20 has the truly important part of the image, the woman’s face, completely compressed into a narrow range - as seen here in FCP’s waveform. In post, when we dig out that detail we’ll be pulling up noise and degrading the entire image. In that process we’ll let those highlights blow out because... who cares??? We want to see the babe!

And this takes me to a discussion I had recently with a colorist friend of mine who opined that he’s tired of the “protect your highlights” mantra. I tend to agree with him. We’ve both recently seen too many filmmakers walk in our rooms with footage that protects for the sky out the window and buries the truly significant detail - like human faces - into the bottom 30% of the waveform. No, not even Red can completely save you.

As Dan Margulis says in his book, there’s highlights and there’s significant highlights. Based on what I see coming through my doors I say filmmakers need to make sure they protect for the Significant Highlights and let the rest blow out. Especially on a camera like the HV20 where it’s far more damaging to try and dig out an underexposed face than to let a window blow out to white.

- pi

 Subscribe in a reader
|