So Everyone Is Wrong...Now What?

You might not have noticed that not everything that everyone agrees on is correct.
Author:
Publish date:

You might not have noticed that not everything that everyone agrees on is correct. I ain't even talking here about trick questions like who the first president of the United States was (George Washington didn't hold the office until 1789) or what a gentleman does standing up, a lady does sitting down, and a dog does on three legs (shake hands, of course).

No, this time what Nellie the Neuron tells me I had in mind is the most basic characteristic of television that there is. But first I guess I ought to announce some of the news.

At Cine Gear Expo last month, an event that many of you probably never heard of, Panavision introduced a new product, Genesis. It's an electronic cinematography camera; not the company's first. Methinks their Panacam, with electronics by CEI (if Nellie is still filing things correctly), came out in the 1970s, around the same time as Ikegami's EC-35.

That's EC, as in electronic cinematography and 35 as in 35mm film. Back in those pre-HDTV, pre-progressive-scanning, pre-24-frame-per-second days, methinks CBS, Ikegami, and Panavision were already claiming stuff that made pix that looked "just like film."

Then along comes HDTV, and my gosh if folks don't say it looks "just like film," so it gets used in a "Star Wars" movie. And then there's an even better version of the recorder that folks say really makes it look "just like film," so it gets used in the next "Star Wars" movie. So now there's the Genesis, and it really most sincerely makes pix that look "just like film."

IN THE BEGINNING

Genesis means beginning, which is fitting, on account of Panavision starting over on this whole electronic-cinematography business. A synonym is Origin, which happens to be the name used by Dalsa for its electronic-cinematography camera. I ain't sure if the Kinetta electronic-cinematography camera has a name yet. It could be the Inception. ARRI could call its the Outset. Lockheed Martin's could be the Exordium (they said they didn't really want to be in that business anyhow). I leave it to Grass Valley, Panasonic, and Sony to use their own thesauri.

Anyhow, between the EC-35 and the Panacam 30 years ago and the current crop of "just like film" cameras, SMPTE tried to wring order from chaos. They created a working group on high-definition electronic production, and the group first tackled that most basic of issues, where to meet. After that, they tackled the second most basic of issues, the shape of the pictures.

This much they knew: Cinemat-ographers sometimes shoot for a 4:3 TV screen. They sometimes shoot for a roughly 2.4:1 CinemaScope-shaped theatrical screen (at the time it was called 2.35:1). They sometimes shoot for shapes in between. Only in extraordinary circumstances do they shoot for any wider or narrower shapes.

The E of SMPTE stands for engineers. So the engineers tackled an engineering problem. What is the shape that will have minimum area loss for any shape between 4:3 and 2.35:1? The answer is roughly 1.77:1, pretty close to 16:9.

Like it so far? There's more. 16:9 is close to the linear (1.76:1) and geometric (1.75:1) means of the world's most popular theatrical projection aspect ratios (1.66:1 and 1.85:1). It's close to a Motion Picture Association-proposed 1.5x anamorphic projection system (1.76:1). It nicely matches a three-perforation length frame of 35mm film. It's close to an old 1.75:1 projection standard. It allows a large 4:3 image to be displayed with three smaller ones stacked next to it. It works with digital component sampling rates and color sub-sampling. It allows square-sampled HDTV with 1920 active pixels per line to fit in common memory sizes. It allows dual-aspect-ratio memory readouts at 4fsc and 3fsc sampling rates. It's 4:3 x 4/3 (and another 4/3 factor brings it up to CinemaScope). Methinks I could go on, but this paragraph is already too long.

NOT SO FAST

With so much going for 16:9, the engineers of SMPTE's WG-HDEP voted unanimously in favor of it the very first time it was brought up. They were probably very pleased that they had solved their little engineering problem. And, with SMPTE's approval, 16:9 spread worldwide.

There's not much that everyone agrees on in television. Frame rate? Hah! The infamous Table 3 of ATSC standard A/53 lists six of them. Number of active lines in HDTV? There are at least 1080 and 720 (with some folks also buying 810, 576, 540, and even 480 under certain circumstances). Colorimetry? Surely you jest. Rec. 709? Rec. 601?

The one and only thing that everyone in the world agrees on for HDTV is that it has a 16:9 aspect ratio. After all, SMPTE's WG-HDEP solved that little engineering challenge back in 1984.

There's just one little problem with the little solution to that little engineering challenge. It wasn't an engineering challenge in the first place.

Yes, 16:9 is close to the ideal shape for minimum area loss for all aspect ratios between 4:3 and 2.35:1. So? Who cares about minimum area loss?

If cinematographers were concerned about saving film stock, they could have switched to three-perforation 35mm film frames long ago. Instead, if they haven't been using anamorphic lenses on 4:3 frames, they've been shooting on roughly 4:3 film frames and using horizontal viewfinder lines to determine what'll be seen in various display aspect ratios. That way, each display has the same left and right edges, and characters enter and leave the frame at the same time regardless of aspect ratio. And, anyhow, even the SMPTE WG-HDEP engineers intended the 16:9 shape only for shooting, not for display.

Kerns Powers, who introduced 16:9 to the SMPTE working group, acknowledged the error in an NAB presentation 10 years later. "Had the SMPTE working group been aware in 1984 of the full-frame (soft mat) protection scheme, it is by no means obvious that 16x9 would have amassed the advantages over 4x3 that persuaded the working group to make that choice." Oops.

"But, Mario, why are you bringing this up another ten years later?"

Heck, that's easy. I'm just reporting the news.

Remember that Panavision Genesis electronic-cinematography camera? Guess what its aspect ratio is. Ready? It's 4:3. ARRI's D20? 4:3. Lockheed Martin's prototype with imagers the size of IMAX film frames? 4:3. In this era of universally 16:9 HDTV, what are arguably the most advanced cameras for high-definition electronic production all have 4:3 aspect ratios.

"But, Mario, are you saying 16:9 is a bad idea?"

Far be it from me to disparage something that everyone in the world has adopted-everyone, that is, except ARRI, Lockheed Martin, Panavision (Dalsa's not 16:9 either), most TV shows, commercials, and movies being shot today, and most TV sets being sold today. But far be it for me to prevent you from drawing your own conclusions.

Too bad we're not living in that early Lumiere brothers movie "The Arrival of the Train at the Station." ARRI, Lockheed Martin, and Panavision notwithstanding, the HDTV aspect ratio train left the station 20 years ago. Like it or not, 16:9 is the shape of things to come, even if we have to put up with fat faces for the next half-century or so.