Skip to main content

To Fix Aspect Ratios, Eliminate Television

You might not have noticed that, in order to solve a problem, it helps to know what it is. Meanwhile, you probably have noticed that gravity exists. Yes, once again, I'm here to rant about aspect ratio.

If you're reading this sorry excuse for fish wrap, chances are pretty good that you're either an engineer or know one. Engineers solve problems.

Ask an engineer how much two and two is, and watch in fascination as a spreadsheet gets built, and various numbers appear on the screen, depending on errors of formulas or formatting. If you see 3.99, consider yourself lucky. It's within "engineering accuracy."

Now ask an engineer to solve a real problem, like people committing suicide by jumping off the Golden Gate Bridge. Education? This is not an area of engineering expertise. Counseling? Likewise. Barriers? Maybe, but potential suicides might climb over them. Nets? Not bad, but there are liability issues. What if the impact with the net causes death or disability? Too dangerous.

Aha! The ideal solution! Tear down the Golden Gate Bridge. Never again will there be another suicide jump off of it. Alternative solution: Kill everyone; no one will ever again commit suicide. Both solutions are presented to management, which will choose, based on preferences, for either the Golden Gate Bridge or life. "But, Mario, isn't that a rather extreme and absurd example?" I'll cop to extreme. Absurd? Have you watched any television lately?


Once upon a time, cinematographers shot film, and all was well—as long as you didn't count the time and effort of processing that photochemical imaging medium. (My last remaining neuron, Nellie, wishes to announce that the last three words of the previous sentence were used to show that the SMPTE Journal has absolutely nothing over TV Technology).

The aspect ratio of the pictures on said film varied starting approximately from Day Two. Maybe Edison picked a screen shape of 4:3, but, by the time the NTSC said it had picked 4:3 to match "motion picture practice," movies had already switched to 11:8 for a decade. And before that, they were narrower than 4:3 because of the soundtrack eating into the frame. And after that, they got wider than 11:8 (the "Academy aperture"), wanting to offer something TV couldn't.

By the time the dust settled around the early 1980s, pretty much the narrowest film aspect ratio was 4:3 (ratios, like digestive systems, have colons), and the widest was CinemaScope (which someone once told me was the first English word with a capital letter in the middle).

Nellie informs me that I could do a whole other column on the shape of CinemaScope (assuming my masters at TV Technology would be willing not to withhold my daily gruel if I did such a thing). But that wasn't the rant I had in mind for this month, so call it "64:27."

'Round about then, folks had been considering modern HDTV for a while ("modern" being thousand-line or more), so SMPTE set up a group to look into high-def electronic production. One of their questions was imager shape.

So they solved what they saw as an engineering problem. Enter spreadsheet formulas. Lo and behold (and I still ain't sure how to lo), the minimum imager-area loss for any shape between 4:3 and 64:27 is 16:9.

Okay, I cheated a little. The engineers on the SMPTE HDEP group used 4:3 for the narrow end alright, but they called CinemaScope 2.35:1, which made the shape for minimum area loss a wee mite less than 16:9. But there's the fun parade of facts if you round upward.

First, (16:9)=(4:3)x(4:3). And (64:27)=(16:9)x(4:3). But, wait! There's more!

A 16:9 aspect ratio allows a big 4:3 image to be displayed next to three stacked smaller 4:3 images on the same screen. The 16:9 aspect ratio is near the geometric and linear means of the then-most-popular global movie-screen shapes, 1.66:1 and 1.85:1. It's close to the shape of shooting 35mm film with three perforations per frame edge instead of four. It was close to the shape of a proposed 1.5x anamorphic-lens shooting system. It was close to an existing standardized cinema screen shape (1.75:1). And there's more!

In adapting Rec. 601 to 16:9, the new sampling rate of 18 MHz wouldn't violate the multiple-of-2.25-MHz criterion. A memory read at 4fsc could create a 16:9 image; at 3fsc it could create a 4:3 image. It provides integer values for multiplexed-analog-component systems. For HDTV, it fits existing memory sizes. It worked in Thomson's Dynamic Pixel Management cameras.

That's all true. It's also a load of hooey. For instance, if the new shape was 5x3, it would allow a big 4x3 picture with four smaller 4x3 images stacked next to it.

The bigger problem is that pictures ain't an engineering problem. Thanks to gravity people move horizontally, which is how come cinematographers shoot film with common sides regardless of picture shape. Folks move across the frame edge simultaneously in any shape. But the 16:9 imager shape is like tearing down the bridge. To get a 4:3 picture, you lop off the sides.

At least during the digital-analog simulcast period, broadcasters could carry full 16:9 on their digital channels and pick letterbox or center cut for the analog. Now that analog's gone (or leaving), there's just the 16:9.

Meantime, most Americans still watch 4:3 screens and get their TV via cable or satellite. Those redistributors don't assign techs to watch each channel and adjust the downconversion instantly based on what's on. They just lop off the sides (lest customers complain of annoying black bars).

Now, then, producers could use AFD to specify how different-shape screens should show every moment of every commercial or program, and that AFD metadata could be recorded and carried right through the whole broadcast chain (just like dialnorm). And cable and satellite ops could use metadata-sensing downconverters.

Gotta stop. Nellie's in hysterics.

Mario Orazio is the pseudonym of a well-known television engineer who wishes to remain anonymous. E-mail him