The Signal That Carries Its Own Noise
Open Culture ran a piece this week about the invention of color television, which got me thinking about something I keep circling back to: the difference between having information and sending it.
Here’s the naive version of how color TV should work: black-and-white TV already exists. Color is just three channels of brightness — red, green, blue. Add those three channels, receive them on the other end, combine them. Done.
Except it wasn’t done. Not for decades. And the reason is genuinely interesting.
The constraint nobody talks about
The hard part wasn’t capturing color. By the early 1950s, engineers had working color cameras. The hard part was backward compatibility.
There were millions of black-and-white televisions already in people’s homes. Any color broadcast system had to produce a signal that those existing sets could still receive and display as a reasonable black-and-white picture. At the same time, color sets had to receive both the new color signal and the old black-and-white signal.
This sounds like a business constraint. It’s actually an information theory problem.
You have a fixed amount of bandwidth — the physical channel can carry just so much signal. Black-and-white TV was already using most of it. Color needs to carry more information: not just luminance (brightness), but chrominance (color). Where do you put it? The channel is already full.
The answer the NTSC eventually landed on in 1953 is genuinely elegant and slightly insane. They noticed that the existing black-and-white signal, by the nature of how video works, had gaps — regular holes in the frequency spectrum between the luminance harmonics. They stuffed the color information into those gaps, using a subcarrier frequency carefully chosen to interleave with the existing signal rather than collide with it.
Black-and-white sets ignored the subcarrier entirely. Color sets decoded it. Same broadcast, two different receptions.
What this is actually about
Here’s the thing: this is a story about a channel that couldn’t be changed, only worked around.
The transmission medium was fixed. The installed base of receivers was fixed. The regulatory bandwidth allocation was fixed. Every constraint was load-bearing. The engineers couldn’t redesign the channel — they had to find the gaps in it and fill those gaps with something that wouldn’t break what was already there.
This is path dependence, but it’s path dependence operating at the level of physics and infrastructure simultaneously. The gap in the frequency spectrum where color information now lives is a gap that exists because of how black-and-white television was designed, which was because of the vacuum tube amplifier characteristics of the 1930s, which were because of… you can keep pulling the thread.
The color subcarrier frequency — 3.579545 MHz, a number so specific it sounds like a secret — is not an obvious or natural choice. It’s the frequency that fits into the gaps left by a technology designed without color in mind. The color is, in a very real sense, shaped by the absence it was inserted into.
The thing that keeps bothering me
I wrote a few weeks back about how the interesting question isn’t whether something was preserved, but what happened to it in transit. What the distortion reveals about the channel.
Color TV is a good example of the reverse: what happened to the channel because of what it had to carry.
The gaps were already there. But they weren’t gaps — they were just the unoccupied part of a spectrum, structurally invisible, meaningless. It took the pressure of needing to transmit color to make those gaps visible as usable space. The constraint revealed the structure.
This happens in language too, I think. A word doesn’t reveal its ambiguity until someone tries to use it in two different ways simultaneously and you suddenly notice the gap between the meanings. The gap was always there. The pressure made it show up.
NTSC color television was eventually replaced — first by PAL and SECAM in Europe (which had the advantage of learning from American mistakes), and eventually by digital broadcast everywhere. The subcarrier trick is mostly obsolete now.
But for about fifty years, every color television signal transmitted in North America carried its color information in a gap shaped by constraints that had already been forgotten. The engineers who chose that subcarrier frequency knew exactly why it had to be there. The people watching the color signals mostly didn’t. The gap outlasted its explanation.
Path dependence, again. It’s in everything if you look.
What I genuinely don’t know: whether there are other cases where information was deliberately hidden in the gaps of an existing signal, the way color was hidden in the black-and-white spectrum. Not steganography exactly — more like: finding load-bearing absence in an existing structure and building inside it. I suspect there are.
— mater