If this is going to be a constant-angular-velocity (CAV) record with the same track density as an LP, I disagree that the record’s area is the relevant measure. It should be about radius.
An LP has a radius of about 15 cm but the label doesn’t need to scale, so it’s only the sound part (about 10 cm or 1e-1 m) that represents 60 minutes (6e1 minutes). Multiply that by 5e15 as in Chronos’ post, and you get 5e14 meters in radius, or 1e15 meters in diameter. (Plus about 10 cm for the label.) This is about 6.25e11 statute miles. The outside rim would be about 3.14e15 meters in length; if the whole thing needs to spin at 33.33 revolutions per minute, we’ve got a relativity problem. Even getting the needle to stay in the groove at those speeds is going to be a challenge.
We really should be looking into making a constant-linear-velocity (CLV) disk instead. Then the area-based calculations would be appropriate. It complicates seeking a bit, though.
A 12-inch 33-RPM disc is sub-optimal. Start with a 16-inch 16 2/3 RPM turntable and extrapolate from there. Those used to distribute 1/2 hour radio shows per side. And since we’re only dealing with speech, go to 3 RPM and piezo cartridges. Who needs fidelity?
That was my first thought - it doesn’t have to 33 rpm, it could be 16 or something even slower. I remember having a record player when I was a kid that had a 16 rpm setting, and my parents explained that they used to use that for spoken word records. (Us kids just used it to make our 33s and 45s sound funny.)
I’m not going to follow the math here, just assuming everyone is correct in their calculations, but the first question is easy, right? 1 groove per side.
You’d think so, but occasionally novelty records had a ‘hidden’ groove that had extra material, including one of the Monty Python records I think. There were also horse-race game records, which had ?half a dozen grooves, so a group of people could bet on the outcome and then drop the needle, which would catch a random groove and play out a particular race.
Not sure how that could be applied to the scenario, but you could maybe bet on alternate historical linguistic scenarios like - the last 2500 years of language if the Romans had lost the Punic Wars; if William hadn’t conquered; if the first Indo-Europeans had turned left at Albuquerque.
I particularly like the issues of the tracks at the speed of light (with apologies to Smokey Robinson). First off, on the purpose of the disk, to be heard I hope even by humans–playback amplifiers will need to calculate the relativistic correction for any number of things, not the least of which is how the damn thing can function in an acoustic environment of any sort.
We’re not talking about playback by transmitting EM or other waves through a vacuum. Let’s get real here.
Similarly, I can see that our imprint-vinyl recording being sub optimal, although taking the Earth-atmosphere acoustic frequencies and using super-duper transportation of some sort to lay down the grooves on the platter is conceivable.
MAD Magazine included a flimsy plastic record like this in one of their monthly offerings. It was a song called “Super Spectacular Day” that started out all nice and optimistic, but then quickly turned dark. I want to say there were at least 5 tracks on that record.
ETA: Found it. (Gawd bless the internet). 8 versions.
Ah, thank you! I had seen TV ads for such a horse-race game back in the 1980s, and always wondered how they could claim it was random if the user could put the needle down in the same spot every time.
An aside, but, last night Mrs. Napier was going on about how wonderful audiobooks are, as if every book should be an audiobook. And so of course I started thinking about audiobook dictionaries…
And audio phone directories, tech manuals, codebooks, atlases, cookbooks, the Voynich Manuscript, puzzle books, field guides, encyclopediae, etc. Blank pages are quiet.
He doesn’t ask for ALL speech of EVERY human but merely for words in human speech. Let’s limit this to Homo Sapiens Sapiens - Denisovans and Neandertals need not apply.
The English language contains possibly upwards of 1/2 million words. (cite) but it’s pretty rich. Let’s say that 100,000 human languages (noted upthread) contain an average of 50,000 words each for a total of 5 billion words. (We’ll gloss-over polysynthetic languages that assemble simple syllables into complex “words”.) Let’s say that a word can on average be spoken in one second. Duration is thus a bit under 1.4 million hours, requiring the surface area of that many single-sided 12-inch 1-hour ELP discs, or about 444,000 square feet, rounded. That gives a diameter of about 750 feet, if my groggy, un-caffeinated brain didn’t muck the numbers. I welcome correction but I think that’s in the ballpark. Or Ikea parking lot.
The OP asked for ‘every word ever spoken by a human’, and you could interpret that both ways [Every Word Once, or All Words by All Humans]. However, some of us spelled out the quantity implied by the All words-All humans version, which he seemed to be very comfortable with.
Curwin might help here by chiming in and clarifying this before we get too much more wacky mathematical minimalism.
