There was nothing unusual about this technique even in 1958.
The Alvin and the Chipmunks page says:
Another contemporary was The Purple People Eater, released in mid-1958, whose page says:
And if you check the Pinky and Perky page:
The Wikipedia page got it right in the first place and supplies all kinds of evidence for the use of the technique other than the Chipmunks. I wonder how all this information got ignored by the OP.
He did. He wrote a three-part harmony and recorded each of the Chipmunk’s voices separately over the very slow orchestra track.
Obviously the basic technology for producing the Chipmunk voices existed in 1959; I got their first album for Christmas that year. It was all of the fancy high-tech stuff mentioned above that I was wondering about.
While it is possible to alter pitch without speed, or vice versa, with modern equipment, it isn’t lossless like a simple change of tape speed is. That’s why autotuned music never sounds quite right.
I don’t think either one is accurate. They don’t exactly record it at *slower *speed either. What they do is:
I see where they’re going with “recorded at twice normal speed” but it’s misleading.
The song itself has to be played at a slow pace, with every note twice normal length.
Forget what Wikipedia says, right or wrong.
Y’all do know that recordings like the Chipmunks involve multiple recordings, at different speeds and different times, before being combined and transferred to a composite mono or stereo mix, don’t you?
The Chipmunks orchestra (rhythm section and whatever sweetening was used) was recorded at normal speed, whatever that was for the recorder used, probably 15ips. David Seville played that back at half, or close-to-half speed, and recorded each chipmunk voice one at a time, probably on separate tracks, at his normal voice register, singing much slower than usual. When played back at 15ips, his voice sounds like a chipmunk.
He overdubbed the “Seville” voice, i.e., “Alvin!!” and other normal-sounding voices at normal tape speeds sometime during the process.
He then mixed, in real time, at normal speed, the multi tracks, to a mono and/or stereo mix for public release, to an intermediate tape, then typically on 45rpm records or as an album, 33 1/3 rpm records.
Is that clear? Cheesh.
I’m surprised nobody has mentioned Walt Disney’s CINDERELLA, which famously used similar methods for several songs. In fact, it was the first animated film to get an Oscar nomination for Best Sound Recording for that very reason.
Sent from my SCH-I435 using Tapatalk
I’m pretty sure the technique was pioneered by Les Paul in the 1940s: https://en.wikipedia.org/wiki/Les_Paul#Multitrack-recording_innovations
I’m confused. Does that mean that the vocalists have to force themselves to sing excruciatingly slowly so the meter (or whatever word you need to use here) will be whatever it needs to be when sped up?
^Yep. But it makes for great diction when sped up.
Yup. A similar thing happens in reverse for the production of some music videos: the music on the set is played at X times normal speed (where X>1, e.g. playing a 33 RPM record at 45 RPM) while the movements of the video actor(s) are recorded, and then during playback the video and audio both get slowed down so the music is heard at its original tempo, and the movements of the actors have thus been unnaturally slowed down. It lends an otherworldly grace to their motion, e.g. hair or dresses moving in the wind. Since the music is sped up while the video is being recorded, it means that anyone who is supposed to be singing in the video has to move their lips fast to keep up with the music; if the original music is fast-tempoed, then it can be hard to do this well. Here’s a tutorial on the technique, cued to the final result; the performer seems to have pretty bad timing, but if you back the video up to 1:27, you can see how fast she has to move her body/lips to keep up with the sped-up song; it’s quite a challenge.
That was extremely common, if not standard. Same principle as the different tape speeds on VHS VCRs – record at a high tape speed for better quality, low tape speed to get more on a standard-length reel of tape but at reduced quality.
(Also, it occurs to me, pretty much the same as having different bitrates for mp3s back in the Napster days – 96kbps was good enough for most people taking hours to pirate a song on a dialup connection to put on their player that had a capacity of tens of megabytes, but if you were fancy and could afford the wait/faster connection and storage space, you’d encode at 328kbps. And also the lossy encoding required to pirate a DVD onto a video CD to sell to your classmates. There just wasn’t a physical switch on the player for the digital formats.)
Not many know this but in Hitchcock’s Rear Window “Songwriter,” one of Jefferies’ neighbors across the way* is played by Ross Bagdasarian, David Seville’s non-Chipmunk name. He has nineteen other acting credits as well as 63 soundtrack credits (natch).
*Hitch makes his appearance winding his clock in the guy’s apartment.
I knew that! 
I was doing a lot of recording in those days. Almost all of the reel to reel recorders had two speeds, 7.5 inches per second (IPS) and 3.25 IPS. If you were recording voice, or something low fidelity you would use 3.25 to save expensive tape. If you wanted higher quality you used 7.5. There were even machines in the high end with a 15 IPS setting when price was no object. I had a court reporter’s recorder that had a 1.75 IPS so the reporter wouldn’t have to change tape during a long session.
And you do realize that Ross Bagdasarian is David Seville right? Witch Doctor was his own inspiration for The Chipmunks.
Nitpick: Your speeds are a little off. Here are the common tape speeds; each is half of the previous:
30 IPS
15
7.5 (7 1/2)
3.75 (3 3/4)
1.875 (1 7/8)
Note that professional recorders had an adjustable speed feature (VSO) and were not limited to these fixed speeds. I think (but I’m not sure) at least one Ampex had a way to speed up or slow down by musical half-steps so the engineer didn’t have to guess or calibrate.