Just a short note to make your day a little more awesome. The Library of Congress has made an extensive catalog of recordings available to the public free of charge.
Oh, very cool. I’ve never actually heard Caruso’s voice before. WOW!
Exploring that is going to chew up some time. Time well spent, I might add.
I just listened to [“Angelo casto e bel”, and it starts out a quarter step high and gets higher and higher by the end.
Surely this is the result of recording equipment. But why did this happen? It it common?
Very cool.
I’ve checked it out, and they need to tweak a few things. The Embed option doesn’t work on Firefox 4.0.1. I would love to see other embed options so we can put these clips on Wikipedia.
Dunno. I’m fairly ignorant of music at levels surpassing high school choir. But I got goosebumps all the same.
Two things: recording speeds were not standardized until well into the 78 era (and even then the standards weren’t rigidly followed), and early recording equipment was primitive enough that the speed could change over the course of a recording.
Many early Victors were recorded at about 76.59 RPM, which is about a third of a semitone slower than the “standard” of 78.26 RPM. Columbias and OKehs were often recorded at about 80 RPM, which is about a third of a semitone fast. The nominal speed of 78.26 was only officially agreed to by the major recording companies in 1925, and even after that they continued to make records at other speeds. A lot of Deccas from the mid-thirties were recorded at speeds higher than 80 RPM.
Early records were cut directly onto wax using special turntables. These turntables didn’t have electric motors - they were driven by falling weights, and their speeds were governed mechanically. The speed was usually pretty stable, but sometimes things would go wrong and the recording would either speed up or slow down.
A professional remasterer should correct for these speed problems. The overall playback speed can be adjusted either digitally or at the turntable. A good digital audio editor will also have features to fix a recording that changes speed over time. My guess is that this Library of Congress jukebox has only straight transfers that were done at 78.26 RPM.
Another thing: the standard of concert A being 440 Hz was not settled on until surprisingly late. 435 Hz was common in the early 20th century, and in general European orchestras tended to tune flatter than American orchestras. I doubt that this is why the Caruso record is off-pitch: it was recorded in New York, presumably with an American orchestra, so they probably tuned at A=440. And if they tuned at some other pitch, it would likely have been flatter than 440, not sharper as you heard on the LOC web site.
Thank you for your very informative answer. I knew about that last part, but not the rest. My ignorance has been fought. I’m also glad that a professional would alter it, as that’s what I was thinking about doing.
And, no, I don’t have perfect pitch. I was using my keyboard to figure out what the top note was, and then tuned it to match, and then it was back out of tune by the end. I only know it’s sharp (and not flat) because I was able to find other people singing it on YouTube.
well that sucked up some time =)
And I will be going back =)
They had a pretty interesting story about this on NPR this afternoon, and mentioned that Sony (the heir to the Victor company) owned the rights to the recordings. But some of the recordings date back to 1900, so wouldn’t a few have fallen into the public domain? Just how long do the copyrights last on recorded music? The NPR story did touch on this, but they didn’t go into much detail.
Copyright law in the U.S. didn’t apply to recordings until 1973. Before then, the rights to recordings were protected by state anti-piracy laws. These laws prohibited unauthorized copying of recordings and didn’t specify a time period. That is, the rights to copy these old recordings never expire.
The right to broadcast an old recording is covered by federal copyright law. Broadcasting a record is considered a public performance of the work, and licensing agencies like ASCAP and BMI collect fees for these performances. I would expect a record from 1900 to be in the public domain for broadcasting, but not for duplication.