Digitizing Rare Vinyl
eldavojohn writes "While the RIAA is busy changing its image to a snake eating its own tail, one man is busy digitizing out-of-print 78s. 'There's a whole world of music that you don't hear anymore, and it's on 78 RPM records,' he stated to Wired. Right now, you can find about 4,000 MP3s on his site, with no digital noise reduction implemented yet."
Cue the purists saying: "But it is supposed to have hiss. That's part of its character."
Slashdot (good ol' Slashdot effect), or the RIAA?
:-)
I hope this guy plans on making a torrent with his stuff
...in bed
The Library of Congress has an archival project:
http://www.npr.org/templates/story/story.php?storyId=1216161
This is going the other way - from digital to 78's. Shellac 78's appear to be the best archival format.
I was going to make a hissstorical pun but that's pointless.
Engineering is the art of compromise.
Someone should download the entire site and post it on bit torrent... then email this guy so he can put the bit torrent link on his site.
I feel bad for his poor server.. its about to get quite a few hits since this is now on slashdot.
For all you whippersnappers who don't remember records: not only were there 78 RPM records, and of course the 33 1/3 and 45's you are aware of, but they also used to make 16's (technically 16 2/3 RPM). I used to own one record in that format (long since lost to the grue in the attic). It was just speech, not music; I think they didn't typically use that speed for music because of fidelity limitations of 16 RPM.
I made the mistake of getting rid of my (admittedly modest) vinyl collection in the 80's when CD's were the up and coming thing. Sorta wish I hadn't, now. I'm not one of the people who think vinyl has superior sound, but it did have a certain charm.
In my many years in Radio, I've digitized a considerable amount of music from LP's and 45's. In most cases, I could get moderately scratchy cuts to sound almost new. The transformation is pretty impressive, to say the least! However, I wouldn't even THINK of compressing it to MP3 until AFTER I had run it through an audio clean-up utility, like Cool Edit or Audacity.
I wonder how badly the MP3 compression affects the music with all of that hiss and crackle taking-up so much bandwidth? Also, how much would the compression artifacts affect the ability of the clean-up utility to do its job?
I think it is a laudable thing to preserve some of this priceless music! Kudos!
Willie...
The same music isn't there in CD or MP3. That's the whole point. This stuff is out of print, never been released in CD. It's the in summary for god's sake! "There's a whole world of music that you don't hear anymore, and it's on 78 RPM records".
And before something about noise reduction pops up. Noise reduction takes time. He rather put the mp3s up first. Notice the 'yet'. If you really want a song to be cleaner, clean it up yourself and then send the mp3 back to him.
Then the purists should invent a way to digitally record all of the information. All the 3D characteristics of the record.
I'd go on a Vegan diet but the delivery time from Vega is too long. --brownkitty
A Russian has been up to this since the mid-90s, digitizing old Soviet LPs (1930s on up) and putting them on his site (http://www.sovmusic.ru/english/) for free.
It's a very extensive collection, and is worth a look, regardless of what you think about Russia's past or current behavior.
Most 78's (there are exceptions, including the very famous and historically important V-discs) are not vinyl.
They are shellac, or rather a mixture of shellac, wax, slate, and a cotton or paper filler.
I personally believe that the decline of the music industry is directly related to the replacement of shellac with vinyl, and that the RIAA must remedy this decline immediately.
78s were not made of vinyl. The substance was much closer to wax, FYI.
The Uncoveror: It's the real news.
DOH, I was wrong.
Please mod parent(me) down.
He has WAV versions of the songs, and created the 128kbps mp3s for the website.
He could use FLAC to reduce the amount of storage that takes up, though.
If I have nothing to hide, don't search me
Surprisingly, if you use a piezo, heavy cell (not suitable to read stereo records), you will get a much better sound, and almost no hiss. I got very good results at a time from a Dual 1010 turnable, unfortunately out of order now :-(
I also have some Jack Hylton songs that do not seem to be present on his Internet tribute site (Bogey wail, Sarita...), for whoever is interested. I guess they are legally in the public domain now, as all of them date from before WW2.
Signature omitted in order to save space. Thanks for your understanding.
Well, that isn't exactly what the summary says. The summary says the 78s are out of print, which is no surprise because 78s aren't produced anymore. There's definitely a ton of music on there that is available commercialy in modern formats. For instance, he has "Caravan," by Duke Ellington. That's an extremely famous jazz tune, and I can't imagine there's ever a time when you couldn't buy a commercial recording of it. You can buy it right now on Amazon in mp3 format for 99 cents, or on a CD reissue. I don't know if it's exactly the same performance or not.
The Wired article also has a discussion of the copyright status of these songs, which basically amounts to, "nobody's sued him so far." I guarantee you that the composition of Caravan, for instance, is still in copyright -- Tizol and Ellington wrote it in 1936, so the only way it would have passed into the public domain would have been if the copyright owner had failed to renew it -- but it was a valuable commercial property (still is), and I'm sure they did renew it. (Nothing from after 1922 has expired in the US except by failure to do the renewal that used to be required.) I don't know about the copyright on the sound recording (is the duration different?), but I'd guess it's still also in copyright.
If copyright law in the US was sane, a composition from 1936 would be in the public domain, but that doesn't change the fact that the law is not sane, it is what it is, and these recordings are not all out of print or out of copyright.
Find free books.
in the thread on the tragedy of the anticommons, but it seems even more relevant to this topic
on the subject of intellectual property and the rare souls reviving old media through blood sweat and tears, the filmmaker vincent gallo said this four years ago:
bottom line: revive old media, bring renewed attention AND SALES to a long forgotten artist and piece of music, and expect the corporate intellectual property assholes to punish you for effort
thats the state of intellectual property today
intellectual property law is philosophically incoherent. it is your moral duty to ignore it or sabotage it
He's doin' Yosemite Sam!!
As a suggestion, how about digitizing the songs several times and then using the redundant data to recreate the originals with no hiss or pop.
As I understand it, pop is sometimes caused by buildup and sudden release of static electricity. This means that the pops will be in different places for different digitizations and can therefore be recognized and accounted for. Scratches, on the other hand...
Hiss is stochastic noise and would average out over several recordings.
It should be straightforward to use a correlation coefficient correction to bring all the recordings into "phase", then use a processing algorithm to remove most of the artifacts.
The artifacts that remain can be removed using techniques more suited to single-images; ie - filtering to remove hiss and pop.
You could always store the raw audio in FLAC, and then use digital filtering when you convert to MP3.
1&1 - Cheap domain and web hosting.
It is not the cartridge itself that matters. The shape of the needle changed from the 78 size to a smaller one for the microgroove recordings. (33 + 1/3 and 45 ) The smaller radius on the end of the later needles means that it will be riding on the bottom of the groove instead of on the two sides ( at 45 degrees). Back in the day (fifties and sixties)the cartridge often had both types and could be turned over to select the correct one.
Of course for best fidelity the single use steel needle is preferred....:) I still have a wind up gramophone of maybe twenties or thirties vintage that uses these. No amplification, no electricity.
Sadly you're right - US copyright law is messed up.
From: http://cylinders.library.ucsb.edu/copyright.php
"Sound recordings were not eligible for federal copyright protection until 1972 and recordings made prior to this date are only protected by state and common-law copyright. All Edison cylinders are presumed to be in the public domain as the assets of Edison Records were transferred to the National Park Service, a federal agency. Other American sound recordings made prior 1972 may or may not be protected by state laws or common-law copyright. Foreign cylinders are all public domain in the country of production and are also presumed to be in the public domain in the United States.
The nature of the various state laws and differing interpretations of these laws in state courts means that the legal status of many early recordings is unclear. The passage of the Sonny Bono Copyright Term Extension Act of 1998 reiterated that all recordings made prior to February 15, 1972 are only eligible for protection under state laws until February 15, 2067, when federal law preempts state law and they enter the public domain. While the Sonny Bono law was intended primarily to extend the copyright protection to the soon-to-expire copyrights of multinational corporations and heirs to songwriters, in effect it meant that all early recordings, no matter what their commercial potential, historical importance, or availability as reissues (with the exception of Edison Recordings) may be protected for well over 150 years after their creation. This is in stark contrast to the original copyright law passed in 1790 which granted a 14-year term of copyright (renewable for another 14 years) or the copyright law in effect for other types of publications when these cylinders were recorded which granted a copyright or 28 years, renewable for another 14 year (28 years after 1909). Not a single person who composed a song recorded on these cylinders or sang into the recording horn is alive today, which suggests that the original intent of copyright to "promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries" has been completely usurped by the Sonny Bono law."
This happens to be another incredible collection of old recordings: http://cylinders.library.ucsb.edu/
some of the song lyrics are racist and at least one of them is x-rated and people have to request it.
The early 20th century had a lot of raw, dry, dark, and offensive humor in their songs. People who didn't grow up during those days will find it horribly offensive, esp during the WWII anti-Japanese years or during when segregation was still a law and songs mocked African-Americans.
Just a warning for people who are easily offended, some of these songs might offend them. So do us all a big favor if you are one of them and don't listen to those songs. Monty Python had a similar warning on their show for the same reasons.
Remember, Slashdot does not have a -1 disagree moderation, and no, troll, flamebait, and overrated are not substitutes.
You mean he doesn't have the CDDB plugin for his KLH turntable? Seriously, none of the files have any ID3 tags. He's also using an ACCESS database. I think the archive gods are displeased with this one.
This is quite an impressive bit of work, and kudos to dude for posting the mp3 version of his archived .wavs. Seeing the whole page of awesome music (and the sub pages of Japanese, Arabic, and Greek stuff as well) really makes me want to see this all packaged up as a torrent - and sooner than later. Spidey sense says many of these will be drawing unwanted interest.
I was expecting someone putting a record into a flatbed scanner
That's been tried, and it sort of works. But ordinary scanners don't have enough resolution. The Library of Congress has a scanner that does. They image the disc at a resolution of 1 micro per pixel, which yields 8 GB or so of imagery. Then they have software which can reconstruct the audio from the image.
Not only is this useful for fragile, unique records, but it will work on cracked or scratched ones. It's even possible to reconstruct a broken record if you have all the pieces.
The current scanner only works for horizontal recording; it can't read depth. So it won't work on vertically recorded records (Edison) or stereo (45/45 Westrex has two components 90 degrees apart.) They're working on that.
Silly question, regarding digitizing 78s. If one can get the right stylus, can't one take a 33 1/3 TT and sample at a 2.34:1 ratio so the net result is like 44.1/48/96 what have you. 78s are likely pre RIAA filters and as such base response shouldn't be that much of an issue.
There is no sanctuary. There is no sanctuary. SHUT UP! There is no shut up. There is no shut up.
If that were my project and I was putting that much work into the data creation I would want a lot more reliable hardware and backups. I'd also work to do more automation.
But, awesome work, and thanks for sharing (:
No sig for you. YOU GET NO SIG!
There are also an amazing number of people who are transferring old Quad albums and tapes from the 70s. They digitize them and then re-release them on Bittorrent as DTS encoded .wav files which can playback with any CD player and any standard 5.1 surround sound system.
I personally possess nearly a terabyte of such albums, and I've hardly scratched the surface of what's out there. It's amazing to listen these old quad albums because most of them were professionally mixed and they enable the listener to appreciate the music more than any stereo recording can, often you get entirely different takes than the stereo release.
Check out http://groups.google.com/group/SurroundSound/ or Demonoid torrent site.
If you really want a song to be cleaner, clean it up yourself and then send the mp3 back to him.
Cleaning an MP3 is rather dubious since the final result will have been encoded, decoded, and re-encoded to a lossy format. The low frequency range of 78's makes it feasible, yet difficult to palate. Trust me, been there, done that.
If you're aiming for a noise floor of a relatively modern recording, even from the 70's, you're looking at about 18+dB reduction. Removing large amounts of hiss is best done in layers with 6-7dB reduction each, so we're talking at least three passes through a good multiband noise gate, each layer leaving artifacts of its own.
It's actually very interesting doing the processes together in realtime. At first it didn't make sense to me that they even made realtime multiband NR, but the best settings for each layer vary depending on the dynamics of the content. The first layer deals with just the louder segments, so you use different settings if they tend to be a vocalist or a drum, for example. The second and third deal with lower level sounds and don't vary quite so much, but the amounts of noise each layer will reduce is a matter of trial-and-error.
In the end, you leave just enough hiss behind to mask the artifacts. Any artifacts present in the source file have to be masked too, so they greatly affect the amount of hiss that can be removed. He definitely should be archiving to a lossless format if he ever expects anyone to work on them at a later date.
War as we knew it was obsolete
Nothing could beat complete denial
- Emily Haines
Dad had crates of them and he picked up cheap players at fetes that you could wind up.
Sadly us young boys wrecked a number of the records and I ruined one of the players with my half-arsed engineering skills. I tried to slow the player down enough to play at 45rpms. The styluses were brass? or silver and would destroy the newer vinyl anyway. We grew up playing the Andrews Sisters, Glen Miller Band and Mario Lanza.
When I was twelve, I visited a friend who played his "Fireball" Album and I left the 78s behind.
Slashdot: Where nerds gather to pool their ignorance
Been there, done that.
What you need to get as a basic setup, is a modern 33.3/45/78 turntable with a ceramic cartridge (or as modern as you can get). Those late 70s and early 80s turntable with strobe speed control is excellent because accurate speed is important.
Try not to use a magnetic cartridge because you will probably need to amplify it. If you amplify it, or for that matter, click the LP/Record option on most audio rippers, you will be applying an eq curve called an 'RIAA EQ Curve'.
What this does is to alter the sound as it is being ripped to disk. This curve is used to help get the tonal balance of records, but was only introduced in the 1930s, so any pre-WWII recordings probably don't have it as the RIAA curve was used in the process of cutting the disk. You'll find that those early records were made 'direct-to-disk' and pressed as such. Having a cheaper ceramic cartridge connected direct avoids this easily. Ceramic cartridges also have a higher output (more volume) and is better suited to sound cards in this case.
So try not to use an amp (or if you have to, then get one where you can switch the RIAA curve out), and plug the T/table into the soundcard. There's lots of free audio ripper software out there and you should get it digitized with no probs.
Don't forget to clean each side - lukewarm water with a little natural soap, 1" paint brush to apply - get the brush bristles into the grooves. Rinse. Don't dry it with anything, but shake it dry. Don't get the label wet. Water on the grooves is ok and some actually flood the grooves when they record as it dampens the needle. :)
The tone arm weight has to be heavy, about 5 grams if you can manage it - or put a small coin on top of the headshell. Experiment with a non-critical record and make sure that the needle is free to move and not jammed up into the cartridge.
Now when you've done all of that, put up a website and let me know the URL
Don't be apathetic. Procrastinate!
It is not the cartridge itself that matters. The shape of the needle changed from the 78 size to a smaller one for the microgroove recordings. (33 + 1/3 and 45 )
Yes. These cells commonly used a commutable needle : one for 78 rpm and another for the microgroove, and a level allowed to switch from one to the other. Needless to say, I supposed the right needle was used.
That being said, piezo cartridges and magnetic ones accepted at a time these dual needles, so using the right needle is necessary, but here not sufficient :-)
Of course for best fidelity the single use steel needle is preferred....:)
That might be. When I was very young we used to have a "Peter Pan" portable mechanical 78rpm player and we had a box of needles, which had to be changed rather frequently. I had the surprise, when reading its user's manual to see that the manufacturer recommendend changing the needle after each record, which seems unbelievable. I always wondered if that really applied to steel needles, or just to former bamboo needles, which I never had a chance to see.
I still have a wind up gramophone of maybe twenties or thirties vintage that uses these. No amplification, no electricity.
What makes me sade evert time there is a technology change is the know-how that it lost with it forever - except perhaps a for a few passionates which allow it some survival. In french brocantes, it is common to find objects for sale, the function of which is ununderstandable, even for its preceding owners :-/
Signature omitted in order to save space. Thanks for your understanding.
I've been digitizing my parents records since I started high school in 1999. I don't spend every waking hour on this, but I've found a few gem. Plus, when I listen to music that is hot now, many of the same "beats/instrumentals" have been taken from these old songs. There is a lost history of these old records out there, you can find the same albums on amazon but they've been re-mastered and put together differently. this guy is awesome.. and I will continue my conversion.
Now that we've established that, when will you be converting all seven thousand-plus files from his site, building a front end, populating it, and giving us access to your obviously far superior solution?
It's early in the week. You'll have it ready by Monday or so, right?
It's all about the information. And what we do with it.
FYI, in the US, it is only material published before 1923 that is guaranteed to be public domain.
-molo
Using your sig line to advertise for friends is lame.
I'm getting nothing but errors when I try to pull any of these up. They were working earlier, but it looks like Yahoo has pulled them.