Does anyone know if a music database of songs with detailed information exists, either free or paid?
I'm looking to access information like
The major stations will have bought in complete systems from somewhere, there used to be CD libraries you could buy like that when I started with this stuff. Only available to licensed stations and all that, and not cheap from what I was told. Where I am there are several networked stations who are all with the same parent companies, they just swap between themselves and cut down on the overall work.
Smaller stations, like the ones I've worked on, have had to do their own. As I said, I've put my tracks into a few now but without charge. All were/are PRS and PPL licensed and it seems that this sort of thing is allowed, again because they are licensed.
It's the same sort of thing as buying in CD's or even records in the past. Each station had it's own library. OK, so now we've gone digital it's easier to consider having a buy-in library, but I'm not aware of any.
Not that I am offering, the 'copyright police' would be on my back in no time, but I wonder what a station would pay to have such a library made available? I have the tracks already edited and tagged, it's my personal collection really. I was a computer analyst/programmer with a lot of database experience and could create such a beast. It would need a lot of work to fill in the info you want though.
Obviously I have artiste and title for every track. The duration is available but would need typing in, as would the album and other info needed. I have never needed/wanted to use intro length, a lot of the output from these tracks was for automation/autoDJ applications. On a side note, PlayIt does a good job of finding the start and end of tracks, especially the full radio shows I'm working with here, but every track in my library was already edited to start at the start and had a 3 second out - be that a sudden end or a fade, I allowed three seconds for end of track so crossover to the next track would be seamless.
On the topic of differing length of tracks, yes I have lots of duplicated tracks - as I mentioned the extended/12 inch versions would be a longer version of a track that may already be in the system (but may not be...) and I covered that by adding the time to the title like SONG - TITLE (6.21) where the time given is the playing time, the track itself would be 6.24 long - a three second outro. That makes it easier to find them in the database, although they are in a directory of their own anyway.
I have a system of segregation for tracks. I have a set of directories that define the decade and so on. 3 digits. The first digit defines the decade so is 5,6,7,8,9,0,1 etc The second digit is 1 for a top 10 hit, 2 for positions 11-40 and 3 for outside the top 40, including album tracks. The third digit is my own rating, 1 for the very best, 2 for good, 3 for deserves airplay, 4 and 5 are for tracks I wouldn't want to hear on air in automation but have been included if required for manual play. This includes football songs, alternative mixes and such.
On top of those I have further directories for REGgae, MOTown, ROCk, NSOul, TWElveinch, INStrumental, COUntry and more. That makes it easier to schedule specialist music if needed, or makes it easier to find for a live presenter. This was created more to aid new volunteer presenters on a hospital radio station. If they were asked to 'play something Country' they could find tracks in that category. They were also told that anything on the system was safe to play as it had all been checked for profanities and so on.
It isn't down to not wanting to share, it's down to copyright. Anyone caught selling such a database would need to be covered legally. I'm not :-)
Some of what you wants with Music Brainz ( https://musicbrainz.org/ )
That has CD metadata and some CD rippers access it (or Gracenote) for information.
The last 3 on your list tend not to be available. The intro length is at least a fairly unambiguous but is stored in different ways on different systems.
The fade treatment is a complex subject and interacts with rotation tools and the better ways of dealing with mixing so is probably best held by the playout system rather than within the file.
Sadly, in practice, you will need to invest the time in your library and your system.
Obviously, making this information tricky to import and export helps to keep users locked into a certain suppliers product range so there isn't much pressure on standards compliance.
Thanks, Mark. I appreciate your post, and yes, it is sad that I have to redo what I'm sure many have already done before me. So be it :)
There is (obviously) a demand for at least the vocal start point for a track and a playout system could provide a way of crowd sourcing that data.
If Jason chose to do it, PlayIt Live could collect those manually set vocal and fade point tags and offer them for the same files in other systems. How to determine the "same" file is more tricky as my 256kb/s rip of a CD would be different to your 128kb/s rip of the same song but perhaps a MP3 tag matches and duration would be sufficient.
It would be reasonable to contribute the tags I create in return for being able to review and accept tags by others.
I have considered this but it is non-trivial and fairly subjective. I would need to implement some kind of Shazam-style detection of points to find the exact point in the song that someone chose. Taking into account radio edits, album edits, remixes and various compression/normalisation of the same song, there would likely be a very high number of incorrect values. I'd rather the system didn't provide the feature at all than provide one that was of a low standard.
Online music databases exist and many of them are free. Gracenote started out as CDDB (CD DataBase) and was free but now is decidedly not so. freedb (all lowercase) replaced it and is now the preferred, at least last time I checked which has been a few years.
freedb is just the database that many programs reference to fill in the data fields. I used to use it while ripping CD's, which is a topic for another day, but there are also media players, cataloguers, audio taggers that can do it. My experience was great with only very rare tracks not already in the databases.
Music for radio is certainly available. TM Century is still available and there used to be another similar CD service whose name I can't remember. I've seen ads in the past that offered libraries on a hard disk drive but can't remember much about that either.
I entirely agree. The only reasonable way of identification would be duration and metadata but it wouldn't be reliable enough to be automatic.
At the risk of drifting off topic, the one thing I would suggest needs to be added to the PlayIt Live ingest process is a subjective loudness measurement. That gain adjustment (combined with any compression/processing on the player) could be applied to the track on replay such that it both "sounds" the correct level and doesn't distort/clip.
The words "You'll be lucky" spring to mind :-(
Those stations/people who have such databases will not want to pass them on, paid or free. I've been doing this for about 20 years now, at several stations I've used my database and tracks - about 18,500 songs that I have personally ripped, edited and tagged for airplay. Over 65Gb of music at 128K as that's what I wanted it at (before anyone starts saying it should be WAV). It worked for the AM station I originally created it for, it's worked on a few FM stations I've been involved with and now I'm using it with Internet radio, hence looking at PlayIt currently.
Happy to report that it all works fine, loads up without fuss and songs start at the start and end at the end with correct ID's displayed. As I say, 20 years in the making, mostly on a voluntary basis. Because most of these music databases are specific to a station or person they tend to be personal choice to a great degree. Mine is mainly top 40 hits over the years but over 1,000 tracks are extended/12 inch versions. There are sections for Northern Soul, Motown, Rock, Reggae and more.
Someone did try to rip the tracks off a station system I had set up, but were caught by management and dismissed. All tracks I've edited and tagged are marked electronically so they can be identified if they were to be stolen again.
Yes, going back to my original comments, "you'll be lucky" - even if someone was prepared to sell their database. I am still adding tracks to the system here but it's not as important that I get every track that ever made the charts loaded in as I don't need that now. I've (more than) enough to keep things going for many years.
Once you get into it you will find that the work flows quite well, ripping, editing, tagging - I was doing 50 - 100 a day, when I knew I had to get quite a lot together to be able to run a station on the proceeds.
Best of luck in getting what you want.
Happy to discuss this if you create a new topic :)
I understand where you're coming from but am still hoping something exists somewhere. If it doesn't, wouldn't that mean that every station across the globe, whether tradition or Internet radio, would each be doing this same work? That seems a terrible waste of time just because people won't share. The Internet is full of information and databases shared and worked on between like-minded folks, so why not track info?
If it's not already a thing somewhere, perhaps some clever chap will set it up.
Thanks for your info and reply. Very much appreciated.
Another excellent reply, thank you. I had not considered the copyright implications, which is perhaps why such a thing might not exist.
Some good examples of directory setups, too. Very practical.
Always happy to explain and help where I can. As long as it's not someone simply demanding I hand over everything I've got :-)
These sort of systems no doubt exist in the big boy world, where a new station manages to get a complete library from one of their sister stations, but in the world of Internet, Community, Hospital and other small stations there needs to be a lot of effort put in to get your own.
As you now see, copyright plays a big part in why you're not finding anything. If I wasn't for that I'd probably have offered to help if I could but I hope you'll appreciate why I cannot. That's apart from having a station to run here that is currently taking twice the time as I have the existing station running and now a mirror is running using PlayIt, and working well :-)
I understand completely and will continue to chip away at my database of oldies and easy listening.
You've got me thinking of how I might store the data in a more universal format. Something for me to look into when I get the time.
Again, I really appreciate your time on this thread.