Out of all the various identifiers we have in music, it can easily get confusing. From an ISNI, IPI, UPC, ISRC, IPN to ISWC it's a lot to take in. Starting with the ISRC, why should you care and what is Muso.AI doing to push it forward
Identifying records in a system has been a standard since the early days of computing. Each row of information has its unique number to locate it almost instantly. Additionally, it helps separate it from other records that might have similar values.
However, when you communicate from one system to another, you won't have the same internal identifiers. This is why you need industry-standard ones, that every system uses.
When music companies send data to each other it's therefore extremely important to adhere to these identifiers, so that we're all talking about the same thing.
Each track has two different pieces of copyright. A recording and a musical work. When we're communicating data for a recording from one system to another, we use the ISRC (international standard recording code).
The International Standard Recording Code (ISRC) system is the international system for the identification of recorded music and music videos.
Most data communication between organizations happens on the ISRC level, as it's the most used identifier and has the highest coverage.
Artists should not create these themself, your distributor or label will be able to generate ISRCs for you and you should not ever change. When tracking the performance of sales and streams, reports are linked to the usage of a particular ISRC.
Each track that is on a streaming platform or able to be bought commercially therefore has a unique ISRC.
If you have a track that is used in several albums, you can choose to use the same ISRC to consolidate sales tracking. However, if you want to track the usage in the different albums separately, you can have different ISRCs for the same tracks
The Problem with Bad ISRC Data
Since a lot of the industry hasn't properly linked these tracks to each other, it's difficult to show credits for the same recording in different albums. Credits are directly tied to recording and it does not matter if the track is one album or twelve, the credits should stay the same.
For example, a track could have come out as a single. Then a year later part of a full-length album. Part of a remix album and then 10 years part of a compilation album. Each one of these tracks could have a different ISRC.
If a change happens on the credits for a track on one album, every other track that has the exact same recording in other albums should be updated. Since they don't always have the same ISRC, this is challenging, but not impossible.
This is one of the big problems Muso.AI is solving with our system. Identifying songs that should have the same data, but since they were distributed at different times don't, and linking them together.
Ensuring that if you ask for credits on a track, that we have the most complete possible data, even if there are multiple distributions of that recording.
Taking a few steps further
Similarly with showing credits on a profile. We think it's not best practice to show a producer credit, for example, on the single, the album, the remix album, and the compilation album.
This creates a messy view of what someone has actually worked on. Instead, we return this credit once and show which albums this track is on. Additionally, the credit count on a profile is only increased by +1 with this credit, and not for each separate appearance in different albums.
Lastly, as we're showing stream count on our platform it's even more important to show a complete picture. If we take the same example as above, I don't want to know what the streams are for the 4 separate distributions of the track, I want to know the total aggregate of streams for that recording - regardless of how many albums contain this track.
Now, all of this is a beautifully painted picture but has proven more difficulties than first anticipated, and further shows the inconsistency of data in the music industry.
Piece by piece, they shall be solved. Starting with linking similar recordings