Adding ISRC Data

Tags: #<Tag:0x00007fa5c2fe0c08> #<Tag:0x00007fa5c2fe0b40>

Has a replacement been made for the “kepstin_magicisrc” batch ISRC add? I wish I knew how to create things like this.


Yea, the tool is still using web service v1, which has been deprecated. Pinging @kepstin to see if he’s free enough to fix this.

I’ve actually already done most of the work to port the ISRC submission tool to the new webservice version, but I’m blocked on missing webservice features.

I’ve updated the page at with details of why it’s not working, and a link to the blocking MusicBrainz ticket: MBS-6033 - add CORS preflight support

Also consider voting for MBS-10053 - this feature really should be part of the MusicBrainz web interface proper.

You can play around with the new (not yet working) interface at (note that it’s linked to rather than the main site).


Does anyone know if there is a way to view a report showing daily ISRC submissions. I am just curious to see the impact to the database after this change. I have tried manually adding the ISRC data. But now it went from a twenty second workflow to about 5 minutes and a thousand clicks. I have stopped adding due to the health risk to my mouse.

You can do an edit search.
Same for me, even from CD.

I need help!

I wanted to merge recordings (lead single and album version) since considering audio data they’re the same. There’s only minor dynamics difference + a short trailing space in one of them. Compare the spectrals for single track:

…and album track:

However they have been assigned different ISRCs (single track: GBAFL1800302, album track: GBAFL1800329).

Also there’s a minor crediting variation. Single featured “Spawn”, an AI algorithm trained to make the song. On the LP they dropped the credit, leaving only Holly Herndon and Jlin.

So, should I leave these recordings separate just because they have been assigned different ISRCs, even though they’re the same?

1 Like

I have seen different ISRCs merged.

IMHO the most important step a merger can do is exactly what you are doing - listen and analyse. You can hear and see that there is no difference. And then you have done deeper research to check that they are the same.

If you can start both these tracks at the same time in two different audio players and not get a problem… then pretty sure they are the same track.

This is more than some users do - some merges are done just on the AcoustID or name and lengths which doesn’t catch the small differences. There was an example a few weeks back of a track that was used in a Bond soundtrack. The difference there was only really spotted by listening as it was a small change in the intro. AcoustIDs were failing in that case as many of them had been attached to the wrong version due to confused compilation disks. Pretty sure that one also had multiple ISRCs.

Your analysis is ideal. Though I would write it into the comment and not rely on this forum. In five years time the forum will probably have moved address (again) and would then loose these notes.

(And now get ready for your post to be moved to a new thread… counting down… 5… 4… 3… )


It would be neat if Acousticbrainz could capture Spectrograms with their audio data capture tool. Then every recording would have a way to visually compare the entire recording along with the more detailed but shorter acousticID.


Maybe I miss the announcement but looks like MagicISRC is working now :wink: Thanks @kepstin for the great work!

Someone needs to combine MagicISRC with so an EAC rip log can be combed for this data.

ISRCs are on my “TODO” list as they are a pain to deal with on a Windows 10 PC. Or at least I ain’t sussed 'em out yet for my 500CDs.


It will be much faster to use to submit both Disc ID and ISRC that to go through a rip and several indirect tools, IMO. :wink:


Maybe not quicker If you have already ripped 500 CDs and have them sitting there as EAC logs with the data in like me. :wink: I’ll work it out at some point - just it is another one of those tools with no working manual. :smiley:

Are you talking about It comes with manual built in (but you literally just run it like (or isrcsubmit.bat under Windows) and it’ll Just Work™ far most of the time):

freso@koume ~> --help
isrcsubmit 2.1.0 by JonnyJD for MusicBrainz

This python script extracts ISRCs from audio cds and submits them to MusicBrainz (
You need to have a MusicBrainz account, specify the username and will be asked for your password every time you execute the script.

Isrcsubmit will warn you if there are any problems and won't actually submit anything to MusicBrainz without giving a final choice.

Isrcsubmit will warn you if any duplicate ISRCs are detected and help you fix priviously inserted duplicate ISRCs.
The ISRC-track relationship we found on our disc is taken as our correct evaluation.

Usage: [options] [user] [device] -h

Please report bugs on

Yeah - I am just weird. I like to read about the [options] before running things. And the link from the github page back to the forum goes no where. Hence it gets kicked down the road for another day. :slight_smile:

That site is ancient. You can see the linked forum thread here (but note that it’s over 7 years old (even if not much has changed)):

Also, online usage documentation here:


Haha - documentation? Summarised by “Install it and run it” :rofl:

For detailed usage see: -h

What a comically trusting world that is from. Thats the trouble with old tools that everyone uses, the manuals get forgotten. Please don’t drag this OT as I’ll get to it another day as I have not even run a python script on that PC before.

I’ve been trying to get this to work and I just can’t. I have ZERO knowledge of python. I try following the instructions and it doesn’t work. When I try to “just use it” by clicking on isrcsumbit I get nothing but errors. I’m not sure why these types of scripts can’t be installed with an .exe. I’m not a computer programmer. I’m about to give up. I’ve been trying for 8 hours to get it work and just can’t. I can’t even get mediatools.exe to run either. This is what I get when I run isrcsubmit.bat

Traceback (most recent call last):
File “C:\Users[user name]\Downloads\isrcsubmit-2.1.0-win32\”, line 47, in
import discid
File “C:\Users[user name]\Downloads\isrcsubmit-2.1.0-win32\”, line 30, in
from discid.disc import read, put, Disc, DiscError, TOCError
File “C:\Users[user name]\Downloads\isrcsubmit-2.1.0-win32\discid\”, line 24, in
from discid.libdiscid import _LIB, FEATURES
File “C:\Users[user name]\Downloads\isrcsubmit-2.1.0-win32\discid\”, line 108, in
_LIB = _open_library(_LIB_NAME)
File “C:\Users[user name]\Downloads\isrcsubmit-2.1.0-win32\discid\”, line 99, in open_library
return ctypes.cdll.LoadLibrary(lib_name)
File "C:\Users[user name]\AppData\Local\Programs\Python\Python38-32\lib\ctypes_init
.py", line 451, in LoadLibrary
return self.dlltype(name)
File "C:\Users[user name]\AppData\Local\Programs\Python\Python38-32\lib\ctypes_init
.py", line 373, in init
self._handle = _dlopen(self._name, mode)
FileNotFoundError: Could not find module ‘discid.dll’ (or one of its dependencies). Try using the full path with constructor syntax.

Press any key to continue . . .

1 Like

I believe I had the same problem when I first tried to use on Windows a year ago. The issue was related to a 64bit Python installation failing to load a 32bit DLL. Unfortunately I haven’t bookmarked or written down my solution, but I remember that I have downloaded a 64bit version of discid.dll to replace the the 32bit version that was included with