Files split using web site data still don't match

Hi guys,

I have been using the information found on your site to split several Frank Herbert Dune audiobooks into chapters. These files were originally single file audiobooks with no chapter markers. I have used the information found on your site to create chapter information and import the information into the file using Chapter and Verse and then use another program that is capable of reading those chapter markers and split the file into the chapters. I have used MS Excel to do the necessary calculations since your site provides file duration and Chapter and Verse expects a time index. I have had a fair amount of success converting a couple of files using this method.

I’ve recently tried to perform the same conversion process for Sandworms Of Dune. The file names and durations all match up but the musicbrainz windows app does not agree with me. For the most part it says that none of the files match and the ones that do match have a yellow status instead of green. I have NO idea why the files are not matching up. I did however see one detail for this entry that gave me pause. Frank Herbert’s name appears to be associated with this audiobook. Frank Herbert did not write this book and thus his name should not be associated with it. Brain Herbert and Kevin J Anderson wrote this book. I am not sure if this is the issue.

Does anyone have any ideas? Any help you can offer would be greatly appreciated. I’m at a lose as to what the problem could be.

Providing links to the MusicBrainz Release(s) in question would be a good start for us/other people to have something to look into more.

These should be single-track Releases in MusicBrainz and not multi-track ones as it sounds like you’re trying to make them.


Thank you for your response, much appreciated.

Firstly, the link you requested to the data I am using as my guide is provided below:

Secondly. the data for this audiobook is in error where the author is concerned. Frank Herbert did not write Sandworms of Dune. He had passed away before this book was written. The book was written by Brian Herbert and Kevin J. Anderson.

Thirdly, I do not like single file audiobooks. They defy what an audiobook is, a book. A book has chapters or some other form of division to make them easier to digest. Hence, the reason for my efforts, tedious as they may be.

I have had success using you the data contained with your database and my method in splitting three other single file audiobooks into their chapters, Dune, Dune Messiah, and Children of Dune. Links to these pages in your database are provided below.

I have taken great care in my method to ensure that file names and file durations match what is contained in your database by copying the information into a text file using copy/paste and then formatting that data into columns which is then imported into MS Excel and the proper formula is then applied to the time duration of each file to generate the correct time index, basically rolling totals. The formula uses relative cell addresses to avoid errors.

The files are correct in both name and length but yet I am unable to get them to match. I have used both Lookup and Scan with the same result. I would appreciate any help you can provide. It is not my intension to be disrespectful in any way, only to convey how through I have been in my efforts.

Welcome VSCurtis!

You can hit ‘lookup in browser’ in Picard, and then search the database/website manually. Clicking the green ‘tagger’ icon should add the release to Picard for you.
You can then drag and drop your files (on the left hand side of Picard) into the right item in the database (in the right hand pane).
edit: or just paste the musicbrainz release URL into the search bad

A few notes:

You can edit the database - please correct any mistakes you find :slight_smile:
But please take note that MusicBrainz strives to document music releases as accurately as possible. This means that sometimes, in order to reflect a specific release as accurately as possible, we might sometimes have information that is technically wrong, but was released that way (by the label or whoever originally released the item).
Meaning: if the cover art credited ‘Frank Herbert’, we might leave it that way in the database.
That said, we do fix some things, for instance if the track order is printed wrong on a CD, so this isn’t a concrete rule - but something to keep in mind.

This just means that something, or multiple things, don’t match up between your file and the database information.
Probably the existing tags don’t match, which is fine. Just make sure they’re being re-written to something you want (check at the bottom)

And for the rest, don’t forget that you can drag and drop files to manually match them!

Hope some of that was helpful, can be hard to guess the problem from a forum post!
Good luck.

Thank you for the reply. I think we may be missing the point of my post. I’m trying to understand why this particular audiobook’s files are not matching up with the entry on which I based my efforts. File names are the same, file lengths are the same everything should match up no problem but it doesn’t. I need some insights as to why this is happening? Fresco asked for additional information and I have provided it as well as providing a detailed description of my process. I would appreciate some insights into what may be the cause of the problem. Perhaps an explanation of how scan and lookup function would be helpful. I might be able to use that information to determine the issue with my files.

Sorry, this hasn’t really helped me understand what exactly you want.
My previous post explains how to seed the release you want into Picard (if scan and lookup isn’t matching it), that you can then drag and drop your files onto the correct track, and then you can check why the file is red or orange by clicking on it and then comparing values in the bottom panel of Picard (the more different the values = more red).

If you’re after something different, perhaps explain exactly what you’re doing, step by step, and what behaviour you’re expecting that isn’t happening.

To answer your other question though:
Lookup searches the MusicBrainz database for the same track/album by reading the existing metadata/ tags on your files, and then checking it against metadata on MB. If your existing tags match MB’s this is usually the way to go.
(If you’re using Lookup you usually want to Cluster first which so that Picard knows all the tracks are meant to be together in the same album/release)
Scan ignores all existing tags, and “listens” to a chunk of each track, and then tries to find that same acoustic fingerprint (AcoustID) in out database. I don’t think we have many fingerprints for audiobooks but not sure. Please don’t add fingerprints to our database from files you’ve edited/ changed the length of btw :slight_smile:

Hope some of that was useful!

If this is the core of your question: The color indicates how well your existing metadata matches the metadata from Musicbrainz. Yellow is still a decent match, but there are some differences. You can see which tags are different by clicking one of the matched files and looking at the tags on the bottom. The left column shows the tags in your file, the right column the tags from Musicbrainz.

Also for some files not matching I think you should read a bit on how to best use Picard’s matching. In short: use lookup instead of scan, unless your files have nearly no usable metadata and lookup does not work. Also use the cluster button before doing lookup, that will group the files by album and Picard will search for the entire album for a match instead of attempting to match each file individually.

Thank you for clarifying these points for me. I was able to get the files to match and all have a “green” status. There was nothing wrong with any of them so I am still not sure what caused the issue. I basically had to match every file individually. It was very frustrating. Your post was helpful nonetheless, thank you

While that’s perfectly understandable, if the audiobook was only made available as a single-file download, that is what should be stored in MusicBrainz. Or at least the only one that should be marked official. I could live with having the cut versions entered as bootlegs, although I don’t particularly like the idea since they aren’t really widely available as such anyway…