Freesound Forums

I'd be surprised if this has never come up before, but couldn't find any related search results. My desire is to download files from freesound.org that are embedded with metadata - particularly useful would be Description, Keyword, Location, Designer, etc., all fields that already exist to fill in when uploading to freesound. If it's possible to use Soundminer software to serve that purpose, then all the better. Is anybody out there doing this already?

Hey, this is a good question.

In order for this to be possible, you'd need to get a BWF (Broadcast Wave File). These are just regular WAV files that have an XML metadata chunk at the end of the file with those fields available and many more.

Currently Freesound does not supply these types of wav files as a standard with metadata in the correct fields because we just supply the original file as it was uploaded, so if it was added as BWF with metadata, then it would be intact when you download it, this also applies to mp3 files with ID3 tags.

Here is an example I just uploaded using BWF MetaEdit: https://freesound.org/people/Headphaze/sounds/574116/

This however is a good suggestion for future feature, allowing input fields that are entered in the description section to be transferred to a download. Not quite sure how this could be achieved, but it would likely require the original file to be .wav and if the field is empty it could be appended with supplied data.

I agree this could be quite useful

Yep, indeed that downloads with metadata fields for Description, keywords, library. Understanding that many amateur freesound users aren't familiar with tagging to BWAV's (as should be expected and embraced in the CC community!), the suggestion is definitely to somehow marry the descriptors queried for by Freesound during uploads into either a BWAV that is available for download on the web, or (and this would certainly require less or no re-encoding of all the content that currently exists on Freesound) a CSV database, which could be referenced by various sound library software for local metadata encoding on the user's end.

Actually, that is probably the simplest solution to provide a CSV table. And quite easy to generate. Wouldn't that be a bit tedious to marry with soundminer or other sound library software for each sound? What's the process exactly with CSV referencing?

Providing a PCM-BWF for every file in the database would more than double the data storage requirements so that option is just not viable.

Here's the Soundminer documentation regarding text import (CSV, tab delimited). I've done plenty of this to tag files that came from CD burns (remember those?), but looking at the linked page, it seems like "Direct Import, which creates a new databases and new additions to an existing database" is the way to go. Providing a CSV or TXT file along with audio downloads would give the user the ability to import the metadata while bringing the audio into their library. It'll honestly be somewhat cumbersome for sounds downloaded individually, but extremely powerful for sounds downloaded as packs, compressed as archive images, and further incentivize users to upload sounds as packs.

https://info.soundminer.com/docs/import-text

Thanks for sharing

This is definitely something to note for consideration

If you need a guineau pig for the development of this don’t hesitate to reach out!

Hi,

This feature suggestion looks interesting. It would be nice if you could open a ticket in github providing details: https://github.com/mtg/freesound/issues

We are already doing some file generation when downloading a pack (the license file), so it should be quite doable to do what you propose. I can't guarantee that it will happen soon, but if the ticket is opened this looks like a nice small task for one of our students

Let me know when the ticket is open so I can properly tag.

done!

https://github.com/MTG/freesound/issues/1563