I'm uh.... Guessing y'all heard about spotify data getting dumped
2025-12-23 35711, 2025
sakgoyal[m]
Would it be legally/ethically/morally a good idea to scrape the metadata to build out the MB database for popular music?
2025-12-23 35739, 2025
volumetrique[m] joined the channel
2025-12-23 35740, 2025
volumetrique[m]
Mapping to MBID would be a struggle just like always and be unreliable, I guess. The only mapping that would be reliable is ISRCs, and while MusicBrainz has a good list (about 5M ISRCs) it would still be pretty meh.
2025-12-23 35740, 2025
volumetrique[m] uploaded an image: (33KiB) < https://matrix.chatbrainz.org/_matrix/media/v3/download/chatbrainz.org/gWicbDAwddRvsGZoOEZLwKDc/image.png >
2025-12-23 35741, 2025
BlastboomStrice[
Btw
2025-12-23 35744, 2025
volumetrique[m]
while on this topic: This is such terrible lack of faith btw. I think it's terribly phrased. Pretty normal for MB to have less ISRCs considering the data quality, lack of full automation of imports from streaming, and ISRCs being submitted through outside tooling only
2025-12-23 35751, 2025
BlastboomStrice[
Did they scrape them in mp3 or flac quality?
2025-12-23 35715, 2025
BlastboomStrice[
I did share the news like 4times yet didnt open the link💀
2025-12-23 35746, 2025
BlastboomStrice[
Oh ok
2025-12-23 35751, 2025
BlastboomStrice[
> For popularity>0, we got close to all tracks on the platform. The quality is the original OGG Vorbis at 160kbit/s. Metadata was added without reencoding the audio (and an archive of diff files is available to reconstruct the original files from Spotify, as well as a metadata file with original hashes and checksums).
2025-12-23 35751, 2025
BlastboomStrice[
For popularity=0, we got files representing about half the number of listens (either original or a copy with the same ISRC). The audio is reencoded to OGG Opus at 75kbit/s — sounding the same to most people, but noticeable to an expert.
2025-12-23 35752, 2025
volumetrique[m]
Most popular tracks are as vorbis 160kbps, while the long tail are in pretty meh opus 75kbps
2025-12-23 35753, 2025
BlastboomStrice[
Yeah I see
2025-12-23 35700, 2025
BlastboomStrice[
I mean ok cant blame them but ok
2025-12-23 35719, 2025
aquatix[m] joined the channel
2025-12-23 35720, 2025
aquatix[m]
"sounding the same to most people" because of crappy Bluetooth earbuds? 😅
2025-12-23 35738, 2025
BlastboomStrice[
I guess theyre good enough to cause a blow to spotify
2025-12-23 35757, 2025
BlastboomStrice[
The average person probably wont care
2025-12-23 35732, 2025
BlastboomStrice[
(And personally Im dont think I can understand the quality difference, maybe if I use better equipment I will)
2025-12-23 35719, 2025
glabnob665589 has quit
2025-12-23 35716, 2025
rogorido has quit
2025-12-23 35710, 2025
virtueisdead[m] joined the channel
2025-12-23 35711, 2025
virtueisdead[m]
<volumetrique[m]> "while on this topic: This is..." <- yeah, assuming they strictly pulled the data off of spotify, i would not trust most of it; a lot of streaming metadata is between unreliable/questionable to outright misleading/wrong
2025-12-23 35701, 2025
notthejohn[m] joined the channel
2025-12-23 35702, 2025
notthejohn[m]
Yeah as anyone who as ever used Harmony can probably vouch, streaming metadata is a total mess (especially by MB standards, haha)
2025-12-23 35712, 2025
PastelSkies1047 joined the channel
2025-12-23 35702, 2025
virtueisdead[m]
<notthejohn[m]> "Yeah as anyone who as ever..." <- i mainly use harmony with bandcamp since i dont listen to music on streaming services so im not likely to discover new music thats only on them, so i probably havent personally experienced the worst of it
2025-12-23 35729, 2025
notthejohn[m]
I think what gets me a lot is just how different one release can look across all the different services it's on; artist credits look different, sometimes featured artists are in different spots, different capitalization standards, etc etc
2025-12-23 35702, 2025
notthejohn[m]
also stuff like (2022 Remastered Version) in every track title on albums that are remasters which needs to get trimmed out
2025-12-23 35715, 2025
NovaAnyElleNyaCa has quit
2025-12-23 35730, 2025
BlastboomStrice[
<notthejohn[m]> "I think what gets me a lot is..." <- And transliteration
2025-12-23 35708, 2025
BlastboomStrice[
Artists on spotify tend to transliterate the words, while they dont do that on their other platforms (like youtube)
