This is not uncommon though.
2 Likes
I would disagree that it is common, although you said “not uncommon” which is a bit more ambiguous than a more affirmative answer like … is common or is not common. I know it does happen, but I do not believe it to be common for releases that are actually different. Same CD in two countries, sure, they really are not different. Two in the same country but paper covers printed in different countries, same. Same CD released on two different dates… again all the same really. Different digital file formats are different mediums. How often do different physical mediums share that barcode? How often do different digital mediums share a barcode, or even just use the same barcode as the CD for example?
There will always be odd ones out. I have seen barcodes used for two totally different products as I am sure you have as well. What needs to be remembered is that there are two different groups in play here. A prior editor here said there some some of us that care about different colors on releases. I have a few such things. On one, the barcode, catalog, disc TOC, everything is all the same, except the case color. That tells me that the vendor and store do not care, it is the same to them. To the user (the editor that cares) they are not the same. A barcode could end up not being unique at all, or a release could be given a different barcode for no reason at all. I think we need to look big picture, note but not dwell on the odd ones out, and look at the majority. Can a barcode be used to identify a CD? Yes, it can. It might give you a different pressing version, or a different printer that did the covers, but it is the same. Can the same be said for digital releases? Absolutely not.
I believe one of the issues here is not comprehending what digital releases are. On physical, you know the difference between a cassette, CD, vinyl, etc. That is easy to do. For digital releases, that same difference applies to file formats / encoders used. A cassette vs a CD is similar to a MP3 to a FLAC. The cassette is a different medium from a CD with different specs and capabilities. The cassette will cutoff around, what is it, like 15kHz or so. The CD will cut at approx 22kHz. The cassette can only be played on a player that accepts cassette, and same for the CD. On digital, the MP3 starts cutoff at 16kHz, the FLAC is 22kHz. A MP3 can only be played on a player supporting MP3 and a FLAC only on those supporting FLAC. If a CD is different from a cassette which is different from a vinyl, then the digital mediums are also different as the same logic of differentiation is used. Once we get to this point, then we need to look at the medium and identify the aspects on said medium that uniquely identify it. I am curious to see if others see it in the same way.
Well, we have the technology.
How many releases in Musicbrainz have UPC/EAN numbers that appear on more than one release?
1 Like
A detail, and not relevant to this discussion, but I thought to comment on this common misconception.
Depending on the encoder (settings), mp3 can contain a full frequency spectrum up till (and even slightly above) 20kHz.
http://wiki.hydrogenaudio.org/index.php?title=LAME#Technical_information
1 Like
Yes, the MP3 CAN reach up to approx 22kHz, but I did say “cutoff starts” and not anything like brick-wall cutoff unless it is set to obviously. It is not really a misconception at all, as MP3 encoder was designed to use lowpass as part of its primary compression factors. The LAME encoder is good (generally considered the best) and does not clearly show signs of this unless it is set to. The Fraunhofer encoder however, even with lowpass disabled, still shows signs of cutoff at 16kHz, the lowpass cutting is by design and turning it off just minimizes it… The MP3 can have frequency of up to 22kHz in theory, no higher, and you can set your encoder to lowpass where ever you would like it to. We know the 22 limitation as it is 44.1kHz sampling, which divided by 2 gives us 22.05, or basically 22kHz. Although it is a theory, it is possible to do with the right environment, settings and musical composition.
The data you point to is theoretical as it is the spec. But as you mentioned, it all depends on encoder, but also source, what is doing the process (front end GUI, encoder direct), etc. What is certain is when audio is analyzed, if you see a cutting at or around 16kHz, you are most likely looking at audio that is or was MP3 at some point. With the most recent version of LAME, encoding with the setting “lame -V0 --lowpass 20” still shows light signs of cutoff when compared to the lossless source or a CoreAudio generated M4A, v256 q2 encoded file from the same source.
I think it is sort of relevant as it shows that there are many misconceptions and misunderstandings of what digital files are and what is actually inside of them. To really understand it, unfortunately, you need more than forums and specs (no pun intended to the prior post). You actually need to do some audio analysis, look at wave forms, spectrograms, etc. There are many fake files out there, even those that people purchase, so just because someone or something says it is a certain quality does not mean it is. Thus the CDs you can find with brick-wall cutoffs and FLACs that are no better than a MP3. This is something that most will not or even cannot do, so what ever MB does I believe needs to avoid such things, or keep the logic of it away from the user in the interface.
1 Like
I can’t keep up with all the separate parts of this discussion, but it is common for multiple disks to exist for many different reasons. This is why I pointed at Dark Side Of The Moon as a good example of many variations of the recording out there.
I’ve been working through my CD collection, cataloging them into MB. Carefully IDing them both here and at Discogs. And it is very noticeable how often different barcodes and catalogue numbers appear as different on different releases. And I only have a small heap of 400 or so disks.
You can’t mash these all into one entry as then the artwork will be wrong. Artwork is more than just a single picture for a track - some people use this MB data with media centres and want all the artwork of their exact release displayed. Just like you need to know exactly which type of audio release it was.
I want all of my umpteen different copies of Dark Side Of The Moon to all sit side by side in my digital collection and MB still be able to tell them apart.
Also a far more important reason for differing barcodes are with re-release editions. Some people like knowing they have in their hands an original release of a piece of music. Bit like First Editions of books. Especially as re-releases often turn up with extra tracks tagged on. Or tweaked audio.
Please don’t let the database go backwards. Don’t throw away data we have now as good and clearly unique.
Audio differences, packaging differences, release year differences, artwork differences - there are many different reasons why different people use this database.
Who is proposing this? There is a conversation on release variations that will be discussed at the next IRC that will address all of this. Being able to take a release and further divide it. The whole focus of all this is to increase the level of detail while not making things too complex. I do not think there were plans of eliminating data discussed anywhere, but I could be mistaken.
That is a different track listing and not even part of the release variation concept. The variation is to handle a release that is the same in track listing. For digital, which is why it was even discussed, that means a release that is in MP3 form and one in M4A form, for example. Same track listing, different formats, but not enough that really justifies a separate release.
I guess this depends on how you define common. If there are 10 million releases, are 1 million of them duplicate barcode releases? That would be 10%, which could be common. Common, meaning by definition often, I do not believe is the case. I would be curious to see MB data on how many releases there in comparison to how many have duplicated barcodes. I seriously doubt that “common” will apply. As was stated above, we have the technology and data, but if something exists even at 1% of the time, that is not common but rare. I suspect that the duplicated barcode is less than 1% though.
Does anyone know how to run such a report?