Do most music CDs contain the needed info about their tracks?

Question

I see that many audio players (I mean multimedia software like Winamp or Foobar2000) have the ability of searching music information on databases like CDDB. Shouldn't such info already be available on the music CD itself? And is it there?

Some players display the CD contents and others do not. Is that info taken from the internet or the CD?

Solution 1

Shouldn't such info already be available on the music CD itself?

I think most of us, as consumers, would say yes.

And is it there?

Almost never in my experience. Certainly the software I have used to rip CDs to MP3s never seems to be able to obtain this information from the CD itself. I have read of a few exceptions (notably Sony since 1997).

There are probably several reasons for this, including:

1. Music-industry business-model
2. Inertia
3. The rise of digital distribution

Business model

The music industry traditionally made money from the sales of vinyl-records, cassette-tapes and audio-CDs. Protection of their copyright was seen by the industry as essential for their survival. To combat illegal copying of tapes they persuaded legislators to impose a levy on blank tape sales.

The music industry felt that facilitating playback on personal computers was facilitating the infringement of their copyright and thus facilitating their own destruction. So decisions concerning audio-CD contents and formats were heavily skewed against making anything easier for personal computer users.

Inertia

The audio CD has been established for a long while and there is no point making new CDs incompatible with existing CD-players. This means that care has to be taken if adding digital content to audio-CDs. Digital data and audio data on CDs use completely different and incompatible underlying formatting. This makes it tricky to mix both - though this can be done.

Given a large population of old CD-players, the industry has evidently not seen a benefit to them of "improving" the audio-CD format.

Their perceived use case is: You buy a CD, you put it in a dedicated audio CD-player attached to an audio-amplifier and loudspeakers. You sit down and read the track information printed on the CD cover.

Digital Distribution

Nowadays the trend is to downloadable content, at least paid-for MP3 files generally contain metadata for artist, album-name, year and genre etc.

It therefore seems unlikely that the music industry has any interest whatsoever in doing anything new with their CD pressing process. Its a dying business after all.

One of the greatest, coolest, but sadly least known and least often used tech things about CDs is "CD-Text." ... This has been out for 14 years and I can count on one hand the number of times I've actually SEEN a CD in my car have text associated with it.

^{From a 2011 blog}

Make that nearly 20-years now and no sign of general adoption by the music industry.

Why did CDs originally not incude metadata?

It is worth remembering that the audio-CD was merely a more durable and convenient-sized replacement for the pressed 12" vinyl album disc.

The latter was a purely analogue form with no digital information on it, just the analogue audio waveform in the form of vertical and horizontal undulations in a continuous spiral groove - with no distinction between tracks other than usually (but not always) a section of silence (no undulations) and wider spacing of the spiral (visible to humans bot not detectable by record-player). Any information about track names etc was present in the printed paper sleevenotes or on the printed cardboard sleeve itself.

So when audio CDs were invented, they took the same approach. They expected CDs to be played in dedicated CD music players, not in computers. Therefore the music was not stored on the CD in the sort of filesystem that a computer would normally use for data files. details of tracks were printed on the paper insert in the plastic CD-case - not placed in the CD contents in any way.

Similarly the audio data on an audio-CD was encoded on a single continuous spiral track. This is very different from the low-level formatting of computer data disks (hard, floppy, CD-data, etc) which typically have a large number of circular tracks arranged concentrically and divided into sectors.

There was no provision for data, probably because this had not been needed for vinyl records and because it would have complicated the manufacture of audio-CD players and made them more expensive at a time when the industry presumably wanted to encourage sales of CDs as a premium, and more profitable, product.

Note that, to identify a CD, programs on PCs have to extract some of the audio data (e.g. the list of song-offsets in the lead-in section of the track or the waveform of part of the first song) and use that as a key for a lookup in a database, typically a remote database elsewhere on the Internet. This is how software retrieves artist-name, album-name, track-name etc.

Some programs do look for CD-Text, sometimes only if they are offline and cannot contact a remote database. So presence of and use of CD-Text is a relative rarity.

There is no computer-readable metadata in most audio-CDs, not even an identifying product number.

Solution 2

The specifications for storing music on CDs is called Red Book.

There is an extension for Red Book called CD-Text. It allows for storage of additional information (text as album name, song name, and artist name) on a Red Book compliant audio CD.

Some hardware players are able to read CD-Text, however not all discographic labels include that information on the CD.

Software players as AIMP, Foobar2000, MediaMonkey, Media Player Classic, MusicBee, RealPlayer, SoundJuicer, Toast, VLC, Winamp (from V 5.31) and Windows Media Player (from V 10) can read CD-Text.

Solution 3

Not intended as an answer, but it doesn't quite fit in a comment. Actually, even the oldest CD Audio standard made it possible for a wealth of meta-data as a very slow bit-rate on the side of the audio data. It's actually so heavily utilised, that not just the track number, and the TOC (with the track offsets) are stored as meta-data, in the lead-in silent seconds, but even the mins:seconds display is actually read from the disc itself, as it goes! The players did not contain too much of a logic to actually count the time from the start of the track, but simply displayed what they have seen in the bitstream meta-data, as the P&Q (PQ) subcode. If you would want to be weird, you could possibly make the time appear to go slower/faster/backwards, or stop. Not that I'm aware of any disc doing it.

It was though possible to influence the track offsets so that the first track was a couple minutes into the disc, so you could scan back, and listen to a hidden track.

Any actual meta-data was very under-used though, sometimes you could see index marks inside a single track, separating parts of the same song/piece. I can only think of excuses why this metadata wasn't used. Most likely it wasn't seen as a differentiator, as most people would be interested in the actual artist's album, and not in another one, no matter how much metadata was on there :)

Perhaps it was even seen as an expensive feature to author the CD with all the meta-data, perhaps the studios didn't want to spend any further money on it on their cost.

Solution 4

What you are seeing is called Metadata, and are basically tags added onto the music file, but they are not part of the music file itself.

They are not taken from the internet, although if you use some internet music services they can be added onto the music file automatically.

Often files on cds have these tags, but most of the time (in my experience) they do not.

Solution 5

I work for a CD duplication/replication company and the users who have mentioned CD text are spot on.

This data is added during the mastering process to provide naming information for hte disc, artist and tracks as well as ISRC codes which help track radio play and royalties to pay the artist.

While many in car systems read CD text, most labels and artists don't even bother adding the CD Text (outside of ISRC codes which don't get displayed anyway) and there are even cases of bad and misspelled titles (See Queens of the Stone Age's "Like Clockwork" original release)

This information is written into the 'Lead In' area of the disc and is aimed at specific programs only.

View more solutions

Comments

[Admin]

Most of my players (in Linux) cannot automatically read the info. But some can, like VLC and Exaile. But do they take it from the internet or from the cd? My impression that the ones that can show this info (like the two players mentioned) will do it in most cases. If that info is from the cd, then most cds do have that embedded.

[CJL]

I know that VLC will take it from the net and the cd, but I am not sure about Exaile.

[jcbermu]

1. It seems that Sony includes CD-Text always, but other labels don't. 2. I know that car CD players can read CD-Text but I don't know about home players.

[Admin]

by music player I mean computer software not the hardware (added that in the question now). As the players that can show the track titles do so in most cases, if they do so from the cd, then most cds do contain that info.

[jcbermu]

Added information about software players in the answer

[Piskvor left the building]

In my experience, most CDs I have played do not contain this info (and net-capable players fall back to pulling the missing metadata from Internet, but that's beside the point).

[Piskvor left the building]

In my experience, VLC can and does retrieve metadata from network - if allowed. VLC will prompt about internet retrieval on the very first start (for privacy reasons); if you have disallowed that, it will obey. (You can still re-enable this in settings)

[Admin]

@Piskvor - you mean in VLC settings?

[Piskvor left the building]

Yes. It's a checkbox named "Allow metadata network access." In Simple mode,in the first settings tab ("Interface"). In Expert mode, it's under Playlist . (checked on VLC 2.2.1/Windows and 2.2.0/Linux here)

[Piskvor left the building]

Quite possibly. From my usage, it found most CDs - haven't compared it with other players.

[Admin]

I will post an answer on my experience too: I have found cds with tracks info, that can be read by some players offline. Doesn't that mean that the cd itself contains the info? And is it not that CD-Text?

[hmakholm left over Monica]

Actually vinyl records did usually have a visible indication of track boundaries; the spiral groove was would much less tightly during the inter-track silences, so tracks on the record would be separated by a thin annular region of (mostly) ungrooved vinyl.

[user5249203]

@Henning: Good point. I was thinking mainly of an indication in a form the player machinery itself could use (e.g. to skip to next song in response to a button-press rather than me having to don my glasses, squint carefully at the platter and lift,move & drop the stylus-head/arm assembly manually only to curse Pink Floyd for not visually separating some tracks. I may be misremembering, must investigate next time I'm up in the attic where my old LPs live).

[Admin]

The presence of CD-Text and the ability of a player to access it can be easily tested by switching off the internet connection and opening the cd with the player. Thus I have identified some CDs that have this offline info on them. Testing them offline with Windows Media Player, MusicBee and Foobar2000 I can say that these players cannot read CD-Text. The other players you mention can.

[Dmitry Grigoryev]

Audio CDs don't have files, they have tracks.

[BrianH]

I would also add that in the mid-late CD era some music companies explicitly tried to create CDs that would not play in computers by design. This was during the time of "Napster" when the music industry was actively against digital technologies, and there was no such thing as widespread "legal" digital downloads. No iTunes, no iPod. Supporting CD-text would have been a bizarre move, then, since it would actually encourage computer use of CDs, which many companies opposed, and digital CD playeres were effectively identicle to computers. en.wikipedia.org/wiki/Compact_disc#Copy_protection

[supercat]

The first CD player I bought would use its four-digit display to display the number of the current track and section thereof (I forget the term used for the sub-sections). That number is 01 for the entirety of all but one of the disks I've played in it; I have a Four Seasons (Vivaldi) disk, however, where the movements show up as 01-01, 01-02, 01-03, 02-01, 02-02, on that machine, but simply as four tracks on every other player. Interesting that sort of metadata used to be supported in ancient times, but has fallen by the wayside.

[Auspex]

@DmitryGrigoryev Pointless semantics. They're read as files by software. Call them tracks, call them files. It doesn't matter. Now, if you were to say that Audio CDs don't have tags, I'd have to agree with you. CD-Text is separate, and the software has to make assumptions about the "file" it belongs to.

[Dmitry Grigoryev]

@Auspex Well, yeah, I meant to say that CDs don't have files to add the tags to. Tracks don't have any structure to hold that kind of data.

[Admin]

on the other hand, talking about "the needed info contained by a CD", that includes also whatever data the software uses to identify correctly the CD and retrieve the proper information from online databases. Do CDs have like a unique ID number?

[Admin]

Great answer, but I think that the part "Is it there? - Never!" needs adjusting: you should say that it is there in some cases, but these are so rare that one might not notice it. In fact one can notice it only if one: 1. uses one of the programs e.g. mentioned in my answer and 2. is offline (which is rare these days)

[Hagen von Eitzen]

It may be worth mentioning that metadata is only useful if there is a method to display that data, which was not necessarily possible with every CD player (though that may be a chicken-egg-type problem)

[nekomatic]

@Auspex it's not pointless semantics at all - CD tracks are fundamentally different from files and are read in a different way. They are not part of a file system.

[Ellis]

Tracks on CDs do not and cannot have tags and metadata the same way an audio file on a computer can do. So your answer is a bit misleading, since it implies that the data is stored in the same way. CD-Text is a different matter.

[Greg Smith]

That you can access audio CD data as if it were a file is a weak and leaky software abstraction layer provided by operating systems. Look at a large sample of CDs and you'll find ones where that file-oriented abstraction breaks. Material hidden between the regular tracks is one example. If you want a correct mental model of how CDs actually work, you can't think of tracks as files. Audio CDs also do not have sectors even though computer CD players show the data that way. That's another leaky abstraction that breaks down if you poke at it hard.

[Auspex]

Of course it's pointless semantics. The tracks are files. The mere fact that they can be read by a computer is proof of that. It's not an abstraction at all. The [digital] data exists in a file-oriented format. That's the concrete implementation. Material hidden between tracks is just more files. Technically, information could be hidden inside tracks, and still be more files. And what "computer CD player" are you using that shows sectors, @GregSmith? That's a hard drive concept that doesn't exist on CDs.

[TripeHound]

For the record (pun intended) "no distinction between tracks ... (visible to humans but not detectable by record-player)" is not entirely true: some "gramophones" could detect the visual gaps between tracks YouTube clip of an 1987 RCA turntable and I had a Sony player from around that time that did the same.

[Lionize]

@jcbermu: FWIW my main home CD player does, but my old CD players (by which I mean, pre-2000) don't have a text display on the front anyway, just numbers. So it's kind of a moot point whether the hardware inside them is capable of reading CD-Text or not ;-)

[ᴠɪɴᴄᴇɴᴛ]

Many Marantz home players are able to display CD-Text. For example the current CD6005: For optimum visual clarity, the CD6005 features a deluxe two line fluorescent display that features alpha-numeric characters, and is compatible with CD Text music discs.

[DSimon]

That's really cool! I can imagine some great "haunted CD" pranks being implemented with this.

[rackandboneman]

Don't forget that the original Audio CD standard predates mainstream multimedia-capable home computers, and inexpensive alphanumeric displays on appliances, by several years. Also, in an old school CD player, there is very little CPU power to deal with such user interface issues - the typical CPU would have been 8-Bit, about one MIPS, and having 2-8K of firmware ROM and 64-256 bytes of RAM (yes, I am describing an MCS48/MCS51 here - still used today!); the "DSP" stuff would have been dedicated and hard wired circuitry.

[TripeHound]

@cipricus "Do CDs have a unique ID number?" In general, no. The lookup to CDDB uses a combination of the number of tracks, their lengths, the total length etc. to compute an often unique fingerprint (see wiki page).

[rackandboneman]

Not pointless. CDDA streams are different from the kind used to encode file-oriented data; a CDDA stream DOES NOT have the metadata that would even allow randomly accessing a byte-exact position in it, which would be crucial for filesystem access. Also, subcode channels do not have an exact timing relationship with data OR audio streams. On top of that, an audio stream is expected to deal with minor degradation due to media defects - deadly for data, otherwise "mode 2" data CDs would be common. Over 10% of a data stream is just extra ECC!

[Joe]

+1 Amazing answer with lots of things I did not know for a question which could essentially be answered by "No."