Hacker News new | ask | show | jobs
by mahoho 1563 days ago
If I understand the article correctly, are (some/most) files for sale on the iTunes store taken from CD rips rather than made directly from the masters?

That sounds impressively sketchy; anyone who has used AccurateRip can probably testify that CD ripping errors and manufacturing errors are surprisingly common.

10 comments

At the time, the CD was often the practical master. Many recordings had come from analog tape, sent to a mastering house, who burned the final master to a CD.

Anyway, I skipped this detail in the original article, but Apple let go of the requirement to use their special “put the CD in the drive” tool. We were able to deliver using master WAV/FLAC files, converted to their AAC requirements, and uploaded.

Yep, I was on the other end of this at the time and had a bunch of those gold colored CDs labeled MASTER for each album
Yeah I was going to say, surely you did not end up ripping 200,000 CDs in a couple of weeks
It was 5,000 clients who paid $40 each totaling $200,000. Not 200,000 CDs.
If they hadn’t, surely there was a scriptable method that didn’t involve re-ripping?
Or I guess you might have been able to do a scriptable method that does involve re-ripping.

That is, stick a CD-RW in the drive, and write a program that would:

(1) Erase the CD-RW, then burn one album's worth of WAV files to it. (Ideally, do it accurately like with a cue sheet file.)

(2) Drive the Apple software's GUI (using AppleScript?) to enter the track metadata, re-rip, and upload.

(3) Repeat until done.

If something ejects the CD-RW, that might mess up the automation. Some drives will pull the CD back in if the tray or disc bumps into something while ejecting, so maybe a strategically-placed heavy object is enough.

Note that this is from the early days of iTunes--things could be radically different behind the scenes now.

I suspect if it isn't listed as 'Apple Lossless' or one of the other fancy labels, its probably originally from a CD rip somewhere. I know from listening to niche-y music, that music catalogs can often be wrong and will be published to multiple music sites. For example Junior Brown's album 'Junior Brown: Greatest Hits' has a track on it that is half glitches AND its the exact same on multiple services and has persisted for years even though I reported it several times on each service. There's also a sea shanty album where half the tracks are static. I reported it to iTunes and Amazon and received boilerplate responses. I then sent an email to the actual band (hard to believe, but its a bunch of old guys) and they contacted their record company...but even they couldn't get it straightened out.

Even when it lists Apple Lossless, there can still be errors in the files.

I found an album from 2001 on Apple Music recently and discovered one of the tracks cuts out at just after one minute, even though Discogs reports the track should be 4 minutes 30 seconds (Slide - Closure (Lounge-Tech Mix), on the Nu Progressive Era compilation: https://www.discogs.com/master/90383-Red-Jerry-Nu-Progressiv...). The album is listed as "Apple Lossless".

I went and bought the original CD version JUST to have that one track in full.

"Mastered for iTunes" is the tag that means someone actually listened to it. (Technically it means the AAC file was checked that it doesn't have more clipping after encode than before it. There's a public PDF about this out there somewhere.)

It's usually missing, but hopefully someone listened anyway.

Back in the early 2000s, I worked for a company that was supplying audio media to Apple, Spotify, etc. and yeah, the record companies would ship them boxes of CDs for ripping, cover scanning, track listing inputting, etc. For some of the companies, it was the only way they had - they didn't have the metadata or cover art in easy digital form, masters available, etc., especially for older stuff.
LoudEye?

They had a cool office and a massive SGI machine.

Yes. I was initially in charge of getting all the music from all the major and indie labels into our system when I worked at what was the biggest competitor to iTunes in Europe. It was 100% from CD. I remember at the end we had a storage unit with hundreds of thousands of CDs. We had teams of young girls and guys working day in, day out ripping CDs.

We did our absolute best to get the high quality rips we could. We sat on the forums and figured out what the best CD-ROM drives were, even if it meant buying really expensive SCSI versions.

But none of the labels had anything in digital format in prior to 2003. I think the majors only started their conversions of their catalogs in about 2004 or 2005.

Just some other background I'll throw out there - the company I worked at opened all the doors at most of the record labels. Most weren't ready to sell their stuff online (WTF) and needed a lot of persuasion. After we got them to sign, Apple would follow us in days or weeks later and have a nice easy job. That was how we found out Apple was trying to build a music store of their own.

And Apple had a good time with the labels. At that time, and perhaps even now, most record labels used Macs for practically everything they did, even admin stuff. So when we went in with a mostly-PC demo, they looked at us sideways. Apple could slide in with shiny stuff and impress them more :)

@sivers: Did we have all your catalog? This was OD2 (On Demand Distribution) in the UK. I have a feeling we did?

It was probably the fastest way at the time to build up their initial catalog quickly; for masters or the best quality recordings, they would need some way to get the music from the masters into the software, and I don't believe Apple had any good audio in ports.

I do recall at some point they had a headphone jack that also supported optical, but don't quote me on that.

Anyway, it would have been better if they had an app that accepted .wav files or something like that.

They could have easily hired a specialized company to receive masters and send them digitized versions. They weren't as ludicrously opulent as they are today, but they were still a pretty wealthy and profitable company.

But why pay, when you can get your own vendors to do it for free after a little song and dance by the Jobster? That's much more Apple.

Not even free. Vendors needed to purchase Apple hardware, did they not?
The iPod helped save Apple. IIRC it, even more than the iMac, helped return them to profitability.
> I do recall at some point they had a headphone jack that also supported optical, but don't quote me on that.

They used to on the Macbook Pro laptops (I have one). Not sure if it's still a thing with the new Pro.

They ditched the TOSLINK port for the 2016 models.
In Apple Music today they indicate if the track is lossless and if it's taken from a master. I believe that lossless studio master rips are 24-bit / 192 kHz (CD is 16-bit / 44 kHz).
The originals are whatever the artist has. There isn't going to be a standard, especially for indie music, and there's nothing magical about 24-bit/192khz that makes it a good format.

CD quality is already perfect, impossible for a human to hear any improvement on. Except of course that it's only available in stereo post-mix.

>> anyone who has used AccurateRip can probably testify that CD ripping errors and manufacturing errors are surprisingly common.

CDs have significant error correction codes so if it sounds right it IS right. Having said that, I have one song I always skip because it ripped badly and I've never got around to re-ripping it and replacing the bad one. But it's obvious that it's a bad rip to the point that I skip the song so I don't have to hear the glitch.

In other words, if they checked each song before uploading it would be fine.

> CDs have significant error correction codes so if it sounds right it IS right.

For data CD formats, yes. For audio CD formats, readers are allowed to interpolate over uncorrectable errors (https://en.wikipedia.org/wiki/C2_error), which would not necessarily result in an abrupt skip or pop.

>so if it sounds right it IS right

No, player will interpolate samples with detected but uncorretable errors. Uncorrectable error rate of CD-DA was deemed too high for CD-ROM, thus it uses additional layer of ECC data on top of it.

> Having said that, I have one song I always skip because it ripped badly and I've never got around to re-ripping it and replacing the bad one.

Heh. I have one that has about five seconds of silence, at the end of an album, then about ten seconds of horrifically loud noise. It still catches me off guard every time, but it’s not in heavy rotation so I still haven’t gotten around to trimming it.

I wonder if this is why there is a Dinosaur Jr song with a massive bad-rip hole about two-thirds in, on all the streaming platforms.
https://en.wikipedia.org/wiki/Cdparanoia

https://xiph.org/paranoia/faq.html#progbar

"""A plus indicates not only frame jitter, but an unreported, uncorrected loss of streaming in the middle of an atomic read operation. That is, the drive lost its place while reading data, and restarted in some random incorrect location without alerting the kernel. This case is also corrected by Paranoia."""

Article was written in 2010 about events that took place in 2003 or so. Seems like a strange approach even then but I don't know much about music distribution back then, or now for that matter.
It's not clear to me if that is still actually the case.

Seems likely some are but as their system presumably grew more automated I'm guessing that's not so much the case anymore? Possibly?

Definitely not, there's a specific website for uploading stuff, it's not done trough iTunes. If you're a musician or a small label you're gonna use something like CDBaby or DistroKid though, which uses an API or something equivalent.
I imagine it hasn’t been the case for any new or remastered music added in at least the last 10 years. They upgraded to 256 kbps in 2009 so CD-originated music surely would have ended by then.
This is why would always dump the CD to an ISO file then mount that and rip directly from the virtual CD.
That is factually incorrect. An ISO9660 is a filesystem on a data CD. An Audio CD is just a stream of bits. That is why you need to rip an audio CD, the CD player needs to transform that stream of bits into blocks of 4096 bytes. It has to remember where the previous block ended and the next block starts. For many years, you had to buy a luxury brand like Plextor to be sure that ripping process would happen without much stuttering and gaps.
Audio CDs do have framing information (in the subchannel). However, the subchannel has no error correction (only basic error detection), so the CD player has to interpolate across subchannel errors (which are normal and common) to figure out where it is, and doing that properly can get complicated.

Also, the audio frame size is 2352 bytes. Those correspond to 2048 data bytes for data CDs (plus extra error correction).

Sure, but many of the cd cloning tools at the turn of the century would name your bit stream dumped to file with a .iso file type, not .IMG . That's what I'm referring to. I think it was wondering the meaning of that file format that led me to even learn what an ISO standard was.