Slashdot Mirror
Apple Says It Doesn't Know Why iTunes Users Are Losing Their Music Files (theverge.com)
Earlier this month, an Apple Music user James Pinkstone claimed that the online music streaming service deleted 122GB of music from his library for no apparent reason. Several Slashdot readers noted they had also faced a similar issue or knew someone who had. The iPhone maker has now acknowledged a bug in iTunes that is apparently causing the glitch, however, it adds that it doesn't really know why "some" users are facing this issue. The Verge reports: The company confirmed that "in an extremely small number of cases, users have reported that music files saved on their computer were removed without their permission." However, Apple was unable to reproduce the bug, indicating it doesn't really know what's going on here. The company adds: We're taking these reports seriously as we know how important music is to our customers and our teams are focused on identifying the cause. We have not been able to reproduce this issue, however, we're releasing an update to iTunes early next week which includes additional safeguards. If a user experiences this issue they should contact AppleCare.
The person in question is a composer. Further a large amounts of their music is in wav versus compressed audio formats to serve as a master of their work.
Thirty four characters live here.
Based on the contents of my 24/96 FLAC folder it averages about 1.2GB/hour. Getting into the hundreds of GB is very doable.
They're really claiming that they don't know why this is happening?
Note to iTunes software engineers: grep through your code and find all calls to delete(). Investigate the code in those areas. If you can't add some basic debug code - or if you're unsure of what "grep" is - maybe you should consider a move to management.
i don't care about your corpus callosum. maybe you should take better care of it?
read further, young padawan.
there's more to it.
lots of bad hype and bad info; and that confuses the issue. but once you remove the bad info, the good info does make sense.
i2s, spdif, receiver chips, dac chips, file processing and DSP on the files.
if I'm going to 'do things' to photo files, I do it at raw and full bit depth. well, audio is becoming like that and even vol control is a dsp math operation. you want high bit depth for doing math on files.
what I agree with you on is: at the final stage, you COULD pump out audio at 44.1k rate and even at 16bit res, once all the maths are done. but there's no harm in leaving the vol-controlled or eq-fixed audio in high bit rate. dacs and stuff like it that way and there's just no down side to leaving it in high res format.
--
"It is now safe to switch off your computer."
http://www.cnet.com/forums/dis...
http://9to5mac.com/2014/12/03/...
http://money.cnn.com/2014/12/0...
It appears they have in several iterations over the years. The GP is likely conflating several instances but i cannot find fault with it. I am somewhat concerned about your strict denial without even a simple Google search that brought up the links i posted and more.
iTunes is a dumpster fire of astronomical proportions, but at some point the user has to take some responsibility for not entrusting valuable data to a flaky consumer-grade application. This sounds like a case where the wrong tool was used for the job.
That's cute. When you're a DJ, 44khz isn't enough. 16 bits is fine. Do you know why? Should I bother telling you?
It has to do with that Fourier transform thing. When you have two different tracks that are in different keys and have different tempos, you need more samples than 44khz to do the pitch and tempo shifting to match the two tracks.
Keep feeling smug, though. Clearly I must be wrong.
He probably didn't. iTunes scans your computer for music files and automatically "adds" them to your iTunes library. And then deletes them for you when it "makes them available from Apple Music."
Actually, there is a very good reason. Anything above half the sample frequency will create aliases in the hearable range. To prevent that, you must use a low-pass filter to remove those higher frequencies but low pass filters are not perfect. They begin cutting before their set frequency. You can use multiple low pass filters, but then you introduce other distortions. Keep in mind, these are analog filters.
So high quality gear is sensitive well into the ultrasonic and uses a cutoff high enough that they begin cutting the signal above the threshold of hearing and so the cutoff is in the ultrasonic range. The sample rate needs to be double that frequency.
Consumer grade gear such as CDs use a "brick wall" cutoff and sample at 44K knowing that the audio won't typically be further mixed and that very few people have headphiones that go anywhere near a frequency response that would reveal the distortions, but that's consumer gear.
Actually 96 kHz makes sense in a mixing setting too if you're altering the pitch of a recording like autotune, then you're effectively stretching/compressing the samples causing interpolation errors in the time domain not just the amplitude. But yes for a final mix 48kHz/44.1kHz is enough, though since both systems are in use another advantage is that you can get perfect sound both for a 44.1 kHz CD and a 48 kHz soundtrack for a movie while a conversion between them would be slightly lossy. But to audiophiles "studio grade" has become a thing like "military grade" is to crypto.
Live today, because you never know what tomorrow brings
Just like they didnt know why 'some of users laptops GPU died' for almost 4 years before admitting fault and reluctantly being forced to a recall.
Who logs in to gdm? Not I, said the duck.
From what I read from those links you posted was music being deleted from the apple devices them self (iPod) when music was put on them through other software or that was using a hacked version of apple's DRM. I actually remember that Real Networks had figured out a way to put a hacked version of apple's DRM on music purchased from them in order to keep their music protected while still being compatible with apple's devices. Apple was able to detect this. But from what I recall and what those links say, it did not hut through the computer and delete them. It would remove them from the device or inform the user of the music with invalid DRM and ask them to restore the device. I have never seen or heard of iTunes simply deleting MP3s or any other audio (or even video) format for simply not having been ripped by it.
Weasel words. When I worked in support, if something failed 100% of the time we officially described the problem as "under some circumstances...".
He probably didn't. iTunes scans your computer for music files and automatically "adds" them to your iTunes library. And then deletes them for you when it "makes them available from Apple Music."
Except it doesn't move them to the iTunes library - it copies them. So unless you delete the originals yourself, you've got another copy of the file... but I can imagine a lot of people would do exactly that, because why would you want two copies on your computer?
I am not really sure how much I agree or disagree with the gp post. People do need to take some responsibility, and we tech-heads know better than most what can go wrong with tech. But most people aren't like us; and all of these companies sell their hardware and software with the message that it'll magically take care of everything for you - Apple perhaps more than most (it "just works"), although Google and Microsoft are certainly not hubris-free.
#DeleteChrome
These two videos are an excellent introduction to audio signal theory.
Highly recommended to all readers enjoying this thread.
i've never really used apple products, but my wife does the macbook iphone thing. we've had two experiences with it randomly deleting her shit.
first time (this was about a year ago, not sure what itunes version):
she got a new iphone, all was well. we wiped the old iphone. one day she dug up her old iphone, and decided to start using it to play music in her car. plugged it into the macbook.
itunes asked if she'd like to sync with the new device. she said yes. it deleted all of the music on her computer, including physical files.
plugged her new phone in, it acted as if it had never seen it before, and asked if she'd like to sync. it then deleted all of the music on her new phone as well.
second time:
she'd been using iphoto to organize all of her pictures (many thousands of them)
fired up iphoto one morning, and all of her shit was gone, it was like she'd never used iphoto in the first place.
no sign of the monolithic 'iphoto store' file, or anything. no original pictures. gone.
there are two things my wife loves, pictures and music, and it systematically fucked her entire collection without warning. unfortunately many of these items had not been backed up. these are just my observations, i don't know why it would do these things, and i don't care. i no longer trust that peice of shit operating system or any of its devices, and i use incremental backups of her entire laptop using rsync now (not time machine, i don't trust it either)
I could never quite comprehend the shear insanity of designing your software to delete the user's files with no warning or confirmation. I think it's unique, even among Apple apps that love to manage your stuff for you.
const int one = 65536; (Silvermoon, Texture.cs)
SJW, n: "Someone I don't like, and by the way I'm a fuckwit" - AC
You've got it mostly right, but not completely right. Any sort of digital manipulation introduces noise into the audio stream due to the nature of floating-point math. The more accuracy (or "headroom") you have in both domains (frequency and amplitude), the closer the final mix will be to the original waveform. That's why many studio engineers use 24bit / 192khz source material... not because you can actually hear any difference in the original source, but because it preserves the pristine quality of the source longer as it moves through arbitrary numbers of digital filters and mixing stages.
It's moderately useful for masters as you stated, because those masters can still be digitally manipulated without introducing any additional audible artifact. As indicated by vel-ex-tech, you actually do need additional samples for clean pitch or tempo shifting, but I wouldn't call that a typical "consumer" application.
Of course, for consumer tracks (simple listening), it's utterly pointless, much like the ridiculous Pono player. No one can actually hear the difference, and nearly all reputable double-blind studies confirm this. It's just a "bigger numbers are better" fetish among audiophiles or people who don't understand the science of audio. And frankly, the "loudness wars" have actually decreased the require dynamic range for most consumer music by pumping up the volume and compressing the shit out of it.
Irony: Agile development has too much intertia to be abandoned now.
any application
Unless you're a DJ. After shifting that track an octave and a half down and slowing that other track down about 50 bpm or so to match, you're really going to wish you had those extra samples. This is an exercise you can do on your own with free software like Audacity. Get a piece of music. Two exercises. 1.) Lower its pitch without affecting its tempo by at least an octave. 2.) Lower its tempo without affecting its pitch by at least 30 bpm. Both of these exercises should give you a hint what will happen if lower by more. It will sound like shit. You will quickly see that those extra samples you think are only for "audiophiles" really do come in handy.
Now, what's truly confusing to me about your posts is that you seem to believe that you need more than 16 bits to adjust volume. If boosting the volume on something with 16 bits makes it sound a worn cassette that's a recording of a recording of a recording, what you're running into is compression. The way a lot of shit gets mastered these days, it's well, shit. You're not getting 16 bits of loudness. You're only getting 5 or 6 bits of loudness at most. At that point, using 24-bit for playback is a soggy band-aid because that waveform that only has 5 bits of loudness with 16-bit is only going to have, what, 7 or 8 bits of loudness in 24-bit.
Please educate yourself. For listening, a 16-bit 44.1khz recording is more than enough. People do more things with recordings than just listen to them in their room or work with them in a big corporate production studio. Well, who knows. You've probably never met a modern live DJ at a bar or rave. They do more than announce the weather and the next few songs like the ones the radio. I doubt an actual DJ would have any reason to wind up in your mom's basement.
Switched to plex and never looked back. Works streamed to your cell, and then on to a bt stereo in the car also. Joy. Ripping though, I still use itunes, and then move the media to the plex server.
H.
Yes for Nyquist but no in the real world. Nyquist theorem assumes that you have a theoretically perfect ADC and DAC stage with theoretically perfectly filtering. That is not the case. In reality in order to properly filter the HF image post conversion you would need an incredibly sharp impossible to achieve filter just above the audible range. While filtering in the frequency domain is easy enough, doing so sufficiently to avoid distortion but without phase shifts below 20khz is impossible in practice.
The simple solution, increase the sample frequency to give you more headroom. You can do this in the recording stage or you can do this in the digital domain. That's why pretty much every piece of equipment has an oversampling DAC and then does the final conversion at 192khz even if it doesn't take source material with those specs. It makes the following circuit design simpler and does actually make a measurable improvement (arguable if it's audible) within the range of human hearing with no gold plating required.
Exactly. It's like RAW images from your camera.
You don't see any difference between RAW and JPGs until you have to save blown highlights, change contrast/shadows/white balance.
No iTunes actually deletes the files. If you want them again you are supposed to redownload them from apple music. But it will also "match" your song to its internal database, and if there is a match then the file doesn't even make it to the cloud, they just serve you the version they have. On top of it, if you stop paying for apple music, you loose access to the files you stored in the cloud.
But, iTunes Match is a completely Opt-In service. Not that losing ones' data is ever cool...
One of the links leads to a story where a guy complains about Apple Music (not mere iTunes). When you subscribe, they actually remove files from your machine and store them in "their cloud", playing you back whichever song they deem a match for your file. And you lose it all when you stop paying.
https://blog.vellumatlanta.com...
lucm, indeed.
Several years back, I made the mistake of checking the option on my iPhone to "sync subscriptions" for podcasts. I later deleted a podcast from my phone, as I was all caught up with it and needed the screen real estate to make looking for other podcasts easier. When I next synced with iTunes, it deleted all past episodes I had stored on my Mac -- years worth of the science fiction podcast Escape Pod. No warning, no recycle bin. Gone. And I wasn't backing up audio files via Time Machine because of their size. Hundreds of files gone. I posted on the Apple forums about what a huge hole this was (seriously, deleting hundreds of user files without warning with no undo is a major UX/UI failing). Never any response from Apple. I'm still angry about it. Even if I were to download all those files again (and since they're no longer in the iTunes feed, I would literally have to grab each one from the website, one by one, and then import them into iTunes, which would probably take days of clicking...), I would still have lost the metadata (play dates, counts). Plus there would be considerable bandwidth for both the Escape Pod site and me, in this age of looming data caps...
One of the links leads to a story where a guy complains about Apple Music (not mere iTunes). When you subscribe, they actually remove files from your machine and store them in "their cloud", playing you back whichever song they deem a match for your file. And you lose it all when you stop paying.
https://blog.vellumatlanta.com...
This is fast becoming an urban legend. But it just isn't true. Neither iTunes Music nor iTunes Match (two entirely separate services) delete ANYTHING from your computer/iPhone/iPad.
I subscribe to iTunes Music, and it hasn't deleted a single byte of data from my work Win 7 laptop, my MacBook Pro, my iPhone nor my iPad.
Pro tip: Don't try to act smart if you don't know what you're talking about :-)
The decibel is a logarithmic unit. When I say a dynamic range of 96 dB, that is the different between the noise floor and a full-scale signal. 96 dB of difference is simple a relative value. Talking about dynamic range in decibel-milliwatts as you're doing, makes absolutely no sense at all. DBm is a measurement of absolute power, so you can't use it to talk about relative dynamic range.
Quantization error exists everywhere, even in analog media. But unless you're editing the audio, 16 bit audio puts the noise floor due to quantization around -96 dBFS, which is more than good enough and very very small indeed. A -96 dBFS noise floor is equivalent to Total Harmonic Distortion less than 0.002%. In other words, completely inaudble. The inherent noise from devices such as microphones is a lot higher and exists in all recordings of non-artificial music. And no, it doesn't turn a sine wave into a stairstep, it just makes it "fuzzy". You can see this with a sufficiently good oscilloscope.
You're either delusional, or you've completely forgotten everything you should have learned about analog and digital signal theory.
Eat the rich.