Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Mastered for iTunes question (Read 12385 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Mastered for iTunes question

Has anyone ever compared a "Mastered for iTunes" VBR AAC file to one they made themselves from a CD or FLAC source?  I'm curious if people can detect a positive different in the tracks Apple sells.

I believe Apple is taking studio masters and converting straight down to AAC.  I wonder if this adds any marked improvement over  just converting a lossless format.

Mastered for iTunes question

Reply #1 –
Apple's software offers clipping detection, so if their encoder sucks, the mastering engineer has a tool to detect it.  Or rather, a tool to provide peace of mind that one hasn't done something terribly wrong? Has anyone found cases where the clipping detection was actually needed? Maybe there is in one of the numerous threads herein.

There could of course very well be a different master used to encode the iTunes tracks, just like there could very well be two different masters behind two CD copies of "Kind of Blue". And Apple asks that the files be encoded from a high-resolution file, meaning that you cannot rule out that someone grabs something else than what ultimately came on CD. 

Mastered for iTunes question

Reply #2 –
Quote from: Porcus on
Apple's software offers clipping detection, so if their encoder sucks, the mastering engineer has a tool to detect it.  Or rather, a tool to provide peace of mind that one hasn't done something terribly wrong? Has anyone found cases where the clipping detection was actually needed? Maybe there is in one of the numerous threads herein.

There could of course very well be a different master used to encode the iTunes tracks, just like there could very well be two different masters behind two CD copies of "Kind of Blue". And Apple asks that the files be encoded from a high-resolution file, meaning that you cannot rule out that someone grabs something else than what ultimately came on CD.


I did read that Apple wants the AAC made directly from the digital master, which is usually 20 or 24 bit, vs 16 bit for CD quality FLAC.  I just wonder if makes for a higher quality track for the consumer.

Mastered for iTunes question

Reply #3 –
Quote from: apastuszak on
I did read that Apple wants the AAC made directly from the digital master, which is usually 20 or 24 bit, vs 16 bit for CD quality FLAC.  I just wonder if makes for a higher quality track for the consumer.


Read it from the horse's mouth.  90% of it is meaningless guff, but it explains what Apple wants and expects, with their reasons why.  In a nutshell they want you to provide more than 16 bits, and more than 44.1 kHz, although they don't want you to just upsample an existing 16/44.1 recording.

I'll offer the following closing quote from the linked PDF.  It sounds good but I'm sure most people here will see right through it.
Quote
Although it’s possible to remaster from a previously mastered CD source with positive
results, in order to qualify as Mastered for iTunes remastered content must begin with a
high resolution digitization of the original analog source and must sound noticeably
superior to the previously released version. Songs and albums submitted to the iTunes
Store as remastered content will be reviewed to ensure that the sound quality shows
discernible improvement.

Mastered for iTunes question

Reply #4 –
Quote from: lithopsian on
Quote from: apastuszak on
I did read that Apple wants the AAC made directly from the digital master, which is usually 20 or 24 bit, vs 16 bit for CD quality FLAC.  I just wonder if makes for a higher quality track for the consumer.


Read it from the horse's mouth.  90% of it is meaningless guff, but it explains what Apple wants and expects, with their reasons why.  In a nutshell they want you to provide more than 16 bits, and more than 44.1 kHz, although they don't want you to just upsample an existing 16/44.1 recording.

I'll offer the following closing quote from the linked PDF.  It sounds good but I'm sure most people here will see right through it.
Quote
Although it’s possible to remaster from a previously mastered CD source with positive
results, in order to qualify as Mastered for iTunes remastered content must begin with a
high resolution digitization of the original analog source and must sound noticeably
superior to the previously released version. Songs and albums submitted to the iTunes
Store as remastered content will be reviewed to ensure that the sound quality shows
discernible improvement.



But is having a higher sample rate that 16/44.1 going to produce an AAC file that's noticeably better quality than just ripping a CD?

My ears can't tell the difference between a 24/192 FLAC bought from HDtracks and ripped to VBR 0 MP3, and taking the 24/192 FLAC, converting it to 16/44.1 FLAC and then making the MP3.  I don't know if having the file in AAC format helps the sound quality in some way when encoding from a high resolution source.

Mastered for iTunes question

Reply #5 –
Read the whole document.  They make their argument.  You decide if you believe it.

Mastered for iTunes question

Reply #6 –
If you have a Mac, grab the Master for iTunes Droplet. You can encode your own samples and compare

https://www.apple.com/au/itunes/mastered-for-itunes/

"The Mastered for iTunes droplet is a simple, stand-alone
drag-and-drop tool that automates the creation of iTunes Plus format masters, allowing you to preview of songs using the same world-class technology used to encode the entire catalogue for the iTunes Library."

Mastered for iTunes question

Reply #7 –
Quote from: apastuszak on
But is having a higher sample rate that 16/44.1 going to produce an AAC file that's noticeably better quality than just ripping a CD?
No. The difference is that the master for iTunes might be different to the CD master. If they're the same you will get the same result.
It's only audiophile if it's inconvenient.

Mastered for iTunes question

Reply #8 –
Quote from: lithopsian on
Quote from: apastuszak on
I did read that Apple wants the AAC made directly from the digital master, which is usually 20 or 24 bit, vs 16 bit for CD quality FLAC.  I just wonder if makes for a higher quality track for the consumer.


Read it from the horse's mouth.  90% of it is meaningless guff, but it explains what Apple wants and expects, with their reasons why.  In a nutshell they want you to provide more than 16 bits, and more than 44.1 kHz, although they don't want you to just upsample an existing 16/44.1 recording.

I'll offer the following closing quote from the linked PDF.  It sounds good but I'm sure most people here will see right through it.
Quote
Although it’s possible to remaster from a previously mastered CD source with positive
results, in order to qualify as Mastered for iTunes remastered content must begin with a
high resolution digitization of the original analog source and must sound noticeably
superior to the previously released version. Songs and albums submitted to the iTunes
Store as remastered content will be reviewed to ensure that the sound quality shows
discernible improvement.




There are two issues there - the possible sample rate issue that seems to have captured and derailed everybody's attention, and the issues related to masteringf a  previously mastered recording  (IMO far more serious).

There are two ways to remaster a recording - one is to start out with a previously mastered recording which is potentially already highly processed, the other is to start out with an unmastered mixdown which is potentially far closer to the original source.

Strictly speaking remastering a previously mastered recording is something that any person who has a previously distributed version of the recording can do. If the originally distributed recording was overcompressed or heavens forbid had something like reverb added, lotsa luck removing those warts.

Better yet is to start out by making a new mixdown, which is often possible.

Mastered for iTunes question

Reply #9 –
Quote from: kritip on
If you have a Mac, grab the Master for iTunes Droplet. You can encode your own samples and compare

https://www.apple.com/au/itunes/mastered-for-itunes/

"The Mastered for iTunes droplet is a simple, stand-alone
drag-and-drop tool that automates the creation of iTunes Plus format masters, allowing you to preview of songs using the same world-class technology used to encode the entire catalogue for the iTunes Library."


I have a Mac and some 24/192 FLACs.  Time to make some files and fire up foobar2000!

Mastered for iTunes question

Reply #10 –
Quote from: apastuszak on
But is having a higher sample rate that 16/44.1 going to produce an AAC file that's noticeably better quality than just ripping a CD?


Not very likely, as long as you are comparing ... uh, apples and apples:
FILE#1 --> converted to 16/44.1, done properly --> pressed to CD, without errors --> ripped from CD, without errors --> converted to 256 kb/s AAC with an encoder that doesn't totally suck
vs
FILE#1 --> converted to 256 kb/s AAC with an encoder that doesn't totally suck.

Apple does provide tools that enable engineers to check for clipping, but if clipping were to occur too often, then there would be lots of listening tests where a 256 kb/s AAC could be consistently distinguished from the original.

But since Apple does not request the 16/44.1 master, chances increase that you do not get FILE#1 but a FILE#2 which has had different processing. So if you are experimenting with this, and you do notice differences, then you should not be surprised - and it is likely to be because something else happened than the apples vs apples scenario.

Mastered for iTunes question

Reply #11 –
Quote from: Porcus on
Quote from: apastuszak on
But is having a higher sample rate that 16/44.1 going to produce an AAC file that's noticeably better quality than just ripping a CD?


Not very likely, as long as you are comparing ... uh, apples and apples:
FILE#1 --> converted to 16/44.1, done properly --> pressed to CD, without errors --> ripped from CD, without errors --> converted to 256 kb/s AAC with an encoder that doesn't totally suck
vs
FILE#1 --> converted to 256 kb/s AAC with an encoder that doesn't totally suck.

Apple does provide tools that enable engineers to check for clipping, but if clipping were to occur too often, then there would be lots of listening tests where a 256 kb/s AAC could be consistently distinguished from the original.

But since Apple does not request the 16/44.1 master, chances increase that you do not get FILE#1 but a FILE#2 which has had different processing. So if you are experimenting with this, and you do notice differences, then you should not be surprised - and it is likely to be because something else happened than the apples vs apples scenario.


Ok, here is what I did.

Got a 24/96 FLAC from HDTRACKS.COM.  I used XLD to convert the FLAC to WAV.  I then used the Mastered for iTunes to make it an AAC file.  I then took the same 24/96 and used XLD to convert it to a 44.1/16 ALAC.  I then dropped the ALAC into iTunes and used iTunes to make a 256K CBR AAC file.

Both went into Foobar2000 with the ABX comparitor.

First 3 guesses were dead on,  It seemed I was able to spot the Mastered for iTunes version.  The next 3 guesses I got wrong.

ABX is saying:

Score 3/6
Probability that you were guessing: 65.6%

So, my questions now are:

1. Is all Lossless the same?  Would there really be any difference between converting from a 96/24 FLAC to a 44.1/16 FLAC or a 44.1/16 ALAC?
2. I'm running Foobar2000 on Linux using Wine.  Is this a problem in any way?  I can't find any native Linux program to do abx comparison.
3. ABX is just testing to see if you can spot a difference between the two files.  Is there a way to tell it which one you liked better?

Mastered for iTunes question

Reply #12 –
3 out of 6 is 50%, which is the same thing as randomly guessing or flipping a coin. Any differences you thought you heard were imagined. You have to get that number down to like 5% before you can say you're really hearing a difference. And if you can't prove you can tell the difference, it doesn't make sense to want to remember which one you liked better.

If you use the same converter, and convert to the same sample rate and bit depth, then yes, all lossless is the same. So 44.1/16 FLAC = 44.1/16 ALAC, regardless of what you converted from.

I don't know what happens to the audio as it passes through the Windows APIs to Wine and on to the Linux APIs. There's a chance it gets converted to another sample rate and bit depth. But your results so far are exactly what we'd expect, so I don't believe there's a problem.

Mastered for iTunes question

Reply #13 –
1. Clearly all lossless are not the same.  Do a diff  But the digital stream that comes out from a decoder of a lossless file is the same.  WAV. ALAC, FLAC, all produce an identical stream of binary data, assuming same bitrate, bit depth, sampling rate, etc. although some other lossless formats produce a fundamentally different type of binary data to reproduce the same sound.  The differences are simply in the format that is used to stored that data in a file, type of compression (or not), tags, encapsulation.  You'll still find plenty of people that claim to hear the difference between, for example, WAV and FLAC, which maybe there was once due to shortcuts in decoder implementation, but I think you have the techniques to test that claim if you feel like it.

2. Shouldn't be a problem.  In the same way that some people think they can hear the difference between WAV and FLAC, no doubt some people claim they can hear the difference between Foobar under Wine and under Windows, but any differences would likely be gross and related to a configuration problem, for example a poor choice of resampler.

3. ABX is a type of blind testing that determines whether you can tell the difference between one file and a reference.  It does this by randomly offering you a comparison between the reference and itself, or between the test file and the reference.  Which you prefer is not really at issue in this type of test, although in most cases it is assumed that any reliably detectable difference is on the down side.  There are other types of blind listening test, some of which allow you to specify a preference.

Mastered for iTunes question

Reply #14 –
Quote from: apastuszak on
Got a 24/96 FLAC from HDTRACKS.COM.  I used XLD to convert the FLAC to WAV.  I then used the Mastered for iTunes to make it an AAC file.  I then took the same 24/96 and used XLD to convert it to a 44.1/16 ALAC.  I then dropped the ALAC into iTunes and used iTunes to make a 256K CBR AAC file.

There are two variables in this test and there should only be one. Encode the downsampled one with Mastered for iTunes too to see if you can tell a difference between encodings made from the original and a downsample of it, without any other factor potentially causing audible differences like a different encoder with different settings.

Mastered for iTunes question

Reply #15 –
Quote from: _if on
Quote from: apastuszak on
Got a 24/96 FLAC from HDTRACKS.COM.  I used XLD to convert the FLAC to WAV.  I then used the Mastered for iTunes to make it an AAC file.  I then took the same 24/96 and used XLD to convert it to a 44.1/16 ALAC.  I then dropped the ALAC into iTunes and used iTunes to make a 256K CBR AAC file.

There are two variables in this test and there should only be one. Encode the downsampled one with Mastered for iTunes too to see if you can tell a difference between encodings made from the original and a downsample of it, without any other factor potentially causing audible differences like a different encoder with different settings.


Mastered for iTunes requires an WAV or AIFF file in order to work, so that's why I went to a WAV.  I was not able to find a way to convert the 95/24 FLAC to a 44.1/16 losless with XLD in any other format other than ALAC.

I had assumed that lossless was lossless.

So, from my tests here, I have learned that, on laptop running Arch Linux and Foobar2000, through unamplified Sennheiser HD558 hadphones, I was not able to discern a difference between the song I picked.

Now, the song that I picked was old.  It was Bellbottom Blues by Derek and the Dominoes.  The song was obviously remastered from analog tape.

So, I'm going to redo the test changing the following:

1. I will attempt to maintain a full WAV workflow.
2. I will select a song that was recorded digitally and more recently
3. I'll use a PC running MS Windows to make WINE out of the equation

I'll post results tonight if my kids give me a break.

P.S:  The reason why I wanted ABX to ask which one you like better is this.  Mastered for iTunes claims to make a superior file.  Which is great.  Let's say the test was a smashing success and I did it 10 times and could tell the difference every single time.  That still leaves one question out there.  Did the Mastered for iTunes version sound better to me, or did the plain old AAC rip?

P.P.S.:  Do I really need to go through the trouble of making a CD and ripping it?  In theory, I should not lose any quality, if I make a 44.1/16 WAV, burn it to CD and rip it again.

Mastered for iTunes question

Reply #16 –
No point burning to CD.  You could theoretically lose a few bits either burning or ripping if you were unlucky or careless, but you could digitally compare the original and ripped versions to see if that was the case.  Any loss should be considered an error in the process, not in the format.

Worry about which is better when (if) you find any difference between the lossy versions from the different masters.  You might also consider a full ABX of the AAC file against the (any) lossless original.  AAC at that level is so close to transparent that it is unlikely you'll be able to reliably spot the difference.  I can ABX CBR MP3 (MP3 CBR is pretty bad, significantly worse than VBR) at 256 kbps, but not on every track and AAC is better in just about every respect.  Of course if you can't ABX the lossy version, the rest of the test becomes either pointless (if you can't tell the difference) or quite remarkable (if you can!).

Mastered for iTunes question

Reply #17 –
Quote from: lithopsian on
No point burning to CD.  You could theoretically lose a few bits either burning or ripping if you were unlucky or careless, but you could digitally compare the original and ripped versions to see if that was the case.  Any loss should be considered an error in the process, not in the format.

Worry about which is better when (if) you find any difference between the lossy versions from the different masters.  You might also consider a full ABX of the AAC file against the (any) lossless original.  AAC at that level is so close to transparent that it is unlikely you'll be able to reliably spot the difference.  I can ABX CBR MP3 (MP3 CBR is pretty bad, significantly worse than VBR) at 256 kbps, but not on every track and AAC is better in just about every respect.  Of course if you can't ABX the lossy version, the rest of the test becomes either pointless (if you can't tell the difference) or quite remarkable (if you can!).


I've always wanted to convert my FLACs to 500kbps ogg files, but my wife's iPhone can't do ogg files, which is annoying.  Right now, I am ripping everything to VBR -0 MP3s using the latest version of LAME.

I'll need to ABX some ogg vs mp3 vs aac.  I assume WMA and MP3pro are dead at this point.

Mastered for iTunes question

Reply #18 –
Quote from: apastuszak on
Quote from: lithopsian on
No point burning to CD.  You could theoretically lose a few bits either burning or ripping if you were unlucky or careless, but you could digitally compare the original and ripped versions to see if that was the case.  Any loss should be considered an error in the process, not in the format.

Worry about which is better when (if) you find any difference between the lossy versions from the different masters.  You might also consider a full ABX of the AAC file against the (any) lossless original.  AAC at that level is so close to transparent that it is unlikely you'll be able to reliably spot the difference.  I can ABX CBR MP3 (MP3 CBR is pretty bad, significantly worse than VBR) at 256 kbps, but not on every track and AAC is better in just about every respect.  Of course if you can't ABX the lossy version, the rest of the test becomes either pointless (if you can't tell the difference) or quite remarkable (if you can!).


I've always wanted to convert my FLACs to 500kbps ogg files, but my wife's iPhone can't do ogg files, which is annoying.  Right now, I am ripping everything to VBR -0 MP3s using the latest version of LAME.

I'll need to ABX some ogg vs mp3 vs aac.  I assume WMA and MP3pro are dead at this point.


Ok, 96/24 FLAC encoded to a VBR AAC file using QAAC.  Seems I could not tell the difference:

foo_abx 1.3.4 report
foobar2000 v1.3.3
2014/10/08 18:35:08

File A: C:\Users\Redirection\xxxxxxxx\Music\Maroon 5\01 - Maps.flac
File B: C:\Users\Redirection\xxxxxxxx\Music\Maroon 5\01 - Maps.m4a

18:35:08 : Test started.
18:36:20 : 01/01  50.0%
18:37:40 : 01/02  75.0%
18:38:36 : 02/03  50.0%
18:40:10 : 02/04  68.8%
18:41:19 : 03/05  50.0%
18:42:27 : 03/06  65.6%
18:43:11 : 03/07  77.3%
18:43:34 : 03/08  85.5%
18:44:32 : 03/09  91.0%
18:45:07 : 03/10  94.5%
18:45:13 : Test finished.

----------
Total: 3/10 (94.5%)

I gave up after 10 tries.

Mastered for iTunes question

Reply #19 –
Quote from: apastuszak on
Quote from: _if on
Quote from: apastuszak on
Got a 24/96 FLAC from HDTRACKS.COM.  I used XLD to convert the FLAC to WAV.  I then used the Mastered for iTunes to make it an AAC file.  I then took the same 24/96 and used XLD to convert it to a 44.1/16 ALAC.  I then dropped the ALAC into iTunes and used iTunes to make a 256K CBR AAC file.

There are two variables in this test and there should only be one. Encode the downsampled one with Mastered for iTunes too to see if you can tell a difference between encodings made from the original and a downsample of it, without any other factor potentially causing audible differences like a different encoder with different settings.


Mastered for iTunes requires an WAV or AIFF file in order to work, so that's why I went to a WAV.  I was not able to find a way to convert the 95/24 FLAC to a 44.1/16 losless with XLD in any other format other than ALAC.

I had assumed that lossless was lossless.

Your assumption is right about lossless. What I was referring to is that you said you encoded the downsampled file to AAC with iTunes instead of the Mastered for iTunes application which you used on the original. Both should be done with the same encoder to remove it as a variable. You can use foobar to convert to 44.1/16 WAV if you need to. Just put the resampler in the processing selector when you convert.

Mastered for iTunes question

Reply #20 –
Quote from: _if on
Quote from: apastuszak on
Quote from: _if on
Quote from: apastuszak on
Got a 24/96 FLAC from HDTRACKS.COM.  I used XLD to convert the FLAC to WAV.  I then used the Mastered for iTunes to make it an AAC file.  I then took the same 24/96 and used XLD to convert it to a 44.1/16 ALAC.  I then dropped the ALAC into iTunes and used iTunes to make a 256K CBR AAC file.

There are two variables in this test and there should only be one. Encode the downsampled one with Mastered for iTunes too to see if you can tell a difference between encodings made from the original and a downsample of it, without any other factor potentially causing audible differences like a different encoder with different settings.


Mastered for iTunes requires an WAV or AIFF file in order to work, so that's why I went to a WAV.  I was not able to find a way to convert the 95/24 FLAC to a 44.1/16 losless with XLD in any other format other than ALAC.

I had assumed that lossless was lossless.

Your assumption is right about lossless. What I was referring to is that you said you encoded the downsampled file to AAC with iTunes instead of the Mastered for iTunes application which you used on the original. Both should be done with the same encoder to remove it as a variable. You can use foobar to convert to 44.1/16 WAV if you need to. Just put the resampler in the processing selector when you convert.


If I took a 96/24 FLAC and made it a vbr aac and couldn't tell the difference, I don't expect the Mastered for iTunes droplet to do anything for me.

Mastered for iTunes question

Reply #21 –
Oh yeah, I wouldn't expect it to. Just if you're doing a test you should only have one variable changed so that you can know if you did detect a difference that it's due to that variable rather than it being ambiguous which of the two caused it. I have no doubt Mastered for iTunes is little more than marketing and that the only possible advantages it could hold, assuming the encoder uses the same settings as the regular iTunes content, is not having the dithering applied from the conversion to 16-bit, which isn't going to be audible anyway, and the possibility of it using a resampler that is superior to some other resampler out in the wild which could have been used, the effects of which are hardly likely to be audible anyway. They encourage good practices in the instructions, but it's stuff that hopefully all audio engineers would know anyway, although I have no doubt there is a ridiculous number who don't.

Mastered for iTunes question

Reply #22 –
Quote from: _if on
Oh yeah, I wouldn't expect it to. Just if you're doing a test you should only have one variable changed so that you can know if you did detect a difference that it's due to that variable rather than it being ambiguous which of the two caused it. I have no doubt Mastered for iTunes is little more than marketing and that the only possible advantages it could hold, assuming the encoder uses the same settings as the regular iTunes content, is not having the dithering applied from the conversion to 16-bit, which isn't going to be audible anyway, and the possibility of it using a resampler that is superior to some other resampler out in the wild which could have been used, the effects of which are hardly likely to be audible anyway. They encourage good practices in the instructions, but it's stuff that hopefully all audio engineers would know anyway, although I have no doubt there is a ridiculous number who don't.


At this point, an audio snob would point out that I need to use a real DAC and headphone amp to truly appreciate the difference.  It's refreshing to have people ground in the real world of audio here.
 

Mastered for iTunes question

Reply #23 –
Quote from: apastuszak on
1. Is all Lossless the same?  Would there really be any difference between converting from a 96/24 FLAC to a 44.1/16 FLAC or a 44.1/16 ALAC?


The two latter would be the same, and different from the former. The signals are different: you need to throw away bits to get from 96/24 to 44.1/16.

However if you store a 44.1/16 CD rip in a 44.1/24 FLAC, it will be the same as in a 44.1/16 FLAC. The last 8 bits will only be padding with zero, just like the number 1.234 is the same as 1.2340000.


And, just in case you should ever want to store floating-point files - often used in PC-software "studio mixing desks" for working with recordings - you cannot store them as FLAC (or ALAC I think)  without altering the content, as they do not support floating-point. (WavPack does.) But that is extremely uncommon in the market - nobody distributes floating-point to the end-user except some artists who without a second thought simply upload the file they are working on. That never happened to me until this year though.