anyone considering/doing this?  Here's a clip from a mastering house that does iTunes mastering...
 
With or without Mastered for iTunes in mind, if you are planning for a release with wide-spread digital distribution, I recommend the following practices during the recording, mixing and mastering stages:

• Record the music in 24-bit 44.1kHz sample rate, or higher, using conservative levels (RMS level of -24dBFS is ideal, peaks should be no higher than -6dBFS)
• Mix to a 24-bit 96kHz sample rate file, using a mixing platform that can handle 32 or 64-bit floating point 96kHz or higher internal architecture, keeping levels throughout the chain to a conservative level (RMS level of -24dBFS is also ideal, here)
• Master to a 24-bit 44.1kHz sample rate WAV file (or higher), using a mastering platform that can handle 32 or 64-bit floating point 96kHz or higher internal architecture (see note, below)
• Master to a 16-bit 44.1kHz WAV file, specially optimized for that bit depth and sample rate
• Check both masters for any clipping, and make any necessary adjustments
• Convert both masters to iTunes plus format, and check for clipping again, then go back and make any necessary adjustments
• Submit the 24-bit WAV master to an MFiT aggregator to submit to iTunes only, and submit the 16-bit WAV version as a separate release for other distribution companies (discard the temporary AAC copies you made in the steps above)

I can't imagine mastering my stuff to RMS -24dBfs.  What level do they push it from there?  It looks like it can only be done also in OSX.

When you submit a file (usually an uncompressed wave) for iTunes distribution, it gets (re)encoded and normalized.
 
What you're accomplishing by mastering to such a conservative standard is making sure that iTunes only raises your volume rather than lowering it. A very hot master suffers when lowered in volume, sounding dull and wimpy. But raising a dynamic master doesn't hurt anything. You're also avoiding artifacts and overs that appear when encoding a hot file using a lossy encoder such as iTunes' AAC (or MP3, for that matter).
 
The only downside to "mastered for iTunes" is that if you also want to put your songs onto a CD, you'll probably want to create separate masters for each. You're going to want a hotter master for the CD, and there is no penalty for doing so as long as you don't go overboard. (At least, not until somebody rips MP3s from the CD, in which case they'll get better results if the CD is a little on the conservative side.) 
 
Then there's the matter of SoundCloud and YouTube. YouTube's normalization standard is even more conservative than iTunes. SoundCloud accepts uncompressed files, but streams everything at 128 kb/s, so that's a whole 'nother can 'o worms. Even though it doesn't normalize, you still have to consider the issue of MP3 artifacts that are exacerbated by loud mastering.
 
Of course, if your tune finds its way to commercial radio or television, there's a whole different standard for that. Broadcasting does the most drastic things to your music, and the -24 (US) or -23 LUFS (rest of the world) standards are strictly enforced so you have no choice but to follow them.
 
Bottom line is that each distribution medium warrants its own mastering standards. However, the overarching message is keep it conservative regardless of the final destination. Limit peaks to no more than -1 dB, which works fine for AAC, MP3, and CDs. For overall volume, you need a LUFS meter and shoot for an average of around -14 to -16 LUFS, a nice compromise that will work for YouTube, SoundCloud and iTunes. -11 LUFS seems to be a nice target for CDs, but that may be too conservative for some genres and too aggressive for others. 

Don't know where that -24dB RMS or -6dB peak stuff is coming from. Maybe that particular mastering facility doesn't want to be handling audio that's already been volume maximised and squished to the max by someone without either the tools or knowledge to limit properly and would rather deal with conservative levels because it's easier for them and they end up telling fewer people their mix isn't suitable for mastering even if the -6dB peak means they're throwing a chunk of dynamic range away before they start.

Apple's technological brief is here - http://images.apple.com/u...astered_for_itunes.pdf and doesn't mention it.

Apple's provided tools are indeed for OS X, but Apple's main concern seems to be helping people avoid digital and inter-sample clipping. The OS X tools basically let you do what Apple do when a file is submitted for iTunes so you can do the conversion to AAC and a clipping check yourself and confirm the result is what you expect.

i mix 24 bit files to about -22db RMS,
and about -6 to -8 for peaks.
 
that is an UNMASTERED final stereo bounce.
 
so, saying the word MASTER, DOES NOT MEAN YOU HAVE TO MAKE IT LOUD!!!
 
can you hear me now?!!
 
LOL
 
i could 'master' that file, by simply doing minor adjustments of eq, minor compressor tweaks with no extra gain supplied, and no brickwall limiting at all, apply dither, and have the file peak at the same peaks, with only a slightly reduced rms average due to the way i used the compressor......
 
but, on the other hand....
i could upload my unmastered file, according to the itunes format, and that would work just fine with their proprietary encoding,
and supposedly my unmastered file would sound just as loud as a professionally mastered file, that was super loud,
that had been dumbed down by the itunes process.
 
in fact, my unmastered file may sound WAY more dynamic than a professionally mastered file that had been mastered for volume.
 
all that said, i'd like to hear a comparison side by side,
no one has really gone there yet, as i can tell..
 
 

batsbrew
all that said, i'd like to hear a comparison side by side,
no one has really gone there yet, as i can tell..

 
Which is surprising since it's surely pretty easy to perform.
 
I just did a quick search and found absolutely nothing on the matter.
 

Sounds like a bunch of hooey to me .  It is just attempting to make the best of a bad format.  I haven't heard a "Mastered for iTunes" album that sounds any better than the original CD converted to mp3.  Different, yes.  Possibly worse, definitely.  

daniel,
i think you are missing the point..
it's to drive folks back to making dynamic music,
by having a system that makes every album play back on an even playing field.
 
this is actually what we all should be demanding.
a stop to this insane 'volume war'
 
 
if you didn't need to 'master' you mixes ( i like my mixes as they are pre mastering,, actually).....
then everyone's mixes would play back at the same volume,
and it would make more sense to NOT compress,
to NOT limit,
to have lively dynamic mixes again,
instead of a homogenous mess that is all one volume, all the time.
 
soap box.
 

if I submit it at RMS -24dBfs, and rely on iTunes to volume match it to match everything else, then I don't REALLY know what it will sound like if they push it 10 db?  
for now i'll just stick to my current ~ RMS -14dBfs and peaks limited to -.5

I recently released a single using CD Baby...to go through the motions more than anything else.  They handled all of the distribution to major streaming sites including iTunes.  I don't really know how CD Baby accomplishes this, but they ask for a 24bit 44.1kHz wave file...and a payment:)
 
The first wave I submitted I had 'mastered' to -0.3 dB I think, which is usually what I export mp3's at.  CD Baby's mp3 sounded harsh to me - which doesn't mean my 'master' didn't in the first place - but since you have a chance to review the song on their site, I resubmitted wave file with levels no more than -1.0 dB.  I don't have the ears of a producer, but I thought the second one was better.
 
Here's that tune at CD Baby:
 
http://www.cdbaby.com/cd/scottholson
 
It's on iTunes under same song title and artist name should anyone want to listen for differences.

you don't allow someone to master an already mastered file.
that ruins it.