You are here

What is so special about MFiT format?

For everything after the recording stage: hardware/software and how you use it.

What is so special about MFiT format?

Postby Kinh » Wed Sep 18, 2019 3:28 pm

I paid an extra $40 for the master guy to do itunes optimize and it basically sounds the same only not as loud. Seems all he did was pull the volume knob down. To my ears there's nothing different here.
Kinh
Regular
Posts: 79
Joined: Tue Feb 17, 2015 1:00 am

Re: What is so special about MFiT format?

Postby blinddrew » Wed Sep 18, 2019 3:44 pm

Have you checked the lufs on both versions? Ideally using something like the Loudness Penalty website?
User avatar
blinddrew
Jedi Poster
Posts: 8102
Joined: Sun Jul 05, 2015 12:00 am
Location: York
Ignore the post count, I have no idea what I'm doing...

Re: What is so special about MFiT format?

Postby Hugh Robjohns » Wed Sep 18, 2019 4:11 pm

Kinh wrote:To my ears there's nothing different here.

Do you really want your mix to sound different on different platforms?

The aim of mastering is to optimise the material for the different platforms, without changing the sound any more than absolutely necessary.

iTunes with SoundCheck is a loudness-normalised platform, so the average volume will need to be lower than for a peak-normalised format like CD (or many download sites) -- hence your comment about 'pulling the fader down' is probably valid to some extent. Potentially, though, the track could also have more dynamic range and a higher peak-average ratio than a peak-normalised version, depending on the genre and the mastering instructions.

There are also various technical aspects involved in MFit to optimise the material for the AAC coding etc... It does involve a separate processing 'pass' with different settings and requirements to other release formats.

H
User avatar
Hugh Robjohns
Moderator
Posts: 24543
Joined: Fri Jul 25, 2003 12:00 am
Location: Worcestershire, UK
Technical Editor, Sound On Sound

Re: What is so special about MFiT format?

Postby CS70 » Wed Sep 18, 2019 5:30 pm

Kinh wrote:I paid an extra $40 for the master guy to do itunes optimize and it basically sounds the same only not as loud. Seems all he did was pull the volume knob down. To my ears there's nothing different here.

Ahah yeah, under the disclaimer that I'm not gonna pay Apple for the privilege of knowing what "Mastered for iTunes" precisely means according to them, there's very little that can be different.

Upon receiving your master, Apple will subject it to lossy compression to reduce the file size (that's nothing to with audio compression, of course). Which type of compression matters little, as there's no magic: the quality of the results will always depend essentially on the result's bitrate (the more,the better, but the larger the file) and the amount of artifacts that may occur when reconstructing the original waveform since - unlike normal sampling - we actually lose information when we size-compress in a lossy way.

A third factor in Apple's compression - the psychoacoustic model it employs - might be a little different than mp3 but having taken Fraunhofer's guys three decades to get to a good result (and let's just say they were a little more genius than your average Apple genius ;-) ), I wouldn't expect it to be that much different.

So anything that is mastered to survive lossy compression, while maintaining good quality, must peak quite lower than 0dbFS (to ensure that the waveform reconstructed from less information does not end up clipping), often a very audible -1dbFS is a good target; and be encoded keeping a substantial bitrate (for mp3s... 320Kbps seems to be the magic number, where the resulting file is smaller but very hard to distinguish from the full PCM).

Perhaps one can go a little less heavy handed with the limiter, or simply master for loudness-normalization rather than peak-normalization.

That means that the result will, indeed, sound a little lower (or quite a bit lower if it's loudness-normalized). In absolute terms, of course - until you reach for the volume knob.

That's it. Apple are indeed good at making money out of thin air :D
User avatar
CS70
Jedi Poster
Posts: 4227
Joined: Mon Nov 26, 2012 1:00 am
Location: Oslo, Norway
Silver Spoon - Check out our latest video  and the FB page

Re: What is so special about MFiT format?

Postby James Perrett » Wed Sep 18, 2019 8:34 pm

CS70 wrote:A third factor in Apple's compression - the psychoacoustic model it employs - might be a little different than mp3 but having taken Fraunhofer's guys three decades to get to a good result (and let's just say they were a little more genius than your average Apple genius ;-) ), I wouldn't expect it to be that much different.

Apple use AAC encoding instead of mp3. AAC is what mp3 would have been if Fraunhofer had been given a few more years to develop it. From things I read at the time I understand that mp3 and AAC development were closely related and Fraunhofer is certainly involved in AAC licensing. AAC claims to give the same sound quality as mp3 with a 25% reduction in file size.
User avatar
James Perrett
Moderator
Posts: 8429
Joined: Mon Sep 10, 2001 12:00 am
Location: The wilds of Hampshire
JRP Music - Audio Mastering and Restoration. JRP Music Facebook Page

Re: What is so special about MFiT format?

Postby CS70 » Wed Sep 18, 2019 8:46 pm

James Perrett wrote:
CS70 wrote:A third factor in Apple's compression - the psychoacoustic model it employs - might be a little different than mp3 but having taken Fraunhofer's guys three decades to get to a good result (and let's just say they were a little more genius than your average Apple genius ;-) ), I wouldn't expect it to be that much different.

Apple use AAC encoding instead of mp3. AAC is what mp3 would have been if Fraunhofer had been given a few more years to develop it. From things I read at the time I understand that mp3 and AAC development were closely related and Fraunhofer is certainly involved in AAC licensing. AAC claims to give the same sound quality as mp3 with a 25% reduction in file size.

Yes, that was why I wrote that sentence.

Apple claims a lot of things and people swallow them (with Apple swallowing their money), but the reality is that a 320Kpbs encoded mp3 is perceptually indistinguishable from the original PCM, and it's really hard to identify a 240Kbps (or even lower) in a blind test, especially if the encoder, like LAME (and most likely AAC), uses good noise shaping. And you can't get blacker than black :)

The basis for compression is gotta be the same (masking, progressive windowing, psychoacoustic model leading to a preference to keep midrange info etc): it's not a radical new method, not even a radical new tech. Besides, MP3 is designed in a way to evolve without breaking - look at LAME over the last 10 years.

Sure space requirements may well be true to a point (and I would bet my car there's lots of ifs - like comparing different implementations based on huffman coding like ZIP and RAR for example).. but disk space is cheap as chips nowadays and who knows if Spotfiy et al. use the same space as Apple Music?

I have a terabyte on a SSD disk that fits in my shirt pocket and costed me less than a dinner out (in Oslo. Oslo is expensive)

Say you need a few petabytes to contain all music that exists - they're not in a single place but over a very large distributed, redundant network, so costs are linear (no specialized controllers, etc). A Petarack is about 375KUSD. 25% more is pocket money..
User avatar
CS70
Jedi Poster
Posts: 4227
Joined: Mon Nov 26, 2012 1:00 am
Location: Oslo, Norway
Silver Spoon - Check out our latest video  and the FB page

Re: What is so special about MFiT format?

Postby johnny h » Sun Sep 29, 2019 1:18 pm

CS70 wrote:
James Perrett wrote:
CS70 wrote:A third factor in Apple's compression - the psychoacoustic model it employs - might be a little different than mp3 but having taken Fraunhofer's guys three decades to get to a good result (and let's just say they were a little more genius than your average Apple genius ;-) ), I wouldn't expect it to be that much different.

Apple use AAC encoding instead of mp3. AAC is what mp3 would have been if Fraunhofer had been given a few more years to develop it. From things I read at the time I understand that mp3 and AAC development were closely related and Fraunhofer is certainly involved in AAC licensing. AAC claims to give the same sound quality as mp3 with a 25% reduction in file size.

Yes, that was why I wrote that sentence.

Apple claims a lot of things and people swallow them (with Apple swallowing their money), but the reality is that a 320Kpbs encoded mp3 is perceptually indistinguishable from the original PCM, and it's really hard to identify a 240Kbps (or even lower) in a blind test, especially if the encoder, like LAME (and most likely AAC), uses good noise shaping. And you can't get blacker than black :)
Sorry, don't want to get all "golden ears" on you, but its quite easy to tell the difference between MP3s and WAVs for certain kinds of music. You lose a lot of the dynamics and transients, even in full 320kbs MP3s.
johnny h
Frequent Poster (Level2)
Posts: 3912
Joined: Mon Jul 24, 2006 12:00 am

Re: What is so special about MFiT format?

Postby Hugh Robjohns » Sun Sep 29, 2019 9:26 pm

johnny h wrote:Sorry, don't want to get all "golden ears" on you, but its quite easy to tell the difference between MP3s and WAVs for certain kinds of music. You lose a lot of the dynamics and transients, even in full 320kbs MP3s.

The audibility of coding artefacts follows a fairly conventional Gaussian distribution, so statistically while the vast majority of people can't tell the difference between wav and mp3 at, say 320kbps (and some not even at much lower bit rates), there will inevitably be some who can still reliably hear the artefacts even in a high bit-rate coding. These people might think they have golden ears... :-D

However, the audibility of coding artefacts can also be increased dramatically in a number of different ways.

For example, it is certainly possible to improve someone's ability to hear coding artefacts through ear training.

It is also the case that different MP3 codecs employ slightly different algorithms with different models of the notional hearing thresholds, and so some challenging audio material can interact with certain coder designs to reveal artefcats that might not manifest in other codec designs (or with different codec settings).

Perhaps the easiest way to reveal artefacts, though, is to replay mp3s at significantly elevated volumes, since doing so completely trashes all the coding assumptions derived from the codec's hearing hreshold models. Hence the reason MP3 files generally sound so obvious when played over chunky PAs at elevated levels....
User avatar
Hugh Robjohns
Moderator
Posts: 24543
Joined: Fri Jul 25, 2003 12:00 am
Location: Worcestershire, UK
Technical Editor, Sound On Sound


Who is online

Users browsing this forum: beakerxvi