Why are masters so different?
Sep 9, 2016 at 6:27 PM Thread Starter Post #1 of 132

miceblue

Headphoneus Supremus
Joined
Feb 23, 2011
Posts
16,391
Likes
2,847
So the sampling rate/bit-depth war has been raging and most would agree that 16/44.1 is sufficient for human hearing.

"High-resolution" music websites tend to distribute different masters of songs.

Vinyl recordings tend to be different masters of songs.

SACDs often contain different masters of songs.



Why can't we just have good 16/44.1 masters all around? I can convert my 24/96 library to 16/44.1 and it sounds exactly the same but when I rip my CDs the master is really terrible a lot of the time. Why don't mastering engineers just distribute their better-mastered music on the CD platform?

I'm just really dumbfounded by the whole concept.
 
Sep 9, 2016 at 7:49 PM Post #3 of 132

castleofargh

Sound Science Forum Moderator
Joined
Jul 2, 2011
Posts
9,664
Likes
5,055
my opinion is that the guys leading the audio industry at most levels are fundamentally Evil corp. they have from the start abused the artists with barely legal contracts. treat audio engineers like factory slaves when they're just as much artists than the band itself. treat consumers like idiots(why not? it works). a few of the industry's achievements so far have been to turn some artists lives into nightmares. to spend millions trying to bribe politicians into breaking down internet instead of understanding that the world had changed and if they weren't going to provide online music, we would get it our own way. to consistently ruin good albums with messy remasters only for the sake of keeping legal rights without a care in the world for the music. to pretend like a new useless format is a revolution only to try and dominate the market to make money from a support in addition to make money from the music. and of course also as a mean to add new DRM and ways to make the consumer's life a nightmare: let's take money from computer/tablet/phones/DAPs selling the rights to use our formats, but let's do everything we can to stop the consumer from making proper use of those devices for music with stuff like SACD.
oh and let's make CDs with no dynamic so that they're radio ready because that's what matters, advertising!
let's claim that file resolution and formats are the proper mean to estimate sound quality, just so that we can sell the same stuff for more money, and from time to time, sell low res music in highres containers for idiots who won't know better and pay high money anyway.
 
 
clearly everything done in audio is done for the love of good sound, and out of concern for the well being of the consumer.
rolleyes.gif

 
 
obviously I would also like nothing more than to have all masters available in all formats with a clear documentation about the origin of a master. in my case, I would actually buy more and the industry is missing out on my money because I'd rather die than pay for stuff like SACD. I only very very reluctantly pay for a few highres masters when I really can't find a satisfying CD versions, but I feel dirty for contributing to milking consumers with a dishonest system. plus it's more work for me as I then need to convert the file into redbook(I'm not keeping 1giga files on my computer just for fun). so when I buy highres, I hate the audio industry, and I hate myself. so much fun!  if everything was available in 44 or 48khz I would buy significantly more music. I might not be the average joe in that respect, but I really would buy more. why would I buy the old albums I've lost along the years from moving, being robbed, having GF leaving with my shiit, or simply having CDs that can't play anymore, when I know the releases available nowadays won't have the sound I fell in love with 10 or 20 years ago? been there, done that, now I keep my money and listen to the old stuff I still own like an old guy.
 
since the Napster period, I've felt like the audio industry was lead by old stupid dinosaurs who enjoyed pissing people off and shooting themselves in the foot.
 
Sep 9, 2016 at 9:05 PM Post #4 of 132

icebear

Headphoneus Supremus
Joined
Jan 9, 2014
Posts
1,640
Likes
360
.... clearly everything done in audio is done for the love of good sound, and out of concern for the well being of the consumer.
rolleyes.gif

 
 
obviously I would also like nothing more than to have all masters available in all formats with a clear documentation about the origin of a master. in my case, I would actually buy more and the industry is missing out on my money because I'd rather die than pay for stuff like SACD. ......

 Actually after the SACD has practically gone the way of the Dodo, you can find these often in the bargain aisles on Amazon, at least for my preferred music genres jazz and classic. The physical SACD cost about 1/4 of high rez. download files especially in case of non-hybrid version, I jumped really late on a SACD player and I love it.
biggrin.gif

 
And for the music industry making money over and over again on the same music content ... OK the game obviously is called don't give away all your crown jewels at once, only show as much bling to attract new sales. Once turnover has ebbed, you reveal some more. Never go all the way in the first round, it's a bit like striptease.
rolleyes.gif

 
Sep 10, 2016 at 1:33 AM Post #5 of 132

dprimary

100+ Head-Fier
Joined
Jan 11, 2015
Posts
363
Likes
122
The job of a mastering engineer is maximize the fidelity of every format the recording it to be released on, be it 96/24 downloads to 8 bit greeting cards and ring tones. If the album is not already in sequence they will put the songs in order, they might adjust the time between songs, clean up the beginnings and ends of the songs, balance the levels between songs, balance the eq between songs, in general make many small adjustment that recording engineer and the artist may have overlooked when they mixed each song. When a song is mixed it all about that song, how it fits in the album is often overlooked.
 
Then of course after all those careful adjustments the label makes them compress every last dB of dynamic range out the recording in some twisted need to seem louder then the next recording. After all everyone knows if you take crappy music and just make it louder it magically becomes better. Ok there might be some truth to that in a live show and the crowd is drunk.
 
In the 80's to early 2000's the mastering engineer would take that master and pretty much do a direct transfer to the CD master both would be 16/44.1 then they would process the master to cut LP, 12" or 7" masters. Then if you were duplicating cassettes that would have different master for its limitations. Now we MP3, AAC, FLAC each with it's own master. Some are direct transfers. Mastered for iTunes is simply the mastering engineer confirming the master they send to apple will encode properly. Apple will reject anything with inter-sample overages, which can happen from excessive dynamic range compression.
 
Sep 10, 2016 at 2:10 AM Post #6 of 132

miceblue

Headphoneus Supremus
Joined
Feb 23, 2011
Posts
16,391
Likes
2,847
The job of a mastering engineer is maximize the fidelity of every format the recording it to be released on, be it 96/24 downloads to 8 bit greeting cards and ring tones. If the album is not already in sequence they will put the songs in order, they might adjust the time between songs, clean up the beginnings and ends of the songs, balance the levels between songs, balance the eq between songs, in general make many small adjustment that recording engineer and the artist may have overlooked when they mixed each song. When a song is mixed it all about that song, how it fits in the album is often overlooked.

Then of course after all those careful adjustments the label makes them compress every last dB of dynamic range out the recording in some twisted need to seem louder then the next recording. After all everyone knows if you take crappy music and just make it louder it magically becomes better. Ok there might be some truth to that in a live show and the crowd is drunk.

In the 80's to early 2000's the mastering engineer would take that master and pretty much do a direct transfer to the CD master both would be 16/44.1 then they would process the master to cut LP, 12" or 7" masters. Then if you were duplicating cassettes that would have different master for its limitations. Now we MP3, AAC, FLAC each with it's own master. Some are direct transfers. Mastered for iTunes is simply the mastering engineer confirming the master they send to apple will encode properly. Apple will reject anything with inter-sample overages, which can happen from excessive dynamic range compression.

Interesting to know. Mastered for iTunes stuff is actually pretty good from what I've purchased.

Still though, why would they make 24/96 masters sound way better if they can just downsample that to 16/44.1 for CD distribution?

...apart from the "high-resolution music having XX times the amount of data as a CD" marketing, which really means nothing for the sound quality.
 
Sep 10, 2016 at 7:31 AM Post #7 of 132

castleofargh

Sound Science Forum Moderator
Joined
Jul 2, 2011
Posts
9,664
Likes
5,055
mastered for itune is nothing special and doesn't deal with the usual loudness war crippled album. they basically suggest to apply about -1db gain as headroom so that inter sample clipping doesn't occur too often. it's IMO a good thing overall and I don't get why all lossy encoders don't include a little margin like that for such a well known problem? but just using replay gain while paying attention to the most dynamic albums(or using a quieter base value) can also avoids intersample clipping. "remastered for myself". ^_^
and very very few DACs keep a little margin just for that too. the answers have always been around, it's only that not enough people care. if someone mistakes clipping for the bad sound of lossy, all the better for highres marketing. and big SNR and crazy high sample rates are what sell a DAC, not that it won't clip your music
rolleyes.gif
.
 
Sep 11, 2016 at 3:31 PM Post #8 of 132

Lex2

100+ Head-Fier
Joined
Jul 20, 2008
Posts
340
Likes
25
  mastered for itune is nothing special and doesn't deal with the usual loudness war crippled album. they basically suggest to apply about -1db gain as headroom so that inter sample clipping doesn't occur too often. it's IMO a good thing overall and I don't get why all lossy encoders don't include a little margin like that for such a well known problem? but just using replay gain while paying attention to the most dynamic albums(or using a quieter base value) can also avoids intersample clipping. "remastered for myself". ^_^
and very very few DACs keep a little margin just for that too. the answers have always been around, it's only that not enough people care. if someone mistakes clipping for the bad sound of lossy, all the better for highres marketing. and big SNR and crazy high sample rates are what sell a DAC, not that it won't clip your music
rolleyes.gif
.

 
Apple sound check aims for a target LUFS of about -16.5 RMS, so anything louder will simply be turned down. Unfortunately not everyone has gotten the memo yet and mixed/mastered music in order to take advantage of that headroom.
 
Sep 12, 2016 at 2:07 AM Post #9 of 132

gregorio

Headphoneus Supremus
Joined
Feb 14, 2008
Posts
4,514
Likes
2,542
[1] So the sampling rate/bit-depth war has been raging and most would agree that 16/44.1 is sufficient for human hearing.

[2] Why don't mastering engineers just distribute their better-mastered music on the CD platform?

[3a] "High-resolution" music websites tend to distribute different masters of songs. [3b] Vinyl recordings tend to be different masters of songs. [3c] SACDs often contain different masters of songs.

 
1. Not really. Most consumers don't know about different sampling rates/bit depth and don't really care. Knowledgeable audio/music engineers know there's no benefit to distributing higher than 16/44.1, a debate settled roughly a decade ago. The "war" is really only in the audiophile community.
 
2. Mastering engineers don't distribute anything. They just create masters fit for purpose as demanded by their clients; record labels mainly. And, what is "better-mastered"? ... A version/master ("A") with a higher dynamic range will sound better than a highly compressed version ("B") on a decent quality sound system. Version/Master "B" will sound better on a poor system, such as on the internal speakers of a phone, tablet, laptop or when using IEMs/headphones or listening to music in a noisy environment, such as in a car or when on the go. In such circumstances, which are extremely common today, version "A" could easily be so bad as to be un-listenable. So which version is the "better-mastered" is entirely different depending on where you're listening and on what.
 
3a. For more dynamic range, more suited to critical listening on a decent sound system and as castleofargh effectively stated, they have to have a different master to differentiate the "HR" version from the "normal" version, to justify the price hike.
3b. Vinyl releases need to be mastered differently to counteract the deficiencies of the media, the RMAA curve for example.
3c. Essentially the same as 3a. SACD is not portable and a SACD player is not cheap and therefore the most likely listening scenario is in a relatively quiet listening environment with a relatively high quality sound system. So the master can be targeted to a small, specific demographic.
 
With the exception of 3b of course, all these different masters could be distributed at 16/44.1 with no loss of quality but then of course it's much more difficult to differentiate between versions and justify double (or more) the price. Actually, the music industry has it easy. In the film industry, a feature film may require more than 70 different "masters"!!
 
Quote:
  Apple sound check aims for a target LUFS of about -16.5 RMS, so anything louder will simply be turned down. Unfortunately not everyone has gotten the memo yet and mixed/mastered music in order to take advantage of that headroom.

 
It's not that we haven't got the memo yet, it's that many/most clients don't distribute exclusively on iTunes. Youtube for example normalises to the equivalent of about -13LUFS. TV is normalised to -23LUFS. Feature film to roughly the equivalent of around -32LUFS, DVD/BluRay (film) releases to about -27LUFS. Even the TV standard is unplayable on many/most portable devices but audiophiles would very much like us to take advantage of that potential extra dynamic range. How many masters should we make? How much time and money do the clients have to make many different masters? Does the genre, production and artists' intention even lend itself to a large dynamic range? What/who is our target demographic?
 
It's not a black and white situation with a black and white solution!
 
G
 
Sep 12, 2016 at 5:33 AM Post #10 of 132

spruce music

500+ Head-Fier
Joined
Jun 10, 2015
Posts
555
Likes
344
   
It's not that we haven't got the memo yet, it's that many/most clients don't distribute exclusively on iTunes. Youtube for example normalises to the equivalent of about -13LUFS. TV is normalised to -23LUFS. Feature film to roughly the equivalent of around -32LUFS, DVD/BluRay (film) releases to about -27LUFS. Even the TV standard is unplayable on many/most portable devices but audiophiles would very much like us to take advantage of that potential extra dynamic range. How many masters should we make? How much time and money do the clients have to make many different masters? Does the genre, production and artists' intention even lend itself to a large dynamic range? What/who is our target demographic?
 
It's not a black and white situation with a black and white solution!
 
G


Well, I can dream can't I?   I would hope for masters without any compression or as  little as possible.  Sell that to the public.  Then let all the various devices provide the option to switch on compression.  Maybe a few different levels of it.  So I can listen at home in the quiet with full dynamic range.  I can engage compression for listening in the car or on the go otherwise.  Or I can do something in between.  That is what I would call a good solution. 
 
For legacy devices without those options, it isn't too big a deal to use software to do appropriate compression before loading music into the memory of whatever device is being used.   Alas I know it is only a dream. 
 
Sep 12, 2016 at 9:24 AM Post #11 of 132

gregorio

Headphoneus Supremus
Joined
Feb 14, 2008
Posts
4,514
Likes
2,542
 
For legacy devices without those options, it isn't too big a deal to use software to do appropriate compression before loading music into the memory of whatever device is being used.   Alas I know it is only a dream. 

 
And what exactly is "appropriate compression"? How many milliseconds of attack and release, what ratio, which settings for which frequency bands, what about the 40 or so other parameters I tweak? Which stems do you sidechain to others (which you can't do with a full mix), how about serial and parallel compression?
 
 
I would hope for masters without any compression or as  little as possible.  Sell that to the public. 

 
Sell what to the public, an unmixed pile of takes? How about sell them a rock album without a drumkit? What exactly is "without any compression or as little as possible"? A drumkit which doesn't sound like a drumkit, lead vocals where you can only hear some of the lyrics and the others are too loud or how about a noise floor which suddenly pumps in and out? Compression isn't just a dumb button you hit to switch on and just dial in how much, it's a complex tool used for creative purposes and then none or some additional amount may be used to make the whole mix louder, depending on genre, how it's been mixed and various other factors. Furthermore, how you apply it to one part of a song may well be different to how you apply it to another. The devices in consumer equipment are dumb, just applying more global compression at some fixed setting. We've dealt with this issue of dumb consumer compression for years in the TV/Film industry and the results are often ghastly, theatrical films being broadcast on TV usually have a separate mix made, to avoid this consumer (or broadcaster) auto-compression issue.
 
What you're wishing for is not possible, neither the record labels, artists nor engineers would even consider it and the results will not be even remotely as good as you seem to think.  Again, this wish comes from a misunderstanding of how mixing and mastering work in practice and is based on a gross over-simplification which only therefore appears to have a simple solution. Gaining some dynamic range at the expense of the sound and feel of the music may be heaven to a handful of extremist audiophiles but the huge majority buy music/audio content because they want to enjoy the content itself, rather than only enjoy one specific technical attribute of the mix.
 
G
 
Sep 12, 2016 at 10:35 AM Post #12 of 132

Lex2

100+ Head-Fier
Joined
Jul 20, 2008
Posts
340
Likes
25
  
It's not that we haven't got the memo yet, it's that many/most clients don't distribute exclusively on iTunes. Youtube for example normalises to the equivalent of about -13LUFS. TV is normalised to -23LUFS. Feature film to roughly the equivalent of around -32LUFS, DVD/BluRay (film) releases to about -27LUFS. Even the TV standard is unplayable on many/most portable devices but audiophiles would very much like us to take advantage of that potential extra dynamic range. How many masters should we make? How much time and money do the clients have to make many different masters? Does the genre, production and artists' intention even lend itself to a large dynamic range? What/who is our target demographic?
 
It's not a black and white situation with a black and white solution!
 
G

 
You make some good points. Most of the problem is probably market forces, each vying for more airplay due to the perceived advantages of a higher RMS level. I don’t know if it is about genre – Metallica evolved from a dynamic range of about 11 in the early 90s to an ear shattering 3 in some of their more recent releases. Then you’d get some mix guys providing already squashed tracks with extreme compression on the stereo buss, leaving little for the mastering engineer to work with. It’s fairly tempting to get more “glue” than one needs, and then dial up that gain reduction on an API or SSL to make the song sound “better.” These are the kinds of things that need to be pushed back on, but easier said than done I guess. Mix and mastering guys (and their producers) have to work together on this. It has to be a “paradigm shift,” although this time it isn’t really a new paradigm because it was commonplace two to three decades ago. I am old school, and long for the days of albums having more than a DR of 16 or 17, as was inherent in some of the early Dire Straits recordings, for instance. I think it would be a good -- albeit incremental step -- just to have something that spans a more “manageable” range from -16 LUFS to -1 dbTP, and have that turned down for radio and/or TV, or whatever regulatory standard they have to follow. One has to start somewhere, and it does look like the big players such as Apple are listening. If online isn’t already the dominant mode of listening these days (at least in the developed world), then I’d say it would be more and more prominent in the coming years.
 
I agree that compression is not something that can be dumbed down with something as simple as a switch. Next thing you know you’d have fairly complex music players with opto, FET, and VCA options, with multple bands!
 
Sep 12, 2016 at 12:02 PM Post #13 of 132

gregorio

Headphoneus Supremus
Joined
Feb 14, 2008
Posts
4,514
Likes
2,542
  Mix and mastering guys (and their producers) have to work together on this. It has to be a “paradigm shift,” although this time it isn’t really a new paradigm becomes it was done before two to three decades ago.

 
It's not just mix and mastering guys, it's the record labels and artists/composers and yes, it was done decades ago but newer artists are generally looking forward, not backwards to music that's decades old. Virtually all the mix and mastering engineers I know try to advise our customers not to crush the living daylights out of a mix and initially most are all for it, right up until they take a mix home and compare it to their other recordings. Then they come in the next day and say it sounds great (hopefully) but obviously it's got to ultimately sound as loud as David Guetta or some thrash metal band with a dynamic range of 1. Unfortunately, if you're trying to mix a rock band, the end result will be a master that's crushed to death because rock doesn't lend itself to being as heavily compressed as EDM and numerous other modern genres which have specifically evolved to include heavy compression. But, they are the customer and my job is to give them what they want.
 
Quote:
  I think it would be a good -- albeit incremental step -- just to have something that spans a more “manageable” range from -16 LUFS to -1 dbTP, and have that turned down for radio and/or TV, or whatever regulatory standard they have to follow. One has to start somewhere, and it does look like the big players such as Apple are listening. If online isn’t already the dominant mode of listening these days (at least in the developed world), then I’d say it would be more and more prominent in the coming years.

 
I think maybe you are confused about what LUFS is? LUFS is an average over the duration of the content, not a limit. Content measured at -16LUFS could in theory have an infinite dynamic range or none whatsoever. If it had peaks at -1dBTP then it would obviously have to contain material below -16 to maintain the -16LUFS average. The part of the measurement which deals with range is LRA (Loudness Range) but as far as I know no one really publishes that information and LRA levels are not mandated. If one had a 16LUFS mix which peaked at -1dBTP and you wanted to broadcast it on TV, you would need to turn it down by 7dB (8dB in North America), so your peak level would now be -8dBTP (or -9). That's a lot of dynamic range to throw away (lost at the loud end, not the quiet end) and TVs don't have a great deal of dynamic range to start with.
 
G
 
Sep 12, 2016 at 1:45 PM Post #14 of 132

Lex2

100+ Head-Fier
Joined
Jul 20, 2008
Posts
340
Likes
25
   
I think maybe you are confused about what LUFS is? LUFS is an average over the duration of the content, not a limit. Content measured at -16LUFS could in theory have an infinite dynamic range or none whatsoever. If it had peaks at -1dBTP then it would obviously have to contain material below -16 to maintain the -16LUFS average. The part of the measurement which deals with range is LRA (Loudness Range) but as far as I know no one really publishes that information and LRA levels are not mandated. If one had a 16LUFS mix which peaked at -1dBTP and you wanted to broadcast it on TV, you would need to turn it down by 7dB (8dB in North America), so your peak level would now be -8dBTP (or -9). That's a lot of dynamic range to throw away (lost at the loud end, not the quiet end) and TVs don't have a great deal of dynamic range to start with.
 
G

 
Not confused, I was being careless with terminology. I was referring to crest factor there. There is an interesting AES paper (Esben Skovenberg at TC Electronic) presenting some arguments for measuring LRA based on different assumed RMS windows to represent different genres, but you're right there are no standards for LRA now. Even if there was, it'd be difficult to implement, assuming one can get everyone to agree to a standard in the first place. You've probably read the paper but I can PM a link if you're interested.
 
Sep 12, 2016 at 3:33 PM Post #15 of 132

LajostheHun

500+ Head-Fier
Joined
Apr 26, 2014
Posts
741
Likes
129
I guess Steven Wilson figured this out awhile ago now since all his mixes are flat transferred, with no additional mastering and tinkering. And this is from the guy who learned this on his own yet it's not his primary job either.Of course he doesn't cater to the lowest common denominator nor he works with big labels either,unless he does a remix of old stuff by request.
 

Users who are viewing this thread

Top