Correct use of loudness metering for CD mastering?

[Master Mi]

I've got a question concerning the right use of the loudness metering unit in my DAW (Samplitude Pro X3 Suite) for CD mastering.

All soundtracks for my CD are already loaded into the music project file on track 1 one after another.

The loudness ratio between all tracks is already set via object volume editing (couldn't build just on loudness metering there - had to do this by ear because of different music genres and different dynamic ranges of the soundtracks).

And now I want to bring the loudest part of this CD mix at around - 23 dB.

For this purpose do I have to:

1) ...check the loudness metering for each track anew from each track start to rise or lower the volume of all tracks together until the loudest part of the loudest track reaches -23 dB?

or

2) ... let the loudness metering do its job by playing the whole CD (all soundtracks in a row) without pausing while checking the loudest part, memorize the highest loudness and rise or lower the volume of all tracks together until the loudest point of the whole CD loudness metering is set at -23 dB?

It's really strange that it can make differences up to 2 dB of loudness in LUFS between these 2 methods.
At method 2) the loudness differences according to the loudness metering unit seem to get smaller and nearly stable/unchanged after the first few tracks.
If you meter each soundtrack track just from the beginning (method 1)) the loudness differences according to the loudness metering unit are much bigger and the loudness metering seems to react much more sensitive.

Although I'm pretty sure method 2) is the right one for mastering the loudness of a CD I want to ask the OCRemix community about this phenomenon and the correct use of the loudness metering for this purpose.

[Master Mi]

C 'mon, guys - really nobody any kind of an idea?

Maybe some further informations about my loudness metering unit and why I guess that method 2) is the right one.

If I would start to messure the loudness right within a track for example it would show me a pretty higher max. loudness than if I messure the loudness of a track or another audio program right from the beginning to the end.
And that's why I guess that I have to messure the loudness of a whole CD right from the beginning to the end, too.

But I'm not quite sure if there might be another measuring technique for CD loudness mastering as well.

[Dalmae]

Hi,

I see your message and I can help you. First of all, I don't understand why you looking for -23dB LUFS (I suppose it's LUFS you're talking about). For the record, Spotify put everything at -16LUFS, Youtube at -12LUFS. A good method is to use the K-system, meaning the majority of your music must be at -12LUFS (or -14LUFS), which give you a PLR (Peak to Loudness Ratio) of 12dB or less.

However, do the mastering track per track, find a setting that sound good to you, and use it track per track. If needed you can change, but the less you change, the better it will, because all your album will be coherent.

I can advice you to read this article:
https://www.soundonsound.com/techniques/audio-mastering-in-your-computer

Don't hesitate if you have other questions.

 

[Master Mi]

Thanks for the answer. ))
 

I had thought pretty much about this topic in the last time and I even talked to a mastering professional - and I guess you 're right.
Although he seemed to be one of the new kind of mastering engineers who are more afraid of given away headroom than given away sound quality or a consistent loudness of audio programs he also said that you would do the loudness metering for each track intead of metering the whole CD in one shot.

And it makes totally sense to me now - because later I had found out that the metered loudness is the AVERAGE loudness of the specific part in or maybe the whole track you have metered (and not something like the momentary loudness).

So, if you start metering the track from 0:00 to 2:32 it will show you the average loudness of this part when finishing the metering at exactly 2:32.
If you start metering the track from 0:32 to 0:50 it will show you the average loudness of this little part when finishing the metering at exactly 0:50.
(That's where the strange measuring differences of the two methods I mentioned in the posting have come from.)

And this also implicates that it wouldn't really make sense to meter the loudness over the length of the whole CD - because the differences of the loudness between each track could be too big (one track a bit too silent, another track a bit too loud could also result in the same average loudness if you meter the whole CD in one shot).

But to factor in the different compression levels of the different tracks (could have been already caused in the production process of the tracks or maybe during the production of some nasty remastered versions) within different genres for my CD I still have to do something.

EBU R128 (that's the loudness metering method at which I wanted to master the tracks of the CD) defines that the (average) loudness of the track should be at a target level of -23 LUFS (dB) +/- 1 LU (dB).
So, I had a maximum range of 2 dB for the average loudness of each track to overcome the problem of different compression levels at different tracks on a CD.
This could be quite enough.
I guess the developers and audio engineers behind EBU R128 could have included this range of 2 dB for exactly these kinds of problems.

But since I don't know if a range of 2 dB will be always enough for mastering CD and since I've mastered all my other track at slightly below - 23 LUFS or dB (rather between -23,1 dB to -23, 3 dB) I guess I will keep it this way for this special CD:

1) leaving the loudness ratio of the tracks to each other as it already is (according to my perception of hearing when I was adjusting the fitting volume/loudness of the tracks back then)
2) metering each track and writing down the average loudness of all tracks to get the track with the highest average loudness
3) adjusting the loudness of all tracks together by keeping the loudness ratio between all tracks until the track with the highest average loudness hits the - 23 LUFS (dB) mark (or rather the - 23,1 to 23,3 dB mark) and all other tracks will be below that (so, the track with the lowest average loudness might be around - 25 or - 26 dB)...

I'm sure this could be a good way to keep it with all CD masterings with highly different and differently compressed tracks in the future.

If I want to bring out a CD with just my own tracks it will be mastered easily at an average loudness of -23,1 to -23,3 dB of each track because I don't use any compressors or limiters in my soundtracks - so there won't arise problems like these.

[Dalmae]

First of all, we are talking about music: the EBU R128 concern TV and radio, nothing more... Don't use as a reference for music CD, you can more further. But if you want to publish your sound in some platforme (like Spotify, Itunes...) I recommend you to put your CD at -14LUFS Long Term. For the loudness, you are several measure: Momentary, Short Term, Long Term. The EBU R128 are very restricted, you only don't have the loudness -23LUFS (-0.5 more or less), but also the short-term, momentary-term, the loudness range and the Maximum True Peak. This was designed for show and podcast with voices and to block the ads to be too loud. But in the music industries, we don't have those issues, we don't care.

For example, for one song, from the start to the end of the title, you must reach -12LUFS or -14LUFS (more or less) in long term, but you can reach +4LUFS compared to your long terme reference on short term with no trouble (meaning, if you're looking to be at -12LUFS, your short term can be at -8LUFS).

You still have to look at your peak, you should not trepassing the 0dBFS. So, fi you want to go at this level, you will need compressors and limiter.

You can trust me on this, I'm also a mastering engineer.

[Master Mi]

And that' s the reason why EBU R 128 can and should be used in music industry as well.

Not just because of uniformly loud music all across the internet - but also because you don't have to bother with useless sound chirugy or use of compressors or limiters for mere loudness gaining anymore - just as it nearly was in the 70s and early 80s.
If you master your tracks always at EBU R128 you won't have to watch the peaks all the time because there's always enough headroom that the peaks can come and flow.

Even in some of my soundtracks with higher dymanics the peaks won't go easily over - 7 dB.
The peaks have always enough headroom to breathe and sound natural.
 

 

[Dalmae]

I agree with you, we use too much compressors today, and I know what you're talking about because I remastered a lot of music from the 70's and even much older, and the sound are (very often) very excellent. It was more "sensible". The loudness war are now finished, but it happens for two reasons:

- if it's louder, it seems better

- if you want your music to be heard everywhere (phones, car ...), too much dynamic can make listening difficult or even betray your music if your low level is too low.

I understand your view, but even the 70's sound had a loudness between -20 and -14LUFS (in general, there was much freedom about music production, including mix and master). A bit of compression is not the evil, it's just a tool and sometimes it's very helpful to make music more coherent, more "together" without taking away the natural. -23LUFS seems very low to me, you need a good HiFi system to enjoy it and it's not the way people listen today.

Everyone does not feel concern by the loudness problematic and it's good to see someone cares about that. I'm curious to hear what you did.

[Dalmae]

One more thing

I looked the video you add in your previous post. The guy made a point but I was skeptical about the LUFS he gets, it seems wrong, very week. I mean Green Day at -18LUFS... there is no way they was so low...

So, this morning I checked several CD I have, here the result:

Coldplay_Square One (2005): Long Term: -9.6 LUFS ; TruePeak: +0.2 dBTP ; Short Term: -6.2 LUFS ; Loudness Range: 10.3 LU

Arch Enemy_Blood On Your Hands (2007): Long Term: -4.6 LUFS ; True Peak: +0.8dBTP ; Short Term: -3.9 LUFS : Loudness Range: 1.1 LU

Nirvana Unplugged_Come As You Are (1994): Long Term: -12.7 LUFS ; TruePeak: -1.1dBTP ; Short Term: -10.8 LUFS ; Loudness Range: 5.2 LU

Daft Punk_One More Time (2001): Long Term : -12.6 LUFS ; TruePeak: +0.6dBTP ; Short Term: -10.1 LUFS ; Loudness Range: 7.2 LU

And I checked one of my previous work:

Serge Gainsbourg_Melody (1971, remastered 2016): Long Term -19.2 LUFS ;  True Peak: -1.2dBTP : Short Term: -12.6 LUFS ; Loudness Range: 13.3 LU (it was the first time I listen on my new monitor, and it still sounds great)

I'm not going to make any statement about the differences, it's normal Arch Enemy sounds like that, every type of music have their own code, it's just I don't think -4.6LUFS is necessary.

So Green Day at -18LUFS, no way, so Michael Jackson so low... I don't think so, I pretty sure if I had the master tape, it will sound fucking great, but the result will be much more than it does shows it in his video.

[Master Mi]

On 5/29/2018 at 9:24 PM, Dalmae said:

- if you want your music to be heard everywhere (phones, car ...), too much dynamic can make listening difficult or even betray your music if your low level is too low.

But couldn't this already be an hint on how much compressed music can damage your sense of hearing - especially in the fine ability to distinguish between loud and less loud?

I mean... I often get the chance to listen to soundtracks on my MP3 player at work - at least as long as I don't have to verbally communicate with my colleagues and if I can work on my own most of the time.
And I've made an experiment today and a few times before.

I was listening to music made without compressors/limters (my own soundtracks I'm working at) - and I was listening to other video game remixes where often compressors and limiters are used.
I've listened to both kind of soundtracks at an adequate loudness level (to do this I had to listen to my soundtracks an my MP3 player on volume level 17 and to the other tracks on about volume level 9 - just for giving you an impression on how big the difference between mastering at EBU R128 standards and modern loudness war mastering relating to the volume/loudness levels can be).

If I listen a whole day to my own soundtracks there won't be a significant difference in my sense of hearing at the end of the day.
But if I listen just a few hours to other video game remixes where compressors and limiters are often used quite a lot and switch again to my soundtracks later on I often recognize that my sense of hearing has decreased a bit - just temporarily... but I won't hear all the details in my tracks for some time (maybe for some minutes until an half hour or so) afterwards and I don't perceive my own tracks as loud as if I would listen only to these the whole day.

Besides - the dude in the video didn't do the mastering for EBU R128 loudness standards in a correct way.
Since the metered loudness always means the average loudness of the part you have measured you always have go through the whole soundtrack before adjusting the master volume.
So, if the loudness meter - after starting the metering at the beginning of the track - shows a loudness of - 20 LUFS (dB) at the end of the track you would have to lower the master volume for about 3 dB to get at least into the right (average) loudness levels for EBU R128 standards.

And then you should also be concerned about the other important EBU R128 parameters like:
- Max. Momentary in LUFS (refering to a time period of 400 milliseconds)
- Max. Short-Term in LUFS (refering to a time period of 3 seconds)
- Max. True Peak in dBTP (...although you won't really have to bother with this parameter if you keep an eye on all the other parameters before)

[Master Mi]

On 5/29/2018 at 9:24 PM, Dalmae said:

Everyone does not feel concern by the loudness problematic and it's good to see someone cares about that. I'm curious to hear what you did.

Since we've talked about the 70s already I can show you a cover of a soundtrack from the 80s called "I'm always here" - by Jimi Jamison which was also played as the misical intro theme of the Baywatch series.
>>> https://soundcloud.com/master-mi/baywatch-opening-theme-master-mi-remix

Or maybe check out the Youtube version (cause I think Youtube offers a slightly better audio steaming quality with an higher audio bitrate of 192 kbit/s (compared to the 128 kbit/s audio bitrate streaming "quality" of Soundcloud):
>>> https://www.youtube.com/watch?v=9xbQHulo8h4

If you want to listen to one of my video game remixes I'm working at I would recommend my remix of the soundtrack The Price Of Freedom from Crisis Core: Final Fantasy 7:
>>> 

...or maybe my Star Tropics remix called "The Fire Of The Southern Cross" - if you want to listen to some more tropically summerly South Sea Island beats. ))
>>>

All the soundtracks are mastered at the EBU R128 loudness standard.

[Dalmae]

On 6/4/2018 at 7:09 PM, Master Mi said:

and I was listening to other video game remixes where often compressors and limiters are used.

If you use the OC Remix Album, I can tell you these are often too loud, at -10LUFS but this is not too excessive (it is, but not too much, I have seen worst). The problem is the dynamics, everything have the same level in those tracks and yes, ears get use to it. I didn't measure the PLR (Peak to Loudness Ratio) but it must not be good.

On 6/4/2018 at 7:09 PM, Master Mi said:

If I listen a whole day to my own soundtracks there won't be a significant difference in my sense of hearing at the end of the day.
But if I listen just a few hours to other video game remixes where compressors and limiters are often used quite a lot and switch again to my soundtracks later on I often recognize that my sense of hearing has decreased a bit - just temporarily... but I won't hear all the details in my tracks for some time

So, this is normal. When you're working in music, this is one of the think you need to be very, very, very careful. If in the morning, I'm working on a pop/rock song which is loud (and this is normal, the style is thought like that) and the afternoon, I got to work on a classical record, I will be very careful and rest my ears one hour before starting to work. Otherwise, the classical will seem to me too low and I will be tempted to push up the level.

I play the devil's attorney, but every style have a history, a background. And if the metal is over-compressed when they play live and in CD, it's meaning something. Put everything at -23LUFS

But I got to say, always listen all day to over-compressed music (and f*****g mp3) can damage ears and can cause attention disorders. If you are used too always have loud sound in your ears, when the silence comes, the brain is freaking out. And this can be reversible, ears can be educated (and need to be educated).

On 6/4/2018 at 7:09 PM, Master Mi said:

Since the metered loudness always means the average loudness of the part you have measured you always have go through the whole soundtrack before adjusting the master volume.

I know how it works, I used it everyday, but even in 5 sec, everything is too low. Coldplay in five seconds, I was clearly not at -18LUFS, so Green Day... and MJ...

On 6/4/2018 at 7:09 PM, Master Mi said:

And then you should also be concerned about the other important EBU R128 parameters

Don't you see my previous post ? I didn't show you the momentary, but I write about the Max Short Term and the dBTP... I always read them and focus on the sound, how it sounds (have I compressed enough or too much?) and find the best balanced parameters.

On 6/4/2018 at 7:09 PM, Master Mi said:

Max. True Peak in dBTP (...although you won't really have to bother with this parameter if you keep an eye on all the other parameters before)

Oh yes you must ! Ok, in mix music this is not the same, and even if I'm not a specialist in TV mix, I have seen how it works (part of my studies was in the movies industry and I could enter in some post-prod studio) and they look at it because they can very, very easily break it. That's why gun fire in TV are always bad, because they need to have a strong voice but they can't put a very loud noise over too much. The cinema can do that, in fact, the cinema surely can have the best mix and master because they can do how they want.

In music, it's the same thing but because the competitiveness, the output level of mobile phones and mp3 player and like I already said, "louder is better" (our perception work that way), producers want to over-compressed the sound. They are not engineers and this is not a good idea, this is just how it is. The AES/EBU is about television and radio, not music. This is why Spotify, Youtube, Itunes have their own Loudness Level, and it's shouldn't be that way. For me -16 LUFS is perfect, because you can do something very balance, or very quiet, or loud (-12LUFS long term at max), it will not alter too much the master.

We should do several master : one for mobile uses, one for Hi-Fi uses, one for radio uses...

On 6/4/2018 at 7:35 PM, Master Mi said:

maybe my Star Tropics remix called "The Fire Of The Southern Cross" - if you want to listen to some more tropically summerly South Sea Island beats.

Nice work, it's very nice. ;) I really like this one ^^