Understanding loudness, LUFS and R128 (using Measure RE)

[sinusfiction]

One of things I struggle with, is how to determine a good audio level for my music. I'm definitely no expert, but I've managed to make an album with tracks that are more or less the same loudness. I used Har-Bal as a tool in which I used one of the album tracks as a reference. My ears and the loudness shown in Har-Bal did the trick.

But, now I want to understand it a bit more, and I want to try to manage it from Reason. A few days ago I read a thread from Marco Raaphorst about loudness, LUFS and R128 and the RE Measure which can show the loudness the right way.

So, how to proceed? I've read that -23 is a broadcast standard, but for digital music platforms -18 is used. So, is -18 the way to go for leveling tracks for an album?

Now, about using Measure, how does this work? I just downloaded it to try, and I've put the display mode on R128 and the Scale Mode on -18. Next I play the track and (in my case using Ozone threshold) I try to get the bars around the 0. Is that the right way? I know, RTFM, but in this case it isn't helping me.

When I collapse Measure, I see the momentary, the short-term and the integrated loudness numbers. How does these numbers relate to -18 or -23? And how does the integrated loudness number (i.e. -13.5) relates to the loudness shown by Har-Bal (for the same track -17,5). Normally I try to get my Har-Bal loudness between -11 and -14, because... I think I read somewhere this would be a nice level. So this is way less loud.

Like I said, really no expert and no pro on this stuff. But I do want to understand how things work, and this is one of them... Who wants to shine some light on this. And, are there other ways to manage loudness from within Reason (how do you do it).

[selig]

Why not simply measure that tracks you want to compare with, and use that level? 

And if you're already mixing all your tracks at the same loudness (something that is typically not easily done at the mix stage and left to the mastering stage), then it sounds like you're already ahead of the pack!

[Tincture]

Hi, I understand you wanting some way of working to and I'm the same, I do like to be able to use a "numbers" system to work to and have a few things that I work to/with now. But Selig's right really, it probably is best to use your ears and a reference and it sounds like you're doing ok anyway.

Just in case you're interested: I use Flower Audio's Loudness Meter as the last thing in my audio chain in RMS mode with 3sec averaging and aim for peaks not to exceed -9.0 dB. I also use Har-bal and aim for the Mid average to be between -12 and -14.

[sinusfiction]

Yeah, thanks, both of you are right. And that's how I do it most of the time, by using my ears. But sometimes it's good to challenge the methods you use. And to learn something on the way. That's why I'm interested in this. Nice to read that you use Har-Bal as well, and aim for more or less the same numbers.

[selig]

I didn't mean to say "use your ears", I meant to say to measure the songs you like and that you want to sound like, and use that value for your mixes.

Personally I've been using the VU + Peak mode on the Big Meter, aiming for a crest factor (difference between Peak and Average/VU level) of around 12 dB - this is based on looking at other songs and finding that value is what I personally "like". Works like a charm and I can't imagine needing anything more unless I was asked to submit a mix at a standard broadcast level (which I've never been asked because I don't deliver to broadcast facilities).

[Tincture]

Hi Giles, can you explain practically how you achieve getting that crest factor? Or is there a bit in the manual I can look up? Or is it as simple as comparing the peak and VU levels and aiming for 12dB diff? Thanks.

[sinusfiction]

Thanks Giles, that helps. I guess the easiest way to achieve this is to aim your mix at this and then use a limiter and it's input control to reach this level further more (or use the Ozone and it's threshold level).

I just checked some tracks I recently made, and these tracks are more or less on this level. So it seems I am doing just fine .

[selig]

Hi Giles, can you explain practically how you achieve getting that crest factor? Or is there a bit in the manual I can look up? Or is it as simple as comparing the peak and VU levels and aiming for 12dB diff? Thanks.

The concept is indeed as simple as the difference between the Peak and VU level. When using a brick wall limiter, 0 dBFS is the highest peak level, which means that all you do it aim for the VU at - 12 dBFS (if you want a 12 dB crest factor, which is fairly common for louder music that still retains dynamics). The VU level will of course vary a bit - I just try to hit around that general area.

Because I use consistent levels for all inputs to the SSL (peaks around -12 dBFS), and also for the mixer (-6 to -3 dBFS peaks), and because I do a little processing at every stage, I can pretty much hit my crest factor target with little brick wall limiting (3-4 dB typical). 

When using the Master Compressor for mix compression/glue and Ozone for limiting, I boost the gain out of the Master Compressor to hit around 0 dBFS (remember you can go over 0 dBFS INTERNALLY and not clip). That way I can dial up the exact amount of maximum gain reduction on Ozone by simply setting the threshold slider at that value. If I want no more than 6 dB GR, I set the threshold to - 6 dB: simples!

[Tincture]

Thanks for the nice explanation. Interesting bit about boosting to 0dB after the BusComp so you know how much Ozone is doing

[selig]

Thanks for the nice explanation. Interesting bit about boosting to 0dB after the BusComp so you know how much Ozone is doing

If I was going to export the mix for mastering, and I wanted to keep the Master Compressor in the chain, I'd obviously not add that extra bit of gain, or add it with the Selig Gain that is in front of Ozone (primarily for metering purposes). This allows me to automatically/quickly bounce a mix that has the desired 3-6 dB of headroom for mastering.

[Gaja]

Sorry to be a little OT here, but I was wondering, is R128 something musicians have to worry about? AFAiK it is a standard for television, so it should be the responsibility of the mixing engineer who mixes te entire TV production to mix according to the R128 standard, no? Just curious...

[Marco Raaphorst]

I love this -23 LUFS norm offering super dynamics. It's like for the first time people start using the full range of audio again.

Sound is improving again because of this. The Loudness War is over.

[Gaja]

I love this -23 LUFS norm offering super dynamics. It's like for the first time people start using the full range of audio again.

Sound is improving again because of this. The Loudness War is over.

well then I guess it's time for the dynamic peace to break out

[JoshuaPhilgarlic]

Here's a pretty good lecture of one of the guys who developed the R128 standard.

https://www.youtube.com/watch?v=iuEtQqC-Sqo

[Marco Raaphorst]

This is the best link to follow, read and bookmark:
http://music-loudness.com

[selig]

One other thing worth pointing out is that loudness is subjective, just like the weather. There isn't a single value that designates loudness - there are a few new standards that attempt to estimate an average that most may agree with for the sake of aiding engineers that will be affected if the standard isn't followed. In other words, here's the standard for broadcast - if you follow this standard then there will be the least amount of processing applied. If you don't, you'll be compressed just like we are all already used to. That being said, if you already have a workflow that assumes compression during broadcast (thinking more radio than TV here), you may be better off sticking with it!

As for loudness, again it is a totally subjective and difficult to define parameter. Like the weather, there is no agreement on what is loud (hot?) and what is soft (cold?), and there is no consensus on what is twice as loud (just like there's no consensus on what is twice as hot or twice as cold). 

[Marco Raaphorst]

The new norm -23 LUFS or -24 LUFS for radio and tv does create a huge shift. It offers much more dynamic range making stuff which is very much compressed sounding boring. Loud mastered music now sounds super bad because its dynamic range is too limited.

Most people are not yet adapting to the new range. Me included. I have only done a few productions for -23 LUFS (using a maximum for short peaks at  -15 LUFS). For the first time we're using almost full scale dynamics which is just lovely to listen to imo. 

[wikholm]

Here's a pretty good lecture of one of the guys who developed the R128 standard.

Thanks for the link. Interesting lecture.

[sinusfiction]

This is the best link to follow, read and bookmark:
http://music-loudness.com

Thanks, great link!

[motuscott]

Would not this -23 standard accentuate the existing problem of quiet dialogue, loud explosions and even louder commercials on TV?

[Gaja]

Would not this -23 standard accentuate the existing problem of quiet dialogue, loud explosions and even louder commercials on TV?

afaik that's what R128 is for. To make sure the dynamic range is not too big, so your neighbours won't hate you for eternity.

[selig]

So, the new -23 standard has its benefits only as everyone´s audio work (TV, film, music) are adapted to this? It has all to do with comparison, right?
Bringing a new standard that allows for more dynamic range. 

This is where some confusing is coming IMO - this standard does not allow for more dynamic range! Everyone is always free to use as much dynamic range as they want, and there is certainly no old standard that was previously restricting dynamic range! 

At best you might suggest that this "encourages" everyone to choose a dynamic range standard that worlds for them, similar to how we currently choose our dynamic range but now there are some "numbers" that attempt to quantify this difficult parameter called loudness.

I am still too off in this topic as I do not really yet understand the difference between the new and the old standard. I do watch the video (and will do it many more times) but one have to know quite a lot already to understand the "revolution". 

There is no "old standard" that I am aware of, which may be why you are experience confusion here!

So, as a Reason user, what does the new -23 standard makes a difference to ones workflow. What will I aim at different in the practical work, than was the case with the older standard? 

It all depends on what you have done in the past. If you have made less dynamic sounding mixes this will pretty much force to you change what you think is "right" in an attempt to conform to a "standard". The "promise" is that if you are broadcast or otherwise played on a system that also adopts the same standard as you adopted, your mix should not have to be "processed" on air and will be played back as is. If your mix registers louder, it will simply be turned down so it matches the loudness of other mixes. This is not such a bad thing when you think about it - your mix will sound the same loudness (according to the "standard") as other mixes/commercials etc. no matter what you do!

I have not yet come so far in my development as to analyze loudness levels to get a final song. All I know myself is that headphone amplifiers for instance, they have a hard time giving justice to a song with lower loudness, while over pressed song is tiring for the ears as well. In that context, where does the -23 comes in?

Not sure it does come in here - if a mix sounds "tiring to the ears" to you, then it's probably not "right". You already have the skills to recognize the mixes that have this affect on you! For me, when I hear a softer song and want it louder, I turn up the volume. But a "harsh" mix won't sound "right" at any volume. So always better to err on the slightly too soft end of things IMO!

What in my mixing and amateur maskering will it help me get a better result? What kind of mindset should I have in this new lodness zen? 

You should simply try to avoid harsh sounding mixes that sound tiring to your ears! Typically sounding, the louder you try to make a mix the more you risk creating a harsh mix. This is because limiting etc. essentially produces square waves as you push it harder, by "clipping" off the peaks. When you turn smoother waveforms into squares you hard higher harmonics, which can result in a "harsh" sound.

Beyond that it's more about the genre and audience IMO. Classical and jazz styles are naturally more dynamic, as that is how they are performed. Dance music is much less dynamic, as that is how it is presented (and why it works!). I think the "issues" arise when a classical piece is squashed beyond recognition, to use an extreme example. OTOH and at the other extreme, no one ever complained about the loudness wars on the dance floor (where loud playback is to be expected). It's about context and knowing your intended audience IMO, and it's up to each artist/producer to choose the sound of their mix first, and the loudness second IMO.

How would one go about adapting a reason song to -23 standard?

Probably use less compression overall, on individual tracks and the overall mix. But this is only required if you are choosing that standard for your own music. It may turn out that you just can't get the "sound" you want at that loudness. The "standard" can be skewed by soft parts of a song, so if your mix alternates between loud and soft sections it will read lower than one that is flat out loud top to bottom. Remember that these standards are just the best we have come up with at present to try to take a very complex concept ("loudness") and try to assign a single value to each audio mix. It's at best an estimate, and at worst just another standard that will one day be replaced by something better (hopefully!).

Right now I only have
1) Trust my ears and try and error, and listen on multiple sources and mixing in multiple volymes. 
2) Compare to music that I do find compelling.
3) Cannot go over 0db in the digital chain.

What else can I do and know as signposts? I really want to deepen my awareness in this field.

I would add this. Try to understand the difference between peak energy and average/sustained energy, and how this relates to what we perceive as "loudness". The ears are slow to respond to volume changes, relatively speaking. That is why a VU meter is simply the peak level run through a 300 ms attack/release time constant. If you play a full level tone into a VU meter, it will take about 300 ms for the needle to reach this level. This is to estimate how our ears "hear" sound as it applies to loudness. 

What this means is that if you have two sounds that both peak at the same level, but one of them sustains like an organ and the other decays like the sound of two sticks hitting together, the organ (sustain) sound will sound MUCH louder than the stick click sound (all other variables being the same). This is because the transient attack of the stick click is so short that the brain barely has time to register it! To make these two tracks "sound" the same we would need to turn up the click sound, possibly quite a bit. But when we do that, the peak level also comes up, right? That's where compression comes in, or more likely "limiting".

By limiting the click sound and then using makeup gain to bring up it's level, we can compromise between the peak and average level and create a sound that is as "loud" as the organ AND shares a similar peak level. But it may also "sound" compressed, and that is where the fun begins! 

Much of my work when mixing is in balancing the transient to average energy so that each track contributes to the overall loudness of the track in the best possible way. A channel that is too "transient" will add peak energy but not be easily heard. A track with too much sustain by also be lost in the mix because there's no "definition" on the transients - exemptions for things like pads that are by design "transparent". Each individual track needs to be addressed in this manner, as does each bus and the mix itself. That's my approach, anyway and FWIW!

I can get an ok mix, but I often find my so called master is often less than so many other songs I hear. And that is even when pushing it way more than I want. So I am in confusion where to change this in the audio chain. I know that mastering is an art in itself and that it is better off to someone else. 

Could be you are trying to achieve all of your loudness in mastering. I've said it before: a loud mix starts with the FIRST element you create. If your mix doesn't already sound loud before mastering, you're probably needing to address things earlier in the process. Programmers use the term "garbage in, garbage out" to essentially say you can't polish a turd. It's the same thing here - a "garbage" mix will result in a "garbage" master!

Each sound you choose all contribute to the loud mix you are after. By addressing things at the root or foundational level, you build a stronger production. IMO the MOST works on a production is done the earliest, and the work decreases as you progress leading you to a mastering session which only has to 'polish' and not actually 'shape' the mix. 

I say: don't fix it in the mix or mastering, fix it in the TRACK! The better arranged tracks are actually a breeze to mix/master in my experience.

Professional mastering has a lot of analog and high-class gear and ears, to bring musicality when blending it all together into fulless. Can emulating that at all be done inside of Reason? I would rather hear some professional music before and after mastering so I can learn step by step difference between a good mix, and a good master.

A good mix shouldn't sound that difference after mastering. In some cases there is very little mastering required. In fact I have a few engineer friends who have had mastering engineers say they didn't really change ANYTHING on some tracks. When the mix is right, there may NOT be any mastering required - this will depend on what "mastering" means. An album of songs will be mastered to all sound right as played, but a single can be mastered differently since there is no other songs to "blend" with. 

Either way, as they say in Nashville: it all begins with a song. Adapted to our conversation: "it all begins with a solid track". 

[sinusfiction]

I'm glad I started this thread! Not only I learned a lot about loudness, but got some terrific mixing tips as well... Thanks selig!

[tibah]

Might be just me, but a lot of the explanation and asking in this thread made me wonder, since the big meter, in its default setting, is already set to show you red indicators once you pass -12dB RMS. It will also show on the *normal* master out meters in the mixer and the rack.

 

I believe this is a very good starting point. A lot of commercial and club music goes way beyond this. I like to have some of my songs hit about -10 when the song is really busy, because this dynamic range just works and this is what everyone else has to figure out for themselves.