EBU R-128 meets K-System v2, a possible future for the loudness debate (Loudness War)

[tanabarbier]

Thanks, it is actualy usefull, I think that the way I am trying to do it is the good way for me, but my problem is how to make the engeneer understand and don't think "yeah, right, I'll put a comp and limiter anyway". I think it's more of a comunication problem that anything else. (communication breakdown, it's allways the same...)

I know that huge dynamic is a problem in noisy environments, but as I am playing the music live and pay a lot of attention to levels, I rise them myself if they go unnoticed because of noise, so, I think that's ok.

I think that a big part of those problems is the lack of experience with that kind of music in live situations. This dynamic isn't bigger than one of an orchestra playing a symphony (between ppp and fff, even if I never measured it, I'm sure there is more than 30 dB RMS), so it is possible to play like that.

[Sounddigger]

So can the EBU R128 standard successfully be used to balance music, or not?

[an excellent explanation for a possible usage following this statement]

This is indeed an interesting concept you show here, and is actually how it was planned to be used. Somewhat at least. Though the focus should be on the ML meters, not the SL meters.

The SL meters (3s measurement time) can give you a more closer readout for short term loudness for individual tracks compared to Integrated, which is a middle way of the whole stream (broadcast). But as you found out yourself, if the SL reaches up to +3, chances are that the ML reaches up to +7 or higher. So the focus should be the ML meter.

It's interesting that you're focussing on the MLk meter. However, not only will this work in the advantage of tracks mixed in the way of the loudness war, as far as I understand the video presentations and the PDFs with technical details written down as in the "EBU Tech 3343" document (page 23) found at http://tech.ebu.ch/docs/tech/tech3343.pdf , the focus should in the end be on the SLk meter.

And really , this makes all the sense in the world, if the objective is to introduce a standard that will counter the effects of over-compression/limiting.

If all you do is set a maximum peak-level (whatever way you set it's average) on the MLk meter, then over-compressed/limited material will always end up louder than dynamically rich mixes. If you reconstruct the experiment and resulting method I described in my previous post, then this should quickly be very clear.

After all, it seems to me that if your reference is only the MLk meter which is based on 400ms ballistics, then you will be tricked by the cumulative effect of the release time setting(s) of compressors/limiters, while it's not uncommon to have a compression/limiting on a master, with release settings higher than 400ms.

Consequently, the idea to introduce a standard based only on the MLk meter, will never be advantageous to dynamically rich mixes. On the contrary! Even if you combine it with a very arbitrary but hardly "absolutely objective" interpretation of the meter's moves, results will at best be dubious.

With all the work you have put into it, I find it very surprising you seem to overlook this point. That is, unless your goal is something else than to counter the "loudness-war". If your goal is merely to skim some dBs off ALL mixes, then yeah, maybe that would be useful in certain cases, such as those you hinted at. But it does nothing to reduce the practice of over-compression/limiting.

On the other hand, maybe you simply have not checked mixes that vary enough, but yet can be found commonly the moment when you look elsewhere than in the categories of pop, rock, metal, dubstep, which you name in your "white paper" (for those who haven't read it yet, see the opening post). You describe in that paper that you selected to check tracks mainly in these categories, and only tracks of "accoutic guitar and vocal in the accoustic department" (as mentioned on page 7). Consequently, you found that (as quoted from p. 7): "Mostly, there was only +2 LU difference. I have to note however, that the read-out can be much higher in theory - depending on the source material."

The point is that read-outs are not just higher "in theory", but really quite frequent in genuinely dynamic mixes, such as are commonly found in particular for classical and Jazz, but also for genuinely dynamically mixed electronic music. Since I included a lot of tracks from all these categories in my tests, I found very quickly that sticking to the MLk meter simply doesn't work to counter the loudness war. After aligning peaks on the SLk meter (as described in my post), I even had a couple of peaks going up to K16v2 +8 MLk LUFS, with SLk max peak at +3 LUFS, a difference of +5 LUFS, which is huge (relatively speaking).

I had to conclude that overall, if the goal is to find a good level between tracks so that all feel as being equally loud, then looking at the level of peaks on the MLk meter is frankly ultimately irrelevant! Let me point out Compyfox, that even you propose that there should be a different standard for (quoting from page 9 of your white paper): " "full dynamic" music content like orchestra, accoustic recordings, etc.". But this means that the level issue remains unremedied (between genres at least)!! And I'm not even raising the question of what should exactly determine "pop" vs "orchestra". I could name quite a few tracks that would fit in both (including hit records). Anyway, if you happen to play a pop/rock track before an orchestral/classical track, both aligned according to your system, then the latter will generally sound significantly less loud than the first.

From my testing, it seems more effective to set maximum peaks on the SLk meter, while using the MLk meter only to analyze if a track is overly compressed or not. If for a given track the MLk generally reads-out within +1 (or less!) to +2 above the SLk level(s), then this is a sure sign of a track that is over-compressed/limited. And this is about all I have found the MLk meter to be useful for, really...

The DAW file with which I tested all this contained no less than 33 different tracks, the last time I counted. And I checked the MLk and SLk read-outs for each and every one of them, experimenting with different leveling approaches. So you can say that I gave it all a fair shot and looked at it quite extensively...

The method I describe in my previous post can effectively deal with the loudness war, but it will still be very difficult to avoid some degree of arbitrariness, as results will differ from one engineer to another.

Therefore, it's not suggested for use during mastering of CD/digital releases (except for leveling tracks within one release, with peaks at levels that are appropriate for the kind of release at hand). I presently simply see no feasible way to introduce a standard that can objectively attain that ALL tracks will "feel" as being at an equal SPL level.

IMPORTANT:
I do not(!) use the K-System while mixing. For this I trust my Digital/VU meter combo with a reference level of -18dBFS (VU) and a maximum peak of -6dBFS (digital meter, per channel) or better said -3dBFS (digital meter on the summing bus).

I only use the K-System v2 while mastering or doing audio montages.

Agreed, as this implies setting an optimum peak-level, which should not be ones concern during mix-down (besides avoiding clipping the track).

However, personally I like to check things at times on a K-meter, as well as on the DR-meter, and others.

By the way, I really like the K-meter found at http://code.mzuther.de/kmeter/ (which I have not seen posted on KVR yet, although maybe it is somewhere...).

Also, for those working with Presonus Studio One it might be useful to know that K-meters can be found at several places, such as on the Master bus meter. To find it on the master bus, click with the right mouse-button on the meter graduation, and select the meter of your choice.

Hmmm... I have no problem with the MLk not having a histogram. I'd rather see a small improvement for the SLk, so that it's possible to disengage the peak "hold" function of the cipher that indicates its level, so that it refreshes even when the level goes down.

As it is now I'm constantly using the "Reset" button... (but the "hold" function is useful the other half of the time, of course).

May I ask why?

Because I tried to align peaks at K16v2 0 SLk LUFS, or +1 up to +3 (depending on dealing with verse/chorus, dynamic or compressed audio). After adjusting the level downwards, the hold function will cause the previous peak still being indicated, which means you have to reset the meter to see (clearly) what you've achieved (with hopefully a lower peak). That's all. Minor issue though...

Else... Yes, again a lengthy post. But it seems to me that you found a solution for you which can be further optimized.

Definitely some good input for me, which might result in an update of the PDF. Maybe even a slight tutorial in either form how to use the meter "in my opinion".

If it has proved useful, then that's all we can ask for

Which leaves the question with the time table I made earlier.

Currently I recommended to drop K-14v2 in favor for K-16v2 by the end of this year. Since I don't see an imminent implementation or release of Loudness Normalisation devices in modern playback devices yet (at least not by end of 2013), is my idea to go down to K-16v2 ASAP (by 2014) to far fetched?

Or should I/we extend the "grace period" until end of 2014 so that even the squallers can adapt?

As you might understand by now from this post, I still think it's not feasible to introduce a standard that can effectively and genuinely but most importantly: objectively, remedy the (arguably) mal-practices of over-compression/limiting.

Even when applying a fairly effective method such as I described in my previous post, you would have different results from one engineer to another. So it can't be imposed as a universal standard. Your own proposals also leave much room for such differences. This lack of 100% objectiveness is a major issue in itself.

And here's another issue that can complicate matters even further: say I'm releasing an album with a collection of tracks that I have carefully balanced so that the entire album can be listened to from beginning to end, with a particular listening experience and without having to touch the volume knob. If I use the method I described in my previous post, then this will fall far away from any method based on only MLk read-outs. Also, I might decide to set an entire track at a somewhat lower level. But this would then deviate from the standard, so when combining that track with material from other releases, we'd have ended up with yet another incompatibility issue.

And I could go on and on finding many other valid problems with the idea of setting a universal standard for the level of tracks, in other ways than that they are not supposed to have a TP that exceeds 0 dBFS (and if they do have peaks that exceed that level, it's only normal these tracks will clip, distort, or be turned down anyway).

But perhaps most importantly, and this goes in the direction of what AudioGuy720 posted, we're dealing with something like 100 years of recorded material, and around 25 years of material that has been released on CD or other digital carriers. So what about that? Some material will be too soft, some too loud. It will always remain a problem, whatever standard would be introduced.

My present stand is that the level of masters should be left to each one producer/engineer's own taste. Since there's already such an enormous diversity amongst musical recordings, it really doesn't matter anymore.

As for the consumer, I think we will surely see more and more technical solutions included in the devices used to listen to music/audio.

Software wise we already have the Melda MAutoVolume, albeit not yet implemented in any device. I haven't tested the MAutoVolume yet, so I don't know how well it would serve as an overall leveler, across all kind of tracks.

What we're looking for in this case is a leveler that particularly takes into account the (average) dynamic range of a track, so that it can compensate for over-compression/limiting in an astute way. The Aphex compellor doesn't do things that way, although it has an astute way to set the final output level.

The other day I found an interesting compressor posted on KVR, the "Autocompressor", which can be downloaded at http://www.elec.qmul.ac.uk/digitalmusic ... mpressors/
It has an interesting feature where it does an auto-calculation of release time based on dynamic range. However, this is not yet effective enough in adjusting tracks that differ greatly in DR as well as overall loudness level.

At any rate, I think that all the research and thinking that has gone into this subject is VERY useful, and should generate some new ideas for how to deal with these issues.

I know that huge dynamic is a problem in noisy environments, but as I am playing the music live and pay a lot of attention to levels, I rise them myself if they go unnoticed because of noise, so, I think that's ok.

I think that a big part of those problems is the lack of experience with that kind of music in live situations. This dynamic isn't bigger than one of an orchestra playing a symphony (between ppp and fff, even if I never measured it, I'm sure there is more than 30 dB RMS), so it is possible to play like that.

Sure it's possible. The only question is how you deal with the audience's attention span. I presume that you generally don't perform in settings that would compare to that of a classical orchestra or ensemble, although maybe you do (don't know your work after all). Even for classical works it's relatively rare to start at an extremely low level. Typically, you would see such a build-up only after a movement or 2. It's a matter of preparing the audience first. If you have an audience that is all put quiet and ready in their chairs (such as for example in a ballet setting where raising the curtain would signal the start of a piece), then a build-up from extremely quiet levels can work wonderfully. A good example of that would be Ravel's "Daphnis et Chloé.

Anyway, experimenting with all this will be your ultimate guide. I can sure see how you could build an interesting "electro-accoustic" ambient-ish tension with a significant time-span. But much of its success will depend on how you play the crowd

Best to all

[Compyfox]

Thanks for joining the debate again, and thanks for the bump.

...as far as I understand the video presentations and the PDFs with technical details written down as in the "EBU Tech 3343" document (page 23) found at http://tech.ebu.ch/docs/tech/tech3343.pdf , the focus should in the end be on the SLk meter.

And really , this makes all the sense in the world, if the objective is to introduce a standard that will counter the effects of over-compression/limiting.

I do see a problem with focusing on the SLk meter compared to the MLk meter in either available tool.

If we focus on the SLk meter for audio CDs, which is indeed a great idea compared to Integrated and having a constant eye on Momentary, then we have to shift down the rules/color codes. By at least 3LU.


As pointed out, if we have a SLk value of 0LU (or -16LUFS), certain sections in a song can still result in a +9LU (-7LUFS) MLk peak. Depending of the density of a mix. Which in turn would mean an exceed of the -1dB True Peak limit and to some extend even needed heavy limiting.

While this production can be indeed darn dynamic, we still introduce something that we want to prevent: heavy peak limiting. To a certain extend, to keep the track in check, even compression again.

But don't we want to get away from that?




I currently see it this way:
Both techniques have their advantages, and their weaknesses.

The MLk meter in my opinion is more true to the K-System v1, if done properly (mezzoforte to forte passages for chorus/soli/breakdown between -3LU and 0LU).

Using the K-System alone to align things, if you're used to it, can work great on a whole CD production. But I always missed some sort of histogram function to see where I'm actually at for possible fine tunings. But it sure is a thing of trusting your instincts.


The SLk meter could work as more unified way for a v2 standard, since it's measuring sections of 3seconds rather than immediate peaks. It's more technical, less "instinctive", especially with the histogram as a visual representation of what's going on. Though locked to the gate functions of the EBU R-128 specifications.

But in turn with the SLk meter, the Momentary Loudness peaks can go haywire and we need (as mentioned earlier) a shift in terms of color codes and strict rules.


So which is the correct way:
Both, and none actually. We have to trust our gut - one way or another.




I think we should indeed also have a focus on the Peak-to-Loudness ratio, or Dynamic Range (Crest Factor) regardless of the used meter. No matter which genre, we can reach up to 15 or even 17dB of dynamic range. Most of the time (with tests I did, and yes, they were mostly based on popular music), I got readouts of about 15dB in terms of dynamic range. With K-16v2 and K-14v2. Overall they were around 9 to 12 (dense mixes), voice over with background music can indeed differ (PLR), could in terms of Dynamic Range (average volume to max volume peak) it can still be 15dB.

I am not saying, that we should constantly try to stick to that value, but I'm saying it is possible while still having a balanced out loudness.

Finding a middle way or to actually "align" things, is the tricky part. Which is the very reason I have that many pages with "concerns".




I do however see certain issues with plain CD productions. As you mentioned, an artistic choice could still be to have a ballad at a lower loudness compared to a more upbeat production. This was always the case back in the days, but dulled out within the last decade or so.

This is also an often overlooked factor in the whole equasion. The question resides: are we allowed to have a fluctuating CD production, or are we forced to stick to a loudness shift of +/- 1 LU?

There are no clear rules yet. And I do not want to make them either as things can go terribly ugly.




What we do in here is actually a good thing:

The thoughts we're currently gathering is something that should be spread, let people actually think about what they're doing, maybe start from scratch. And not letting go things out of hand. I mean, the last AES (2012) already got note regarding this issue thanks to Mister Katz and to a certain extend, the implementation of the K-System v2 in EBU Loudness as well.

Of course we can simply say "funk all that, as long as -1dBTP is not exceeded and the loudness doesn't go higher than -16LUFS/-14LUFS with either meter, I'm fine.

But is that really it?
Should we really trust future loudness normalisation means to work that flawlessly?

We still bend the rules either way.




Regarding the timeline:
I think this is something that needs to be redone. I maybe shift it for about a year or so. More than enough time to adapt, create more awareness in between.


I don't say this white paper is a holy grail. But if it did something right, then it's creating a platform for open minded debates. And this is congruent with my original ideas.




So again, thanks for the input.
Definitely appreciated and adds a lot of value to the whole thing.

Now it's down to fine tuning.

[Sounddigger]

...as far as I understand the video presentations and the PDFs with technical details written down as in the "EBU Tech 3343" document (page 23) found at http://tech.ebu.ch/docs/tech/tech3343.pdf , the focus should in the end be on the SLk meter.

And really , this makes all the sense in the world, if the objective is to introduce a standard that will counter the effects of over-compression/limiting.

I do see a problem with focusing on the SLk meter compared to the MLk meter in either available tool.

If we focus on the SLk meter for audio CDs, which is indeed a great idea compared to Integrated and having a constant eye on Momentary, then we have to shift down the rules/color codes. By at least 3LU.


As pointed out, if we have a SLk value of 0LU (or -16LUFS), certain sections in a song can still result in a +9LU (-7LUFS) MLk peak. Depending of the density of a mix. Which in turn would mean an exceed of the -1dB True Peak limit and to some extend even needed heavy limiting.

This is very right. In fact, I had 2 tracks with peaks on the K16v2 MLk of over +8 LUFS. Haven't checked if it was +9, but might be the case. For the sake of my test (which focused on finding a satisfying method to align gain levels of different tracks in a way that felt "right"), I actually limited their peaks by 1dB. As one of the tracks merely clipped on an ugly transient, it was unnoticeable, the other was on a drum hit, with limiting barely noticeable.

So yes, if you're talking about aligning ALL tracks in the world in a universal way, than the "basis level" (for "verses") should be at the K16v2 SLk -1 or -2 LUFS. Maybe it should all be transformed into a K18v2 system.

However, again, I believe there are simply too many differences between tracks, too much of a recording heritage, and too many differing views on the matter to want to introduce any mastering levels as a standard. The method I described should merely be seen as a technical approach that can have real practical use in leveling different tracks when used in a release or AV production.

While this production can be indeed darn dynamic, we still introduce something that we want to prevent: heavy peak limiting. To a certain extend, to keep the track in check, even compression again.

But don't we want to get away from that?

This depends on how purist you want to be about it.

I think there's a place for all approaches, although I personally will tend to stay away from over-compression, because I definitely experience listening fatigue much faster in those cases, although much will depend on the quality of the music as well. If you have bad/mediocre music + uber-compression/limiting, then it's definitely hell...

I currently see it this way:
Both techniques have their advantages, and their weaknesses.

The MLk meter in my opinion is more true to the K-System v1, if done properly (mezzoforte to forte passages for chorus/soli/breakdown between -3LU and 0LU).

Using the K-System alone to align things, if you're used to it, can work great on a whole CD production. But I always missed some sort of histogram function to see where I'm actually at for possible fine tunings. But it sure is a thing of trusting your instincts.


The SLk meter could work as more unified way for a v2 standard, since it's measuring sections of 3seconds rather than immediate peaks. It's more technical, less "instinctive", especially with the histogram as a visual representation of what's going on. Though locked to the gate functions of the EBU R-128 specifications.

But in turn with the SLk meter, the Momentary Loudness peaks can go haywire and we need (as mentioned earlier) a shift in terms of color codes and strict rules.


So which is the correct way:
Both, and none actually. We have to trust our gut - one way or another.

The question is what use you have in mind precisely for each meter.

My comments about the use of the SLk readings are strictly to be understood for how to go about aligning the levels of different tracks so that they all "feel" as being equally loud. It's not to be understood as any recommendation for mixing down or mastering. On the other hand, I found the method particularly useful for maintaining a constant SPL during mixing, with all the advantages this offers.

When mixing down or mastering, for me personally I don't feel there is a significant difference between the K14/16v2 and the original K-meters as such, except that I prefer to look at the "old" K-system meters, as I definitely experience the difference between the 400ms and 600ms ballistic. The 400ms is simply a bit too hectic for my taste. It gets on my nerves, really... I'm just as fine reading a normal FS meter.

I think we should indeed also have a focus on the Peak-to-Loudness ratio, or Dynamic Range (Crest Factor) regardless of the used meter. No matter which genre, we can reach up to 15 or even 17dB of dynamic range. Most of the time (with tests I did, and yes, they were mostly based on popular music), I got readouts of about 15dB in terms of dynamic range. With K-16v2 and K-14v2. Overall they were around 9 to 12 (dense mixes), voice over with background music can indeed differ (PLR), could in terms of Dynamic Range (average volume to max volume peak) it can still be 15dB.

I am not saying, that we should constantly try to stick to that value, but I'm saying it is possible while still having a balanced out loudness.

Finding a middle way or to actually "align" things, is the tricky part. Which is the very reason I have that many pages with "concerns".


I do however see certain issues with plain CD productions. As you mentioned, an artistic choice could still be to have a ballad at a lower loudness compared to a more upbeat production. This was always the case back in the days, but dulled out within the last decade or so.

This is also an often overlooked factor in the whole equasion. The question resides: are we allowed to have a fluctuating CD production, or are we forced to stick to a loudness shift of +/- 1 LU?

There are no clear rules yet. And I do not want to make them either as things can go terribly ugly.

Again, it seems to me that with the present state of technology and meters at our disposal, even if they are wonderful tools of precision, there is no way to have a 100% objective method to align tracks at a level that feels equal for all. And then there's also the extra complication when tracks are mastered at levels that are below the standard, which is as problematic as when tracks are above the standard, really.

That's why the only solution viable must happen at the consumer level, whatever way you look at it.

So again, thanks for the input.
Definitely appreciated and adds a lot of value to the whole thing.

You're welcome

Take care

[rolander]

Compyfox, thank you for this thread. It's a little step against the loudness war. But I fear it's a futile struggle, tilting at windmills. We have had convincing initiatives against it, e.g. the K-System from Bob Katz and the Dynamic Range classification from Tischmeyer. Now we have the EBU R128 recommendation for broadcast and more and more TV- and radio channels in Europe adopt it.

But look at this:

(Sorry, video doesn't show up here. In the preview it works )

This guy's video was seen by more than 400.000 people and the overwhelming number of comments were positive to enthusiastic. I don't want to be rude and offensive because this guy wants to educate people to produce better music and he does it for free. His intension and some of his tips are quite good, but he clips all the transients off and ends with a very loud master with about 3 to 4 dB dynamic range. He and all his followers on Youtube think that the song sounds much better after mastering. That is the crux! I don't blame the Ipod generation (to which he seems to belong to) for it. The younger people are grown up with the over compressed music of the last 20 years and they had never the chance to listen to clear, transparent and dynamic music. They know nothing else than heavily compressed, distorted and overloud music. And they wouldn't recognize the benefits of dynamic mastering because of the huge loudness difference. Imagine that you have a playlist on your Ipod with a lot of temporary music or remastered loud classics and one original master like "Stairway To Heaven". Imagine, you never heard that song before. You have adjusted the volume according to the louder contemporary songs. I think that you would skip "Starway To Heaven" because it sounds dull and much to soft in comparison. Without a loudness normalization, dynamically mastered music can not compete against over compressed music. But loudness normalization can not be installed by law but only by the power of persuasion.
I think we have to educate the listeners themselves rather than the mixing and mastering engineers. They know about the problem but they have to do what their customers want.

What we can do is to show comparisons between dynamically mastered and over compressed songs at the same loudness. We should train people to discover the richness of music with lot of transients, we should demonstrate them that harsh clipping is no good to the quality of music.

Your aim is to achieve a new standard value for music of -16 dB weighted (K16 V2). That is a good idea. There should be enough headroom for dynamic music. But I also see the point that it is not so easy to produce a consistent album when setting the maximum of every song to the same level. You guys mentioned crescendos or calm sections or songs like ballads. I therefore wouldn't be so strict. I would recommend a loudness level standard for a whole album so that some songs could be louder or softer. I think the louder ones wouldn't be mastered too loud because that would destroy the balance and flow of the album. But for music distribution platforms in the internet there should be a strict standard, else wise it would be a mess to make a mixed playlist of downloaded songs.
There are some platforms like Spotify who use loudness normalisation yet and Itunes has a preference setting called "soundcheck" which is described as working in the same way as ReplayGain. Good news! I hope that more will follow. But there are other popular web pages like Youtube or SoundCloud and a lot of sites where you can publish your music for free. There, the loudness war is still taking place. So education especially on Youtube is essential. Problem is: no one cares about your education videos if you are not popular

My background: I am not a pro but a skilled hobby producer with thirty years of experience. I use the K-System as well as the DR-Meters and I will give the ToneBoosters EBU Loudness Meter a try. I have learned, that the music I produce (mostly rock and folk rock) has a dynamic range of about 10 to 14 dB after proper mixing. When I use the term dynamic range I don't mean the maximum loudness difference (macro dynamic) but the difference between the RMS-value at the louder parts of the song (e.g. chorus) and the highest peaks (micro dynamic). The overall dynamic of my songs is much greater than 10 - 14 dB.
In the mastering stage I usually try to preserve this dynamic range, making only EQ-adjustments (e.g. bass cut) and a bit of sweetening. In case I use a bus compressor or multiband-compressor I apply only subtle compression without affecting the transients. In the last stage I have a very good limiter (Fabfilter) which is set to -0.3 dB FS threshold and its meter scale to K12 or K14 depending on the material. I then set the input level so that the output at the loudest parts of my songs shows 0 dB or slightly above at fortissimos.

OK guys, that is what I think about loudness war and new standards. Sorry for my "Denglish" (German English). Schooldays are long ago

[GeckoYamori]

I thought the whole point of these standards was to impose loudness normalization through as many channels as possible, in turn exposing how lifeless compressed music really sounds.

I have no insight on the legal mumbo jumbo but from what I've been told EBU and ITU are going to be implemented pretty much everywhere. For YouTube in question, I could sort of see it falling under the same legislation as TV seeing as it plays video advertisements.

[Dean Aka Nekro]

I am not at all interested in its place in audio/music creation, All it would do is compromise the individual's choice, I do not want an audiophile 'experience' and the so called "loudness wars", Well it is all a bit silly along with being a superb partisan. However for sorting out the differences in volume between a program on the television and the adverts that pepper the breaks gimme, gimme, gimme already. I do not listen to radio so do not have a clue if that is plagued with the same problem

I would conform if it became a necessity begrudgingly but still treat my own material/work exactly how I want to, Unless that happens (enforced by law to comply) then not a chance. I really believe the crusade to be a waste of time. A lot that would take me a epoch to say is covered here, Very well I must add - http://createdigitalmusic.com/2010/05/t ... d-by-fact/

It just never fails to make me smile as TC was one of the pioneers in selling people a relatively cheap solutions to begin this "loudness war" to begin with if you believe things are so bad with their finalizer range but are now once again at the front except doing the exact opposite (Still pimping a cure/the solution and without any doubt doing rather well by doing so)

That is my personal point of view/perception of it and I an not attempting to be at all disrespectful just in case it seems like I am

[GeckoYamori]

Nobody is going to force you to produce your music in a certain way. Nothing is required on your end. All this does is turn the tables on average loudness compensation during playback.

[Compyfox]

Sorry folks, I saw this yesterday but was way too tired to respond. Time for me to catch up.

Compyfox, thank you for this thread. It's a little step against the loudness war. But I fear it's a futile struggle, tilting at windmills. We have had convincing initiatives against it, e.g. the K-System from Bob Katz and the Dynamic Range classification from Tischmeyer. Now we have the EBU R128 recommendation for broadcast and more and more TV- and radio channels in Europe adopt it.

Oh, sometimes I like being a Don Quijote in the audio realm. And to be honest, it's not "that" much walking against windmills.

I actually didn't do much on that behalf (read: using the EBU R-128 meter for mixing music), I only wrote it down properly (still doing a refined v2 or something) - and now it's even debated amongst those engineers that originally created the EBU R-128 standard.

With positive response if I may add.

I think we have to educate the listeners themselves rather than the mixing and mastering engineers. They know about the problem but they have to do what their customers want.

Actually, we have to educate both fractions.

The listeners are already aware of the problem. Most popular example is the disaster with the Metallica "Death Magnetic" album (the A/B with the album and the Guitar Hero release is a prime example). And engineers finally want to get rid of overcompression as well.

The problems are the clients - they still think "louder = better = more listeners". But taking a short dive over to the charts, you realise "this can't work - my ears bleed".

Compressed content, run over a compressed stream (youtube, radio, TV) makes things worse rather than better.

Your aim is to achieve a new standard value for music of -16 dB weighted (K16 V2). That is a good idea. There should be enough headroom for dynamic music. But I also see the point that it is not so easy to produce a consistent album when setting the maximum of every song to the same level. You guys mentioned crescendos or calm sections or songs like ballads. I therefore wouldn't be so strict. I would recommend a loudness level standard for a whole album so that some songs could be louder or softer. I think the louder ones wouldn't be mastered too loud because that would destroy the balance and flow of the album. But for music distribution platforms in the internet there should be a strict standard, else wise it would be a mess to make a mixed playlist of downloaded songs.

This is what the debate over the last two pages was all about (with Sounddigger and me).

My initial idea was to focus on the MLk meter (Momentary Loudness), which uses a rise/fall of 400ms, and is what some people consider more close to the original K-System.

It turned out however, that the SLk meter gives a way better response in terms of loudness analysis. Not only over the course of a song, but also a whole album. Which is the main reason why I need to overhaul my white paper - but I didn't find some time so far.


One of my last projects was mixed in the K-System v2, and you can follow that here:
Chameleon Jazz Connection promo album "Highway to Hell" - mastered by Studio Compyfox

I used K-12 after a longer debate with the client. The SLk meter doesn't exceed +3LU on forte fortissimo parts (mostly chorus and breakdowns), the overall Loudness of the whole production (per track and for the album) is within the +/- 1LU spec (sometimes it's -1, sometimes +1). The tracks are still dynamic as hell. Even at that loudness.


This album is a prime example/proof/showcase that this standard is working.

There are some platforms like Spotify who use loudness normalisation yet and Itunes has a preference setting called "soundcheck" which is described as working in the same way as ReplayGain. Good news! I hope that more will follow. But there are other popular web pages like Youtube or SoundCloud and a lot of sites where you can publish your music for free. There, the loudness war is still taking place. So education especially on Youtube is essential. Problem is: no one cares about your education videos if you are not popular

The trick is the promotion, word of mouth, forwarding ideas.

Since my white paper was released, I usually have 100-200 hits on my page every month, just for this white paper. And I'm sure the word is spread even more on other pages. Even this KVR thread here has way over 7500 reads!

The response so far is mostly great, with very constructive critism which helped me improve certain things (see last two pages). And like I said, even the creators of the EBU R-128 standard know about this paper and debating the ideas amongst themselves. We help each other out.


Granted, education is still(!) an issue. It will always be. It would actually help if not every other company would say "make your sh*t loud with this tool" but rather advertise as "it's a tool to keep your stuff in check". Unfortunately, that won't happen. Marketing these days is focusing on "loud" and "punchy" and "analog warmth". Why else do you think that there are so many compressors and limiters released these days?


So who is really to blame?

The software/hardware companies? The mass media? The labels? The users? Or the radio stations?

I say, we all are to blame. Instead of saying "no - stop it", we tagged along.

In the mastering stage I usually try to preserve this dynamic range, making only EQ-adjustments (e.g. bass cut) and a bit of sweetening. In case I use a bus compressor or multiband-compressor I apply only subtle compression without affecting the transients. In the last stage I have a very good limiter (Fabfilter) which is set to -0.3 dB FS threshold and its meter scale to K12 or K14 depending on the material. I then set the input level so that the output at the loudest parts of my songs shows 0 dB or slightly above at fortissimos.

You pretty much understood how the K-System v1 is working, and you have no problem whatsoever using the K-System v2. With only two differences:

1) focus on the SLk meter in EBU Loudness. The amber zone for forte fortissimo parts are +3LU now - ignore the MLk meter. According to the EBU R-128 standard, it's allowed to peak (momentarily) up to +9LU (and at K-16v/K-14v2, that's totally fine). Even at K-12v2 with +7LU MLk ever so often, limiters don't barf as much as if the signal is constantly there.

2) Use a ceiling of -1dB True Peak (oversampled brickwall limiter should do)


EBU Loudness has a DR readout built in. Actually two. One is reading the loudnes range, and one the RMS to Peak value (crest factor).

So upgrading does make sense!

OK guys, that is what I think about loudness war and new standards. Sorry for my "Denglish" (German English). Schooldays are long ago

If you haven't noticed already, I'm also German. And your english is understandable. So don't worry about it.

I thought the whole point of these standards was to impose loudness normalization through as many channels as possible, in turn exposing how lifeless compressed music really sounds.

In short, yes.

I have no insight on the legal mumbo jumbo but from what I've been told EBU and ITU are going to be implemented pretty much everywhere. For YouTube in question, I could sort of see it falling under the same legislation as TV seeing as it plays video advertisements.

This is the original plan as well, but... YT is not declared as TV station (yet). So people can do whatever the funk they want. And they do.

Take either music videos, or even better, short films. They have a way different loudness span. Here education would work to clear one step of the equation. Then it's easier to implement a loudness normalisation scheme without having that "f*ck, what's going on" factor.

I mean, most video editors these days can use VST plugins. The users only need to educate themselves in terms of how to use metering tools.

But then again, if this (Loudness Normalisation on Playback) is only happening in YouTube, and people snag the videos regardless (with certain tools) - the effect is gone.

So... rather than fixing the problem at the end of the pipe, it should be considered to be fixed at the source as well.

I would conform if it became a necessity begrudgingly but still treat my own material/work exactly how I want to, Unless that happens (enforced by law to comply) then not a chance. I really believe the crusade to be a waste of time. A lot that would take me a epoch to say is covered here, Very well I must add - http://createdigitalmusic.com/2010/05/t ... d-by-fact/

That article brings up a good and valid point:
Whoever controls the mass media market, controls how things are done.

But you(!) are still in control how your music is mixed. The focus should be on the final loudness of the production. Not how a track has to be recorded/mixed.

It just never fails to make me smile as TC was one of the pioneers in selling people a relatively cheap solutions to begin this "loudness war" to begin with if you believe things are so bad with their finalizer range but are now once again at the front except doing the exact opposite (Still pimping a cure/the solution and without any doubt doing rather well by doing so)

That is my personal point of view/perception of it and I an not attempting to be at all disrespectful just in case it seems like I am

You're not disrespectful, you're right.


Though you can't completely blame it on TC electronic alone in terms of the loudness war. Limiters are still there to keep certain signals in check - that is still their purpose. However in the 90ies, certain engineers realised "we can tickle more out of our systems", and certain marketing "guru's" were like "we have to be louder than everyone else to be heard".

And from then on, the spiral started. And it was possible, since the CD technology offered that! And it was even more possible, since the DAC's in portable devices and PC's simply ignored the clipping, especially with newer revisions of CODEC's (first codecs barfed on too strong loudness, some still do).



In the end, noone is forcing you to anything (as GeckoYamori said). But you, as user, can help fight against that nonsense. Make music enjoyable again, while only having an eye on "what maximum" to shoot for.

Now what would you do if you created an awesome track, or even album, it's as loud as you want (like: modern loud), but you want to print it on vinyl? There would be a problem.


Now with the idea, or concept (however you want to call it), to drive back on loudness prior to playback systems, you can cover a lot(!) of platforms right from the start. Vinyl, HD Audio (PABD, SACD, DVD-A, HD WAV), consumer audio (MP3/AAC) - whatever. Without thinking about it like "does it even port over on other systems?!".

And this is the good thing about it.



Granted, the whole "loudness normalisation on playback" is a huge issue these days. But we musicians and engineers, can make it easier for all fractions - if we consider the loudness issue right from the beginning. And as we speak, there are more and more tools coming out that help on that behalf (like: Wavelab 8's built in Loudness Normalisation tools - as soon as WL8 is out I mean).


Example:
Instead of mixing everything at the loudest volume possible, one excellent step might be going back to using a reference level of -18dBFS (a lot of mix engineers already do so again, and certain plugins force you to do so as well). Setup once, mix to your hearts content, don't exceed 0VU/-18dB RMS.

Then on the mastering process (those that do this regulary), do not go higher than recommended (K-16v2/K-14v2/K-12v2). I consider K-16v2 as the ultimate goal. And actually iTunes also shoots for that (-16,5LUFS) with their SoundCheck scheme.

That is still 2dB louder than -18dBFS. And with the focus on the SLk ballistics and -16LUFS (or K-16v2 for that matter), we can still have dynamic passages ranging up to -7LUFS (MLk) on forte fortissimo passages. Every limiter can handle that without artifacts or pumping. And it also makes agressive multiband compression obsolete.


This, in my opinion, is the best of all worlds. And we're talking about the sound/enjoyable loudness from the early to mid 90ies again. Problem solved.

Though at least IMO. YMMV.




What we need to do:
- educate
- debate
- recommend

With who:
- interested parties
- clients
- fellows
- the record label
- friends that listen to our music content
- our children(! - they are exposed to loud noises/music, without ear protection - it's a similar widespread disease like wearing glasses, only that the outcome is much more dangerous)



As much as we debate movies, music or TV shows, we should debate loudness just as much. And not just the political aspects of that (street noise). But also the important stuff like: listening volume with earbuds and headphones, going to a concert or larger party without ear protection, etc.

We can't just sit there and do nothing, say that it doesn't affect us. If we stay like that, nothing will happen and it's the same stuff all over again.

[Dean Aka Nekro]

Edit: I do apologize it has taken me a good week or whatever to reply, Life is being...Well life for a change! Not intentionally being rude/late

I won't do an epic quote but there are parts I do agree with you on but again there is some areas that remain grey and/or not an actual problem. At least the way I see/hear them personally. A disclaimer/reminder that I'm an engineer and musician whom is a metalhead lifer so the bad press it has received in these 'loudness' debates must be taken into account as in apart from some exceptions quality has not suffered, It is not really that dynamic by nature! Some of its offspring such as post-rock/post-metal can be rather dynamic to point of it actually being very unsuitable to listening to on headphones whilst out and about without some reduction in the loud and quiet passages to make them somewhat not be near silence to deafening or at least a pull headphones off as a knee jerk instinct reaction, So I do not keep that sort of material on my iPod as it loses too much and I'd also again personally say not the sort of atmosphere drenched soundscape it is without the right environment to take it all in if that makes sense (For arguments sake I'll use Sigur Ros as a decent example). So some pieces in no particular order that you might find useful for what it is worth. First 'The Making Of Death Magnetic', Its clear that it could of and does sound like a regular modern rock recording during the tracking sessions even on youtube and not in HD:
Secondly three sort of interrelated articles in order of date I think, First an oldy from SOS: http://www.soundonsound.com/sos/sep11/a ... udness.htm
Then a sort of follow up/laymans overview from Ian whom is behind dynamic range day: http://productionadvice.co.uk/loudness- ... mic-range/ along with another short but to the point issue with the tools available:http://productionadvice.co.uk/how-loud- ... nt-enough/

Live sound, I do attend live gigs when I can do and there are laws already which dictate how much a venue can push levels without causing problems. I can not say for certain they stick to it always but honestly I've not had any problems with my hearing and regular attending 'extreme metal' live gigs. I've had the misfortune to be in some awful clubs having to put up with what junk they play to get the dance floor moving and I have heard some dreadful moments where the PA system's built-in safety limiter kicks in and the whole thing falls apart...The crowd never seemed to notice anything though

Common sense does help, Unfortunately its not that common it seems

All the best and to all as always

Dean

Just for a quick idea of how after I've finished tracking/engineering one of my own or collaboration tracks will sound you are more than welcome to listen to this brief example. Hope it is not too 'death magnetic' for anyone https://soundcloud.com/nekro-sounds-aud ... uick-sound or a finished one which is the guy I works with mixdown of a fairly recent track so it does sound different as its not the same drumkit...etc. but from the same project: https://soundcloud.com/pstevenson/scorc ... equency-of
- He seems to go for more low-end than myself but are tastes are fairly similar Cheers, Maybe I am simply not the person best qualified to debate 'loudness'

[codec_spurt]

I've been on a massive trip the last few days about this subject.

First of all. Big respects to Compyfox for taking the time. And also Jeroen for making the tools to make it all possible.

I've downloaded and read all the relevant white papers and the r128.pdf.

I bought the Klanghelm VUMT a while back and have only just learned how to use it. What a piece of work it is. A beautiful, considered piece of work. Bravo.

I currently have a 10,000 word document on the subject - about how to callibrate your meters. The difference between PPM and Vu...

40- 50 percent of it is Compyfox sharing his knowledge.

I also quote some bits by rane, so I can't make it public.

It is available to anyone that asks me for it via PM. And anyone that sees bits of their stuff in there, that they do not want in there, I shall delete. Or I'll be quite happy to make quotes and give references. But it would spoil the flow in what is quite frankly, the most concise and to the point essay on the subject. Sourced from KVR and Gearslutz. Quoting from the aforementioned Compyfox to Bob Katz to many more. Most of this stuff is public domain.

It shows what tools to use and how to use them. Read it once. And you will probably be good to go.


Dean, it's good to see you around.

Would it be possible for you to tell us briefly what VU/PPM meters you use, software wise, and how you callibrate them? And then maybe, quickly, how you would use them from tracking the guitars from recording, to mixing the thing, prepared for a final master. And a bit about mastering too would not go amiss, that is whether you master yourself or maybe job it out to someone else.

cheers.

[Dean Aka Nekro]

I'll gladly have a read as I have read a lot, It is sorely needed something concise and to the point I would say, If the audience is going to listen then compact is about as much of their own time/patience majority of people would care to spend...Sadly I will add. So if you guys could do a concise write up to spread then it has the potential that no white paper has on giving people a means to get to grips with the facts, the fiction/myths and areas left to listeners to decide to their own preferences. Cool, If you would PM me with it I would appreciate it codec my friend

As for me with VU and PPM meters, I tend to simply not have anything peaking over -12 on the peak meter whilst tracking, The RMS will obviously fall quite lower than that (Honestly I've not really given it analysis much, Maybe I should just so I know what I'm doing? So its easy to explain as well as do). That way it to me at least seems to leave plenty of headroom if/when needed later down the line along with not pushing converters to much, As Compy has mentioned and we've all either heard or experienced - When they "barf" its bloody awful and should always be avoided at all costs. If there is any outboard used during tracking then the VU meters are set to represent +4dBu = -16dBFS, Same for many plugins that work as hardware does (sometimes a trim is needed) as in the signal they expect to see for optimal results and processors/effects that are not setup that way is no problem as they to me at least do not seem to mind if they see -16dBFS to 0dBFS so it doesn't affect my workflow. That is about it really I aim for my mix downs to be around -6dBFS (peak of course!!! ), A lot of stuff I'll try my best to nail in the mix so that only minor adjustments is needed at the next stage. So yeah I do do my own 'homebrew' mastering for the sake of making a set of tracks that belong together sound cohesive but not at all the full side of mastering such as minding my p's and q's and so on as I do not release my stuff or plan to, If it ends up as going for release then it would/does certainly get sent to someone with the skills to take care of those details my/our (the dude I work with) setups/joint setup just is simply not equipped for or my Neanderthal man/early man/missing link can only do simples tbh . I'm under no delusions that I'll ever be a mastering engineer either as much as I admire the skillset it takes and you know as its stuff not intended to release there are no funds to spend on getting the full treatment, I'll never say never though
Software meter tools I use mostly are Sonalksis FreeG (for the gain/trim part also), The PPMulator+ by zplane (can't justify the upgrade to XL), PA/brainworx bx_meter, VUMT by Klanghelm (for the gain/trim part also) plus the old Elemental Audio/RND InspectorXL set sees use still, I do like the crest factor on the old Finis peak-limiter too but its not a patch on Pro-L or Limiter#6 so I do not use it for peak-limiting very often these days, Martin Zuther's K-Meter I've dabbled with and his latest one 'TrakMeter' may well go well with the next I'm going to mention. The only Loudness meter I've got is by Jeroen from the ToneBoosters plugins, I've not had a chance quite honestly to setup a mix and try to implement/place LU in the place of how I work. I will get around to trying it at the very least to give myself further understanding of how it functions outside of the manor in which I've briefly used LU to measure/analyse mix downs and mastered material. That is about it, If I send signal out the box then I aim for same as going into plugins that expect to see that same +4dBu = -16dBFS, It allows enough headroom for driving certain processors if the affect is desired I find. Obviously it might be a bit on the hot side for some but that is how I am used to operating. If the material is different from my usual stuff such as a post-rock type deal then I'll adapt by going down to +4dBu = -18dBFS and peaks drop by about 3 or 4 dB and that is due to application of delay and reverb being a big part of that sound plus synthesizers or some form of drone/root note or pedal tone may well be used so it has to fit somewhere.

One thing I have noticed which I do not know has been looked into is the stereo width of more modern material whereby it seems much louder and is padded out with more sounds at source or perhaps mix also to give it the appearance of/sonic feel of being a lot, lot more dense and full but on metering/inspection it doesn't show up as being 'louder', Anyone else noticed that out of curiosity? May go a bit of the way to explaining the fatigue that can happen to the listener? Again I really do not know the answer but am interested to hear what everyone thinks

All the best, Hope that is what you wanted codec and all the best your way and to all as always

Dean

[Compyfox]

I'll respond tomorrow as it's a bit late over here (and Dean's posts are a tad hard to read currently).

But codec_spurt, can you also send me a PM with the PDF? That'd be nice.

[Compyfox]

I'll simply double post, just to let the readers know that there is an actual response. I hope that is fine - won't do that often but I was way too tired yesterday.


@Dean:
I totally understand your points in terms of loudness, or better said "uber-dynamic" productions on an iPod/iPad/iPhone or whatever playback system you are using. I have the same issues with my HTC while I'm outdoors.

Either I have to constantly reach for the volume knob (since it's either too quiet or too loud), or certain tracks simply drown in the environment noise. Especially classic stuff. So the logical step: reach for the volume knob. Have a loud section, reach for the volume knob, change the environment, reach for the volume knob. That game goes on.


There are several solutions to this dillema:

a) each track needs to be treated dynamically for mobile playback. Meaning: compression. But this is a ton of work and mostly also involves re-encoding.

b) in order to have a consistent loudness, we need some sort of loudness normalisation. Most of the time, that is ReplayGain schemes.

c) even then, there should be a dynamic control system integrated (compression), according to the environment setup. But this is not integrated in playback devices yet.




As I was still not too knowledged of that whole ReplayGain thing, I sat down the other week and did some closer look/analysis. Especially if I want to use the normalisation schemes in Wavelab 8.


It turns out, that ReplayGain is also a two edged blade.

First and foremost, the ReplayGain uses a reference level of -14dBFS, setup with an output level of +89dB SPL (c-weighted). Much like the K-System's setup. But the analysis timeframe is not 3 seconds (like EBU R-128), or 300ms (like standard VU/RMS meters), but 50ms. The official specs about ReplayGain (HydrogenAudio Wiki) simply draw their ideas from... somewhere (IMO!), but not based on working facts (again, IMO).

So if you have ReplayGain activated in your player, and you have "auto analysis" going on, the volume jumps are drastically. And due to the ballistics, it's also mainly focusing on peak values rather than average loudness. This can be clearly seen in the Sound-on-Sound article.



Now... you can use tools to measure in EBU R-128 and write the appropriate tags for ReplayGain. But this is also a two edged blade currently.

For example:
The official reference level of EBU R-128 is -23LUFS. Most tools that can't shift the reference level will stick to that. The ID3 tags for ReplayGain however are written after the Integrated Loudness analysis. While this is actually a good and working thing for a complete album (or a certain stream), it is not for individual tracks (here, shifts of up to +/-4LU can happen).

So not only do we have an extreme low volume setting (and most playback devices only have a certain maximum output volume and then also a very limited battery life), the loudness shifts will still be drastic in some cases.



As we can clearly see - both systems are not yet ready for fixing the issues with portable players (especially) or home stereo systems. We still need to reach for the volume knob - though not as much as without the analysis. There are still missing links.

That is one issue.




The other issue is:
When is loud too loud? Or when is quiet too quiet?

We have to find a middleway in terms of engineering. And yes, drastically limiting and compressing a signal while production doesn't help. Our ears like the transients - they say "yo, I have the impression that this thing is dynamic (as in form of 'punchy')". Cut them away, and you have a tonal porridge. Definitely not good.


I still say that a good worklevel (reference level), may it be -18dBFS, -20dBFS or even -14dBFS if you're up for loud mixes (though I do not recommend that!), is a good starting point. And then go from there. You can always squash stuff to sh*t in the end.

But if it's squashed already, you usually can't fix it. And dynamically limited (compressed) material might be easier on the ears in loud environments. But it's not sounding better or louder than a track with retained transients.

At least in my opinion - YMMV.




Which brings me to the "Death Magnetic" dillema.

The problem wasn't that the production was crap to begin with. Hell no, I know the Death Magnetic documentary, I know the Death Magnetic Multitracks from Guitar Hero. All is fine in there, and pretty much what Metallica did in the last decades. With a ton more lowend, since CD's and MP3s are not limited in that section compared to Vinyl. And they sticked to reference levels both while mixing and recording (just take a look at the faders and certain peak readouts in the videos - that gear is calibrated to -18dBFS!).

The problem was the raised loudness - or better said, the actual mastering. Some bloke was like "this thing has to be loud", and they pushed it to their limits. And at that point, it fell apart. Even more so while playing on the radio (even more compression applied).

The album should be officially rerelased, and not by fans with the GH:Metallica mix. Then it's like it should have been.


Same with the last two Albums by Johnny Cash... the production is okay (though I can't stand the chain rattling in pretty much every song), but the finished CD is pushed too much into it's limits. So we can not blame Rick Rubin, we have to blame (in this particular example) the recording industry and the accustomed listener base.






SUMMARY:
Tracks that are mixed/mastered dynamically are too quiet on playback in loud environments. Here, dynamically limited tracks definitely prevail - but fall apart at home stereo envrionments.

ReplayGain was always considered as a step towards the right direction in term of unified loudness on playback, but ultimately fails due to crazy setup ballistics and reference levels that make no real sense for music or playback (IMO).

EBU R-128 measurement in combination with ReplayGain tags/metadata do work better (IMO), but they are currently focusing on the Integrated Loudness only. While this works for the "album" tag, it doesn't work for the "track" tag. In my opinion, the ReplayGain tags need two additional metadata blocks for proper playback analysis: "Reference level used" and "Metering System used". Only then, we can think of a consistent and enjoyable level on playback, since the player knows "oh, the tracks have a reference level of -14dBFS, I'm setup to -18dBFS - the ballistics for real time analysis is 3ms with k-weighted filters, else I simply shift everything by -4dB and trust the mastering".

And this is what the "Music Loudness Alliance" is actually trying to do.




Is there a suitable solution yet?
Unfortunately not. But at least already steps towards the right direction. And the K-System v2 does contribute to that.





SIDENOTE:
I definitely need to find some time to update my white paper. Especially with the ReplayGain thing. And if I have enough time left over, I can start that Metering series on my techblog that I'm constantly talking about since the beginning of this year.

[codec_spurt]

As for me with VU and PPM meters, I tend to simply not have anything peaking over -12 on the peak meter whilst tracking, The RMS will obviously fall quite lower than that (Honestly I've not really given it analysis much, Maybe I should just so I know what I'm doing? So its easy to explain as well as do). That way it to me at least seems to leave plenty of headroom if/when needed later down the line along with not pushing converters to much, As Compy has mentioned and we've all either heard or experienced - When they "barf" its bloody awful and should always be avoided at all costs.

There is that figure again - -12. I take it you mean -12dbFS?


This is almost like a magic number. Again and again it crops up.

I think the main difference between what we do is that you track live stuff like guitars, and though I do do that occasionally, I haven't got that far yet.

Perhaps it might be related to the K-12 scale?


I'll send a copy of my notes when I am done, to you, Dean.


cheers.