Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Advanced loudness normalization (Read 8830 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Advanced loudness normalization

I'd like to share a new whitepaper that discusses loudness normalization techniques. The whitepaper has been posted to Music Loudness Alliance website and describes what we believe are improved techniques for loudness normalization. The techniques are applicable to computer and portable file-based playback. For portable applications, hearing loss prevention (as legally mandated in some countries) is integrated with normalization.

We believe BS.1770-2, the revised album normalization, variable target level and the hearing loss prevention features are potentially useful improvements for ReplayGain. With the publication of this paper, these ideas are being put into the public domain.

The MLA is interested in sharing these ideas to improve loudness normalization implementations to the extent that they may enabled by default, or even always on, in players. Intelligent normalized audio playback improves the listener experience, can help protect people's hearing and eventually may lead to improvements in mastering and production quality i.e. the loudness war.

We are interested in any comments (technical or otherwise) and would like to work with the Hydrogen Audio community to help further the adoption and use of loudness normalization in players.

Advanced loudness normalization

Reply #1
There are lots of good ideas in here, and I hope it does well.

Album level, BS.1770-2 (EBU R128), on by default - these are all good.


I'm concerned that many audio systems don't have a single digital gain control, but have some analogue gain as well (or multiple digital gains). Where this is true, it's wrong to assume that 0dB FS at a given point in the chain represents the loudest sound a system will ever produce. It would be nice to integrate the loudness normalisation into the only volume control (I suggested something like this in 2001), but you've got to anticipate that this won't always be possible.

Where it is possible, I don't like the idea of giving up the loudness normalisation in an effort to avoid peak limiting or compression (pages 6-8). I'm thinking about how this would work on my portable mp3 player. I'm listening to a classical track that's basically quiet but has a few loud peaks. These loud peaks mean that the top part of my volume control effectively doesn't work. I visibly change the volume control, and the volume doesn't audibly change - because those peaks (which might not even be in the part of the album I'm listening to at the moment) are preventing it from going higher. Then I skip to a pop track (I got bored with the classical track - I couldn't turn it up loud enough to hear the quiet parts!) and BANG the volume control is working and the level jumps up by 12dB. You could put some logic in to prevent the jump, and/or visibly lock out the top of the volume control range while the classical track was playing - but either way I still can't turn up the quiet parts of my classical track (or the quieter parts of any dynamic album) to listen to them properly.

If, instead, you allow peak limiting, then if I really want to turn my mp3 player up that loud (maybe because this part of the album isn't that loud!), I can. You could solve the concerns over the audio quality when the limiter acts in a couple of ways - either restricting the range over which you can push it, or putting a "limit" indicator on the display when the limiter kicks in - people would notice that it sounded better when you reduced the volume in this case. (though this only happens during the loudest parts with the player absolutely maxed out - there might be an expectation that it won't sound as good maxed out, so maybe nothing needs doing)


I like the approach to hearing protection - though given that headphones are replaceable items with hugely varying levels of sensitivity, if the hearing protection is finally something sensible which I might want to keep in place (rather than the current idiotic hard limit which, if it cannot be defeated, means I won't be buying the mp3 player!), it might be worth exploring how headphone sensitivity can be reflected. Maybe dial the sensitivity level into the player, and adjust the calculations accordingly?

If it worked sensibly and didn't get in the way, users in the rest of the world might choose to enable the hearing protection. Some might value the display, even if they disabled the absolute limit.


I don't think getting the portable players to calculate the values is a great idea, but I could be wrong. I don't always have full albums on my cheap little player. Many people don't even own full albums - though I suspect that, for the non picky, just calculating the value over the tracks they do have, while not "right", is probably good enough.


Finally, I doubt audiophiles will accept "always on". "On by default" is fine, but some of them will insist on using the full range of their DAC.


Hope this is some help. What you have is basically a great idea. If you can take away the bits that might cause some people concern, the only thinking stopping you is inertia. While there are valid reasons to object, even trivial ones, people can hide behind those.

Cheers,
David.

P.S. Think about metadata, and how you'll market this solution.

P.P.S. If you introduced some kind of dynamic range management at the same time, you could keep everyone happy. Don't forget that Dolby essentially got all this worked out for movies many many years ago.

EDIT: EBU R128 / BT.1770-2 is slowly taking over the world:
see right panel here: http://tech.ebu.ch/Jahia/site/tech/cache/o...dness-faq-radio
see first external link here: http://en.wikipedia.org/wiki/Commercial_Ad..._Mitigation_Act

Advanced loudness normalization

Reply #2
Thanks for your comments David. Very helpful.

With improved performance of DACs, and advent of digital user interfaces, the variable analog gain stage is seen much less frequently now especially in portable players. Even where this is not the case, it's hard to argue that multiple gain controls for the same signal is of value to the user. The proposal challenges designers to make it look like there is one control even if there are multiple. Either they'll fail and (hopefully) we'll be no worse off or they'll succeed and the listening and human-machine experience will be improved.

I believe there is room for improvement for how the the control responds towards the top of the range. We think, however, the best solution is to avoid the problem altogether and build systems with more headroom. The hearing protection features make this safe to do. One thing we did not want to do is try to tackle the dynamic range management issues. We recognize that dynamic range management in the player is useful for playback in noisy environments etc. but defining how that should work is not something we could take on at this time. You can argue that specifying a limiter is a more tractable problem but based on crappy limiter implementations we see in players, we decided the best approach was to make the system linear.

The focus of the HLP provisions is more on meeting requirements of current (EU) legislation than on actually protecting hearing. True HLP is another piece that we considered but decided we could not, in good conscience, take on. Current legislation calibrates SPL based on factory earbuds. There's no known solution to the loophole of using sensitive aftermarket transducers. While a configurable sensitivity parameter is an interesting improvement. We worked hard, however, to reduce the system to a single knob.

There are no hard requirements specified in Annex 4. The purpose the annex is to make player designers aware that there is no authentication mechanism in many metadata systems. Blindly trusting metadata can enable a different kind of loudness war where content providers fudge loudness metadata to make their media more compelling.

Advanced loudness normalization

Reply #3
I have been in the vinyl mastering business from 1972 and took up CD mastering in 1986.
Since the late 1990's it has become more and more frustrating.
Clients come in and ask to preserve their dynamics. Amazingly the next question usually will be: "Can't you make it any louder?"
No argument helps, the (unfounded) fear of being not as loud as "the competition" on the radio or any other medium leads to masters that are not our, nor the musicians choice.
Comparing to any other unrelated musical genre forces us to threat ballads just like grunge rock.
And even if we stick to our guns and deliver a decent sounding master, some 25 year A&R manager at the record company will sometimes reject the mastering demanding a musically completely ruined over-compressed and even clipped distorted master.
The general public is by now so used to this aggressive distorted crap that we can almost speak of a lost generation.
The Loudness War is far from over, there is a growing awareness, but the fear of the artist is bigger than the courage to resist.
This has now been dragging on for decades, unless Loudness normalization must be implemented by law also on the radio and in every possible music player, portable or in the home, there will be no end to the Loudness War.
If Sound-check, Replay gain or any other system should be ON by default, the situation would already improve enormously.
When will Apple, Microsoft and the others finally take steps? How can we convince or even force them?
Why is the Dynamic Range of a live song on TV DR11 and of the same song on CD or iTunes DR6?
I have devoted my life to better quality in music reproduction and now that we have better tools music quality is worse than ever before.

Advanced loudness normalization

Reply #4
If Sound-check, Replay gain or any other system should be ON by default, the situation would already improve enormously.


I doubt it. I think most people will rather think that this "thingy" is broken and makes everything too quiet and even destroys "sound quality". They might even blame volume cap laws and they'll post instructions on internet forums to disable the "crippling volume limiting setting".

I don't think the loudness war can be won without some amount of education. For me, the test to perform is to crank up the volume. If you feel completely immersed in the music, if the drumming makes your body vibrate, then the recording is good. If everything sounds flat and "gooey" and it gives you a headache after a minute, or if the drumming makes your headphones / speakers sound like they're made of cardboard paper, then it's too fucking loud. Compare a good recording and a bad recording at high volume, and the difference in actual sound quality will be more evident (IMO).

Edit: there's already a lot of misunderstanding about what Replaygain does, particularly in audiophile circles. Notably, that it "kills dynamics".

Advanced loudness normalization

Reply #5
I have been in the vinyl mastering business from 1972 and took up CD mastering in 1986.
Welcome to Hydrogenaudio. Glad to see a mastering pro on this forum (where dynamics are still much appreciated)!

Advanced loudness normalization

Reply #6
If Sound-check, Replay gain or any other system should be ON by default, the situation would already improve enormously.


I doubt it. I think most people will rather think that this "thingy" is broken and makes everything too quiet and even destroys "sound quality". They might even blame volume cap laws and they'll post instructions on internet forums to disable the "crippling volume limiting setting".

I don't think the loudness war can be won without some amount of education. For me, the test to perform is to crank up the volume. If you feel completely immersed in the music, if the drumming makes your body vibrate, then the recording is good. If everything sounds flat and "gooey" and it gives you a headache after a minute, or if the drumming makes your headphones / speakers sound like they're made of cardboard paper, then it's too fucking loud. Compare a good recording and a bad recording at high volume, and the difference in actual sound quality will be more evident (IMO).

Edit: there's already a lot of misunderstanding about what Replaygain does, particularly in audiophile circles. Notably, that it "kills dynamics".


I think I will have to explain what "Loudness Normalization" (like Sound-Check) does: The system reads the file's metadata and brings all tracks to the same loudness.
This can even mean that an over-compressed track sounds lower than a decently mastered one, keeping the dynamics of the latter. Read: http://music-loudness.com/PDFs/Loudness_Al...er_final_v1.pdf
If Sound-check or Replay gain is on by default most people won't even notice it is on and enjoy the music at the same loudness without having to reach for their volume control between tracks.

Educating the general public wil not help, or very little. If have been trying to educate my clients for more than 10 years now, only to find myself with less work because I do not want to get into mastering with less than DR6 on any kind of music. Most clients agree that the general quality has gone down year after year, but are to afraid to be the first to resist this crazy loudness battle.
When will the big names in the music business take a stand and have their next album recorded and mastered with decent dynamics and prove that it will sell as wel or even better.

If audiophiles say loudness normalization "kill dynamics", the are sadly very misinformed.



Advanced loudness normalization

Reply #7
I have been in the vinyl mastering business from 1972 and took up CD mastering in 1986.
Welcome to Hydrogenaudio. Glad to see a mastering pro on this forum (where dynamics are still much appreciated)!


Thanks. Those that really appreciate dynamics seem to be a small minority, but we will have to stick together to fight the battle.
Best regards,
Paul Van der Jonckheyd
Flemish Masters
http://www.foon.be

Advanced loudness normalization

Reply #8
I think I will have to explain what "Loudness Normalization" (like Sound-Check) does


Not here, you don't. FYI, the author of Replaygain is an active member on this board. Did I sound like *I* misunderstood the concept? I thought it was clear that I was pointing out widespread misunderstanding of other people.

And yes people will complain that their device isn't loud enough with some headphones, or when connecting their DAP to a hi-fi or their car stereo
  • , and thus will look for ways on the internet to disable that "crippling" feature. People routinely report that "problem", which is even more pronounced on EU DAPs that already come with volume caps that can't always be disabled (like iPods).

  • Most DAPs probably output only 1V at max volume, which is half that of a proper line-out. So if you further reduce the signal by 6dB, or even 12dB (which is not at all uncommon), the problem gets significantly worse.

 

Advanced loudness normalization

Reply #9
And yes people will complain that their device isn't loud enough with some headphones
...even I do that!  The solutions discussed earlier in this thread would stop this happening though. It's true that a simple ReplayGain + clipping prevention + an arbitrarily low target level that's independent of the volume control can stop you from turning the volume up, and force you to listen to some tracks far quieter than they were delivered. However, a target level that's integrated into the volume control, plus a peak limiter, will (if you push it, and the device allows you to push it) allow you to make your music even louder than it started out (though it'll sound truly awful if you do). A device configured like that, with sensible limits and useful headroom, will never sound quiet at full volume. Whereas mp3s with ReplayGain applied at 89dB can do on some devices.

Cheers,
David.

Advanced loudness normalization

Reply #10
I think I will have to explain what "Loudness Normalization" (like Sound-Check) does


Not here, you don't. FYI, the author of Replaygain is an active member on this board. Did I sound like *I* misunderstood the concept? I thought it was clear that I was pointing out widespread misunderstanding of other people.

And yes people will complain that their device isn't loud enough with some headphones, or when connecting their DAP to a hi-fi or their car stereo
  • , and thus will look for ways on the internet to disable that "crippling" feature. People routinely report that "problem", which is even more pronounced on EU DAPs that already come with volume caps that can't always be disabled (like iPods).

  • Most DAPs probably output only 1V at max volume, which is half that of a proper line-out. So if you further reduce the signal by 6dB, or even 12dB (which is not at all uncommon), the problem gets significantly worse.


Sorry, I see what you mean.
And yes, they will have to turn their volume up. That might be a problem with some players.
But I' am convinced Loudness "Normalization" is the only way out of this war.
Loudness has always been a big issue, also in the vinyl days, but reducing the dynamics of our music below that of a wax cylinder is all to silly.

Advanced loudness normalization

Reply #11
EBU R128 / BT.1770-2 is slowly taking over the world


Doesn't that standard have a target of 84dB, a full 5dB lower than Replaygain? Which means that "I'm With You" by Red Hot Chili Peppers gets attenuated by 17.36dB (!), which in the world of volume capped iPods, is bordering on insanity.

The solutions discussed earlier in this thread would stop this happening though. It's true that a simple ReplayGain + clipping prevention + an arbitrarily low target level that's independent of the volume control can stop you from turning the volume up, and force you to listen to some tracks far quieter than they were delivered. However, a target level that's integrated into the volume control, plus a peak limiter, will (if you push it, and the device allows you to push it) allow you to make your music even louder than it started out (though it'll sound truly awful if you do). A device configured like that, with sensible limits and useful headroom, will never sound quiet at full volume. Whereas mp3s with ReplayGain applied at 89dB can do on some devices.


Apologies, I should have read the whole thread before answering. Integrating normalization into the volume control sounds like a neat idea (if I understand it correctly), but how would that be much different than attenuation in the digital domain, and conversely, disabling tag-based Replaygain on the playback device in order to get maximum volume?

I simply fail to see any technical solution to a cultural problem, which wasn't there in the 80s. Loudness normalization is not something that content producers and consumers want. The former (artists and producers) will keep competing on making their recording sound as loud as possible, and the latter (clueless consumers) will keep bitching that the recording is too quiet. Most of the time, I feel like it's a lack of taste (especially from artists themselves!) and education.

I remember reading such complaints from Steven Wilson fans on his facebook page. The artist mastered his album himself, and made a point of it to retain a decent amount of dynamic range (the result is a recording that sounds great, particularly at full volume). His fans loved the album but failed to see the benefit of his choices, and even saw it as a problem (and went so far as to actively complain about it!).

The only solution that I see, is to have all parties listen to a great recording on a good system, and the same material but with all the usual excessive compression and peak limiting applied, hoping that they will realise that some amount of dynamic range makes the recording sound a lot punchier, in contrast to what almost everyone has come to conclude (for all the wrong reasons). I have no idea how to do that on a massive scale, however.

Advanced loudness normalization

Reply #12
Again there's also the problem of the inadequacy of high dynamic range material in noisy environments. A solution to that, as mentionned earlier, is a live DRC DSP, but then I don't see people using it appropriately (enabling and disabling it whenever it makes sense), and mostly I really don't see content producers leaving that up to the consumer (because that would mean losing control over their competitive edge).

Edit: perhaps a notable counter-example, is Daft Punk's latest album, "Random Access Memories", which is number one pretty much everywhere. It has an album gain of just -6.07dB, and while its DR8 score isn't all that impressive, it sounds really good, nothing like a smashed RHCP album. And in this particular instance, everyone is raving about how good it sounds.

Advanced loudness normalization

Reply #13
EBU R128 / BT.1770-2 is slowly taking over the world


Doesn't that standard have a target of 84dB, a full 5dB lower than Replaygain? Which means that "I'm With You" by Red Hot Chili Peppers gets attenuated by 17.36dB (!), which in the world of volume capped iPods, is bordering on insanity.
You have to deal with the volume capping in an intelligent way at the same time. Say the volume cap is 85dB. Give the user free-run of their volume control, and put a limiter at 85dB. Job done. The benefit is that "quieter" masterings can actually be turned up to 85dB now. This is much better than the present implementations, where you have a digital full scale = 85dB limit, meaning quieter recordings (i.e. those with some headroom!) can't even be turned up to 85dB.


Quote
The solutions discussed earlier in this thread would stop this happening though. It's true that a simple ReplayGain + clipping prevention + an arbitrarily low target level that's independent of the volume control can stop you from turning the volume up, and force you to listen to some tracks far quieter than they were delivered. However, a target level that's integrated into the volume control, plus a peak limiter, will (if you push it, and the device allows you to push it) allow you to make your music even louder than it started out (though it'll sound truly awful if you do). A device configured like that, with sensible limits and useful headroom, will never sound quiet at full volume. Whereas mp3s with ReplayGain applied at 89dB can do on some devices.


Apologies, I should have read the whole thread before answering. Integrating normalization into the volume control sounds like a neat idea (if I understand it correctly), but how would that be much different than attenuation in the digital domain, and conversely, disabling tag-based Replaygain on the playback device in order to get maximum volume?
You need to think about the user experience when you have (invisibly) cascaded the original audio track + the ReplayGain adjustment + the user's volume control + some limiting/compression to handle content above 85dB (EU) or above the limit of the player's DAC/amp (rest of world). The result is that the clipressed recording can be made to play just as loud in this scenario as it would without volume normalisation. Quieter content can be made to play even louder than it would without volume normalisation. There's no loudness loss.

Quote
I simply fail to see any technical solution to a cultural problem, which wasn't there in the 80s. Loudness normalization is not something that content producers and consumers want.
Nobody wanted seatbelts mandated either. But as that was people, rather than music, being killed, they weren't give  the choice.

Quote
The only solution that I see, is to have all parties listen to a great recording on a good system, and the same material but with all the usual excessive compression and peak limiting applied, hoping that they will realise that some amount of dynamic range makes the recording sound a lot punchier, in contrast to what almost everyone has come to conclude (for all the wrong reasons). I have no idea how to do that on a massive scale, however.
You have volume normalisation on by default.

Cheers,
David.

Advanced loudness normalization

Reply #14
Thanks for your comments David. Very helpful.

With improved performance of DACs, and advent of digital user interfaces, the variable analog gain stage is seen much less frequently now especially in portable players. Even where this is not the case, it's hard to argue that multiple gain controls for the same signal is of value to the user. The proposal challenges designers to make it look like there is one control even if there are multiple. Either they'll fail and (hopefully) we'll be no worse off or they'll succeed and the listening and human-machine experience will be improved.

I believe there is room for improvement for how the the control responds towards the top of the range. We think, however, the best solution is to avoid the problem altogether and build systems with more headroom. The hearing protection features make this safe to do. One thing we did not want to do is try to tackle the dynamic range management issues. We recognize that dynamic range management in the player is useful for playback in noisy environments etc. but defining how that should work is not something we could take on at this time. You can argue that specifying a limiter is a more tractable problem but based on crappy limiter implementations we see in players, we decided the best approach was to make the system linear.

The focus of the HLP provisions is more on meeting requirements of current (EU) legislation than on actually protecting hearing. True HLP is another piece that we considered but decided we could not, in good conscience, take on. Current legislation calibrates SPL based on factory earbuds. There's no known solution to the loophole of using sensitive aftermarket transducers. While a configurable sensitivity parameter is an interesting improvement. We worked hard, however, to reduce the system to a single knob.

There are no hard requirements specified in Annex 4. The purpose the annex is to make player designers aware that there is no authentication mechanism in many metadata systems. Blindly trusting metadata can enable a different kind of loudness war where content providers fudge loudness metadata to make their media more compelling.


This last paragraph is one of my concerns as well. I someone can tamper with the loudness metadata a new loudness war will start all over again!

Advanced loudness normalization

Reply #15
Again there's also the problem of the inadequacy of high dynamic range material in noisy environments. A solution to that, as mentionned earlier, is a live DRC DSP, but then I don't see people using it appropriately (enabling and disabling it whenever it makes sense), and mostly I really don't see content producers leaving that up to the consumer (because that would mean losing control over their competitive edge).

Edit: perhaps a notable counter-example, is Daft Punk's latest album, "Random Access Memories", which is number one pretty much everywhere. It has an album gain of just -6.07dB, and while its DR8 score isn't all that impressive, it sounds really good, nothing like a smashed RHCP album. And in this particular instance, everyone is raving about how good it sounds.


This is indeed a good sounding album, with DR's between 6 and 9, but still with a lot of inter-sample overs, which I can hear very clearly distorting in some places on my studio-monitors. It maybe would have been a great album with a DR of around 11 or 12 and -1dBFS for headroom. Normalized Level Control could then make up the extra gain people want?

Advanced loudness normalization

Reply #16
... This is indeed a good sounding album, with DR's between 6 and 9, but still with a lot of inter-sample overs, which I can hear very clearly distorting in some places on my studio-monitors. ...


I was under the impression that audible digital overs are due to the ADC designer not having allowed enough headroom in the reconstruction filters and/or analogue stages. (Ideally, there shouldn't be any inter-sample overs in the first place, but...) If I'm wrong, I'd appreciate being (politely) corrected.
Regards,
   Don Hills
"People hear what they see." - Doris Day

Advanced loudness normalization

Reply #17
There's no evidence that these distortions are due to ISOs and not introduced during production/mastering.

Advanced loudness normalization

Reply #18
I'm concerned that many audio systems don't have a single digital gain control, but have some analogue gain as well (or multiple digital gains). Where this is true, it's wrong to assume that 0dB FS at a given point in the chain represents the loudest sound a system will ever produce. It would be nice to integrate the loudness normalisation into the only volume control (I suggested something like this in 2001), but you've got to anticipate that this won't always be possible.


FWIW (and assuming I understand you!), in pretty much any portable or PC situation (although not necessarily something like HDMI), you will always have both digital and analog gain under you control. So in practice, you can make something like this work for most things outside of home theater.

I'm thinking about how this would work on my portable mp3 player. I'm listening to a classical track that's basically quiet but has a few loud peaks. These loud peaks mean that the top part of my volume control effectively doesn't work. I visibly change the volume control, and the volume doesn't audibly change - because those peaks (which might not even be in the part of the album I'm listening to at the moment) are preventing it from going higher. Then I skip to a pop track (I got bored with the classical track - I couldn't turn it up loud enough to hear the quiet parts!) and BANG the volume control is working and the level jumps up by 12dB. You could put some logic in to prevent the jump, and/or visibly lock out the top of the volume control range while the classical track was playing - but either way I still can't turn up the quiet parts of my classical track (or the quieter parts of any dynamic album) to listen to them properly.


What we do in rockbox on portable players is to apply replaygain digitally, and then allow analog gain above 0dB (that is, gain that would push a FS signal above the analog levels supported by the amp).  In this way a user can apply digital correction to avoid clipping in lossy formats and to perceptually normalize volume, but a user can still pump up analog gain to compensate for digital attenuation.  In practice this works somewhat like how you are proposing, you can pump up the volume on quiet tracks if you want, and peaks are (possibly) clipped in the analog domain which is usually mild or at least tolerable.