[sideshowb]

We are already starting to win the loudness war now thanks to iTunes SoundCheck. I'm not an iTunes user but as far as I understand it, it causes iTunes to adjust all tunes in a mix to the same perceptual loudness. If everyone adopted this or a similar technology (such as the open ReplayGain) there would be no incentive to master loud any more, for albums at least, because the player will only turn it down again meaning the net result is only loss of quality. Hopefully iTunes support for this tech is a turning point in the loudness war.

[SwellJoe]

That can't actually account for perceived loudness, which modern mastering processes do account for. The absolute amplitude of a song has always been fixed, by the technology delivering the audio. Radio, the primary medium for music discovery throughout most of our lives, has a very hard upper bound (set by both the technology and the FCC), thus a compressor/limiter is employed at the final stage before sending the audio out of the radio station to go up the tower and out over the waves. There has always been that sort of leveling going on.

Modern tools provide an entirely new dimension in the form of multi-band compression, digital phase alignment, etc. It is now possible (and being done in nearly every genre) to make a recording perceptively louder than other recordings by maximizing amplitude in specific bands (those humans are most sensitive to), reduce phase cancellation between speakers, and hype the sound (boosting high and low frequencies, which tricks the ear into hearing it in the same way as louder music...but also causes listener fatigue faster), often all at once.

Amplitude compression, even when it's smart enough to recognize that there's more activity across a broader spectrum as provided by SoundCheck, does nothing to restore the damage to dynamic range, natural frequency response curves, and "real" sounding recorded music. The music is broken by these processes...the listener has no power to fix it, other than to not buy it, and choose music that hasn't been mutilated in such a way.

[sideshowb]

The main weapons in raising perceived loudness are brickwall limiters such as Waves L3 which raise RMS (root mean square) amplitude while keeping peak levels constant. I think soundcheck is based on RMS, please correct me if I'm wrong - if it were based on peak levels then as you say it would indeed be useless.

I don't know how much or whether the other technologies you mention really help to boost perceived loudness much beyond what is measured by RMS. Multiband compression raises RMS. Boosting high and low frequencies (aka "bare fat bass and mad amounts of high end" [1] - beyond what is wanted for a good sound) could be defeated by measuring RMS based on equal loudness/frequency curves. I've never really used the other things you mention but whatever they are they can be defeated by technology that measures their effect in the listener device. That's if they can cheat RMS anyway - I'm not sure they can but if you do have data on that I'd be interested to see it.

As you say, sound quality is still lost, but if normalization (over the album length where necessary, of course) becomes the defacto standard in playback technology then the incentive for making bad quality/loudness tradeoffs in album mastering is gone. The standard if adopted would restore the dynamic range by taking away the engineers' incentive to compromise it. The engineers would doubtless breathe a sigh of relief as most of them are more bothered about this than we are.

Analog radio would still be mastered stupidly loud. I would assume final stage compression in radio broadcast is keyed on peak not RMS level, again please correct me if I'm wrong. But it's a dying medium anyway.

[1] https://www.youtube.com/watch?v=7nlzwDfxVSg

[SwellJoe]

That's if they can cheat RMS anyway - I'm not sure they can but if you do have data on that I'd be interested to see it.

RMS is a measure of power, not of perception. You don't have to defeat RMS to defeat perceived loudness.

But, you're right that a playback device could take measures to defeat the perceived loudness baked in (though they'd get quite complex, as the tools for baking it in are extremely complex these days, and I don't actually understand half of them, despite having studied audio in college and worked in the industry). I don't know if it would actually improve listener experience to do so, though. Ending the "Oh shit that's loud!" and "Why is this song so quiet?" problem would be positive for listeners, of course. But, at what cost?

"Analog radio would still be mastered stupidly loud. I would assume final stage compression in radio broadcast is keyed on peak not RMS level, again please correct me if I'm wrong. But it's a dying medium anyway."

Analog radio (and digital broadcast radio, as they still compress, despite some of the technical reasons for doing so being gone, inertia is strong in broadcast; I have noticed that Pandora and Spotify do not do any sort of compression, however, which is nice...but also annoying when the playlist has new and old music as the difference can be massive and unnerving) was among the earliest adopters of various technology to make music perceptively louder. Aural Exciters (a very early salvo in the loudness war) were available in a broadcast targeted version from very early on, etc.

I haven't been in the broadcast industry in a long time, and I worked in television rather than radio (though the station I worked for shared a tower with several radio stations and another TV station), but I'm reasonably confident radio stations in major markets still tend to have the most modern "make it loud!" devices available. Loudness=listeners in radio. That's why the loudness wars are happening.

[sideshowb]

I agree with most of these points. I should have been clearer in what I meant, RMS is a measure of power and a proxy for perception - according to you, a worse proxy than I thought, so thanks for educating me on that point.

"I don't know if it would actually improve listener experience to do so, though" - except inasmuch as it would remove the incentive to master albums too loud. And if it were the default in all playback devices (digital radio, codecs) it would also remove the incentive in broadcast. The only catch is it relies on the broadcaster not owning the means of playback, which with internet radio they sometimes do. But if it's someone like Spotify they are in a position to prioritize sound quality - are people really going to leave Spotify because it's too quiet when the thing has a volume control you can just turn up if you want? It's not like surfing analog radio channels used to be.

[clay_to_n]

Though it also has drawbacks, especially when listening to full albums - suddenly the ambient interlude track in the middle of an album is sounding as loud as the loud tracks around it.

[FreeFull]

Can't it be set to have a single per-album volume setting?

[KeytarHero]

> Hopefully iTunes support for this tech is a turning point in the loudness war.

Don't hold your breath - iTunes has had Sound Check since 2002.

[sideshowb]

Damn. But I can still hope.