Waves Audio, engineers, and producers on the evolution of the Loudness Wars

[EERecordist]

Waves Audio is an early and continuing successful digital audio plugin maker. They made this short chatty documentary of the evolution of the loudness wars.

Howie Weinberg • Bob Katz • Dale Becker • Gavin Lurssen • Reuben Cohen • Jesse Ray Ernster • Waves Co-founder & CTO Meir Shashoua chat with examples of artists, genres, eras and the specific Waves plugins L1, L2, L3, and L4. I was irritated by one quote "that it sounds like it sounds on the radio" is a good thing. The theme is competition. They also briefly discuss streaming.

A vehicle is a terrible listening room with a very high noise floor. Headphones and earbuds are much better for noise floor, same concerts and DJ events, but they are still subject to unnecessary loudness maximization. One engineer points out that today much of the loudness compression is done in at the mixing stage, far before final mix then mastering.

It would be nice to be an article with references, including loudness studies of released tracks and a technical comparison of the streaming platform normalization.
Still it's a fun short.

 

[jEDGEc0m]

Who's in charge of the volume knob at the very end? WE ARE! so stupid isn't it. when californication came out, we just turned down the dial, and were "rewarded" with still thinking our equipment might have been broken. yet it wasn't our fault at all.

It's still saddening how widespread super processed ultra loud so much audio is even now with loudness normalisation being widespread.

Some modern productions I guess do a better job of sounding good despite being so limited, and low end is generally better/stronger these days, i guess. maybe. but doing that without ultra limited processing would only be even better.

It is kinda depressing. occasionally there's a remaster that is actually an improvement, but that's probably an exception rather than more common.

 

[dasdoing]

I never realized that the limiter was introduced in the digital world.

They urgently need to fix LUFS. I had Bob Katz once replying directly to me on the topic, but once I talked about how broken LUFS is he didn't reply.
LUFS has 2 problems:
- it penalizes strong bass. that's why pop music now has so dense bass that is very "noisy"....because you can't have any peaks there anymore.
- a very dense midrange.....and a loose midrange can have the same reading, but the first will sound louder.

 

[NTK]

I never realized that the limiter was introduced in the digital world.

They urgently need to fix LUFS. I had Bob Katz once replying directly to me on the topic, but once I talked about how broken LUFS is he didn't reply.
LUFS has 2 problems:
- it penalizes strong bass. that's why pop music now has so dense bass that is very "noisy"....because you can't have any peaks there anymore.
- a very dense midrange.....and a loose midrange can have the same reading, but the first will sound louder.

It is not like they haven't done any homework. If you want to "fix" it, you better bring data.
Note that in the ITU recommendation, LUFS is referred to as LKFS.

 

[DVDdoug]

I was irritated by one quote "that it sounds like it sounds on the radio" is a good thing. The theme is competition.

Well, music is a business (as far as professional mixing/mastering engineers are concerned). And in the end I "blame" the "average consumer". If more dynamic recordings were selling better and winning Grammys, everybody would be copying that style. I also suspect that a couple of generations of musicians have grown-up learning and believing that music should be "constantly loud".

With streaming and loudness normalization on the playback side (I use ReplayGain with my digital files), being louder than everybody else on the radio isn't that important anymore. Of course, loudness normalization doesn't fix the "damage" done by compression & limiting. It just an automatic linear volume adjustment makes all tracks about the same overall loudness.

I never realized that the limiter was introduced in the digital world.

At the beginning of the video somebody says digital accelerated the loudness war. Digital just provided more powerful weapons... One of the advantages of digital is that it can use "look-ahead" so it can limit without distorting the wave shape.

Of course, digital has a limit... Everything has a limit. And digital has more dynamic range capability than analog if they choose to use it that way. I was around when CDs were introduced and I wrongly expected that musicians and producers would take advantage of the available dynamic range.

There were analog compressors and limiters (and they are still used). And sometimes they would simply saturate the console mixer. Tubes tend to soft-clip when overdriven so that was a characteristic that could be used in the days of tube mixers. Somewhere I read about line-level signals being fed-into mic inputs! Analog tape makes a pretty good limiter and there are tape emulation plug-ins. Tape tends to soft-clip ("limit" more than hard-clip) and the NAB playback equalization tends to further-soften the distortion harmonics. Like RIAA EQ on records, with tape, the highs are boosted during recording and reduced-back to normal during playback.)

There was (is) a kind of look-ahead that can be used when cutting vinyl records. In the analog days it was a tape delay. It's automatically adjusts the groove width so they can use smaller grooves on the quiet parts and wider grooves on the loud parts, allowing a louder overall recording or more playing time. That does NOT compress or otherwise alter the sound.

The rumor is that Motown Records was winning the loudness war with a process called Loud And Clear

LUFS has 2 problems:
- it penalizes strong bass. that's why pop music now has so dense bass that is very "noisy"....because you can't have any peaks there anymore.
- a very dense midrange.....and a loose midrange can have the same reading, but the first will sound louder.

LUFS is a measurement. It doesn't change the sound.

If you want to hit a certain loudness (LUFS, or by-ear, or whatever) or "win" the loudness war you may have to compromise the sound quality to get there.

 

[dasdoing]

you better bring data

It's common knowledge that almost all songs in popular genres on Spotify are mixed much louder than -14LUFS. Why would you do that if the algorithm will bring it down again? They do it because it will be louder at -14LUFS if mixed at -9LUFS than if it were mixed at -14LUFS. (*)
there is a list out there in the net anywhere. can't find it now. but someone analysed the top 200 songs on SPotify or something similar, and all were mixed much louder.

(*) so, -14LUFS can have different loudness levels. The goal was that -14LUFS always is the same perceived loudness.

 

[Ktacos]

For me loudness has always been a genre thing. Some stuff sounds better when it's smashed, some stuff sounds better when it has a lot of headroom.

 

[IPunchCholla]

It is not like they haven't done any homework. If you want to "fix" it, you better bring data.
Note that in the ITU recommendation, LUFS is referred to as LKFS.
View attachment 503093

Looking at that chart, an I right in understanding that for most of the range there is a 10dB variation in measured loudness to subjective loudness? So a -14 iLUF measured could be -9 to -19? While highly correlated, that also seems pretty loose and gameable.

As someone who has been producing my own music for fun and loves deep bass, I find I spend a lot of time gaming the low end (using harmonics to make the bass sound louder than it is and having soft cymbal hits if done just right, can lower my loudness without touching anything else). I should probably read the algorithm really closely to figure out what strategies to use.

 

[Philbo King]

Bob Katz has some well written articles on his web site too.

Articles - Digido.com

Digital Domain - we ensure musicians, independent artists and record labels get the best sound possible. Bob Katz provides the articles to help.

www.digido.com

 

[NTK]

Looking at that chart, an I right in understanding that for most of the range there is a 10dB variation in measured loudness to subjective loudness? So a -14 iLUF measured could be -9 to -19? While highly correlated, that also seems pretty loose and gameable.

As someone who has been producing my own music for fun and loves deep bass, I find I spend a lot of time gaming the low end (using harmonics to make the bass sound louder than it is and having soft cymbal hits if done just right, can lower my loudness without touching anything else). I should probably read the algorithm really closely to figure out what strategies to use.

The ITU recommendation does say the science is not exact, and is open to refinements of the method (see Note 1).

 

[ThatSoundsGood]

Interesting video.