There is peak, RMS, and LUFS normalization. In that same order there is a difference between volume, gain, and perceived loudness. Peak normalizing all songs to the same peak is not going to alter the dynamic of any of the songs or match the perceived loudness of them at that peak. RMS and LUFS (loudness) normalization screws with the weight and the crest factor of the songs in addition to the just the peak. It will either compress or decompress songs to match their target, and that is totally destructive. I've been fully aware of what loudness normalization is. There's a lot of information available to anyone who might assume I just talk off the cuff.

Here's one that shows what loudness normalization does. It kills dynamics.
https://www.pro-tools-expert.com/lo...4/loudness-normalization-part-2-the-standards

 

My take on this is that I really like amazing masters. And any good content, even if I manually need to pump the volume up - which makes it sound lesser than amazing, but the action makes up for it. And too loud is too loud anyway. So I just don't mess with it from a Spotify listeners point of hearing.

I'm no fighter, but war never changes ;)

In other words, I'd rather hear it how the peeps intended it to sound like, instead of changing anything inbetween other than output volume. In the realm of Spotify alone that is (which has pretty good quality imo). But that's just me as well and I respect your global approach. I am more specific on tracks rather than the ideology I guess.

 

Last edited: Oct 31, 2018

It does not kill dynamics. For some reason you have got it backwards. Loudness normalization makes sure loudness stays consistant, but allows dynamic content to remain dynamic. More dynamic than the compressed music.

Loudness normalization does not apply compression or expansion on the material. Loudness normalization only changes the gain of the music, so that loudness is consistant.

 

Last edited: Oct 31, 2018

But isn't anything below 100% not bit-perfect* anymore? Like losing actual sound data in the name of consistency?

For example: a really fat and well selected/personalised Spotify playlist sounds stronger with normalization off. And turning it on messes up the vibe imo.

*or close to that idea at least

 

No. What loss would that be?

Not really sure what you are referring to. Are you talking about "louder sounds better" and the equal loudness-curve (Robinson-Dadson). In that case you are just fooling your ear, so to speak.

You can still turn up the volume (with loudness normalization enabled) and it will do the same. BUT the dynamic songs will then be more dynamic and "punchy" than the squashed/over-compressed songs. It's all relative.

 

Last edited: Oct 31, 2018

All you had to do was click on the link and read. I put that link because I already got the notion you don't want to accept any information directly from me that's contrary to what you think for whatever reason. It even shows you visual images. Like I said, there is a difference between "peak" normalization and "loudness" normalization. Loudness normalization handles not just the peak, but will also compress according to the target you set.
https://www.pro-tools-expert.com/lo...4/loudness-normalization-part-2-the-standards

 

Last edited: Nov 1, 2018

Like a loss of the capacity for content?
I hope I am understanding or at least hearing this right - for example: if you turn down the volume inside the Spotify software or the Windows mixer, bits get lost as sound gets reduced? Like ASIO, WASAPI or Kernel Streaming is the way to output sound at 100% for a bit-perfect reproduction. So messing with volume before the DAC sounds like fucking up someones hard work spent in the studio.

Nope. And I totally agree that overcompressing into blocks rather than gluey fishbones is bad. And that lo-fi vinyl rips sound thin as well. But in my opinion, normalizing takes away the shine from really well mastered tracks. In contrast to what you believe:

We may have different tastes, but normalizing won't fix squashed/over-compressed songs anyway, so I would stick with what the author had in mind and rely on keeping the original volumes.

 

I have read it. I read it before you posted it and now when you posted it (if I really had missed anything).
But I didn't miss anything.

I know what peak-, RMS- and loudness normalization is, ofc.

Loudness normalization does not compress any material. It adjusts gain to match a common loudness. Jeez.

 

Naw. It's a miniscule dBV difference.
You have a monitor controller before your monitors? You adjust volume on your smartphone? Same thing.

Sounds like your equal loudness curve is speaking again.
The "shine" is still there.

Again, loudness normalization just adjusts the volume/gain of less dynamic tracks. The loudness is homogenic and consistant to 1) keep you from adjusting the volume all the time and 2) work as a remedy against the loudness war.

The "loud tracks" will still be "loud tracks", while the dynamic tracks will have the same loudness as the "loud tracks" but still have more transients/punch/dynamics.

 

1) So currently both big stars and unknown artists upload to online stores(itunes etc) their pop songs mastered to as low as about -12 LUFS?

2) At the same time, CDs are still being sold (and pop songs on CDs are mastered to rather louder levels like to -6 or -5 LUFS)
So artists who want to sell their song both on itunes and on a CD make two versions of mastering? (1st version around -12LUFS for online, and 2nd version e.g. -5LUFS for CD?)


3) So when people buy some songs from iTunes they got tracks with say -12dB LUFS, and when they rip songs from CDs and copy them to their mp3 players, they have (on the mp3 player) a collection of songs that vary in loudness a lot...? (louder songs ripped from CDs and quiter ones from itunes...)
(Sorry for my ignorance)

 

CD is basically a dead format.

What will end up happening in the future (and it is already happening) is that by default the media player on your phone will scan the files and match the volume of the tracks. Just like the streaming services do.

I would also like to say is that -12LUFS is very squashed. -6 or -5 is pure insanity.

While the loudness war may be over in the traditional sense, the battle still rages on. These recent somewhat shady decisions by large online streaming services will hugely influence the way audio engineers all over the globe will do their job. Yes, it is a step in the right direction, but shit will still be too squashed unless you are listening with your laptop speakers.

For comparision the EBU R128 standard recommends normalization to -23LUFS. (what this means is that when the insane loudness war era mixes are played on european broadcast tv, they will be turned down like ~20db)

 

Last edited: Nov 1, 2018

Technically yes. In practice no. In a real world scenario you would need to listen at volumes that cause permanent hearing damage (in extreme cases maybe even kill you) to actually hear/experience the loss of potential dynamic range caused by these loudness normalization schemes.

 

Dang, I could swear that I hear a change in sound as in lesser meat on good recordings. Also keep in mind, that Spotify has 3 different modes/levels of normalization, so there is some kind of difference by default. Maybe it's my monitoring, but sometimes the sound clicks just right at a given level. Even in like foobar, changing output, decoding or oversampling settings makes a huge defference for my taste, so normalizing is a big no-no. And again, an important argument for me is keeping 100% volume for bit-perfect reproduction. I can hear the difference and I'm still alive But hey, let's agree to disagree anytime.

 

This may sound like a dumb question but... can't they just turn down the volume if you give them a louder file?

 

Peak peak, not talking about the "loudness" itself. Yeah that's a pretty simple explanation, I'm dumb, thanks ^^

 

This is not how the dreaded inter-sample peaks work. They don't actually even exist in the digital domain - there is no data between the samples - they are only created when converting from digital to analog signal.

Like I said in the other thread: Most DAC's can handle them just fine and even if they can't the distortion should be inaudible, and even if it's not the end product will most likely not be delivered at -5dbfs or however far the absolute highest inter sample peaks may reach, so there is nothing you can do.

Well, except to turn down the volume before the signal hits the converter. So yes, they can just turn down the volume if they think inter-sample peaks are a real issue.

 

320 mp3 is always worse, but it can be just a little dull or get really tragic.
Sometimes I also prefer lossless wav over flac due to weird decoding, btw.