With the resurgence of good Hi-Fi, where more and more people buy good headphones and listen to a lot of music thanks to streaming, it’s sad to see the things the mastering engineers (and in the end the artist) do to music. Yet, the recordings are getting worse. Even old records that gets the “Remastered” or “Expanded” treatment doesn’t sound as good as the original. Why?
As Loud As Possible! The Loudness War
Everyone wants to one up on another by making their record “hotter”, making it overly compressed. It compresses the dynamic range of the music, so it sounds equally loud, which makes it sound perfect for people listening to music on a 1-inch speaker on a transistor radio. This has been going on for a long time and It’s popularly known as the “Loudness War“. This made sense back when people consumed their music through transistor radios in their cars, where the engine sound drowned out the songs. Or maybe their boomboxes at the beach. But look around, most people listen to music with headphones today. A lot of people use the crappy headphones that comes with their phones, but more and more people upgrade to better sounding headphones. But even with crappy headphones, some dynamic range makes music sound so much better.
Mastering Engineer, Guess What? It Doesn’t Work!
All streaming services uses ReplayGain (or equivalent), which calculates the dynamic range, maximum loudness and compensates when played back, making music being more equal in loudness when played back. So if you do it the old, “let’s crank it up to max to make it louder!” doesn’t work anymore. Quite the opposite. Your song will actually play back at lower volume than a more dynamic recording on streaming services like Spotify or Apple Music.
This isn’t just a problem with new recordings done today. Every time you see the dreaded “Remastered” or “Expanded”, you can be almost certain that some asshat has cranked everything to 11, and in the process squeezed the life out of the music. You’ll won’t be able to listen to more than two songs before you get listener fatigue.
So let’s look how bad it is. Let’s take a classic alternative Rock album “Goo” by Sonic Youth. This is from their released CD on Mobile Fidelity. Play close attention on the round ring, where the green line shows variations in the dynamics of the recording. The highest peak is just under the 0dBFS recoding level you can have on a CD, because unlike analog recordings, digital sounds horrible when clipping.
Ok, so let’s take a look on a really high resolution 24-bit, 192KHz “Remastered” version from HDTracks.com. Check out how the mastering engineer (or artist) managed to compress the living hell out of the dynamics of the recording, making it equally loud almost the entire song. Check out that green line. Like a brick wall. Don’t get me started why on earth they released this in 192KHz either. There’s only extra signal to heat up your tweeter if you have a really good speaker system. Or maybe having bats flying into barns…
The thing is, you can hear this easily when you compare, and will never touch the “remastered” version again.
As you can see from the frequency bandwidth, it seems like the mastering engineer actually had access to the analog master tape, because there is information beyond 22KHz. But they managed to totally destroy the album. The album is from 1990, so it pre-dates digital recording. If you listen to the CD version and then switch to the “high-res” remastered version, you can’t stand listening for many seconds before giving up.
This is not an anomaly, most recordings gets this treatment when they do “Remastering”. And if you wonder why vinyl LP’s are having a comeback, here’s one reason. You can’t make vinyl too hot, because the pickup will jump out of the groove.
Want To Check?
I’ve used to use the free program Audacity, but found a German company called XiVero that does an excellent suit of audio programs. MusicScope is an excellent application (and AU-plugin) for checking out why albums sounds better or worse. You can see the amount of compression, dynamic range and frequency response. It’s also great to check for another, far less known problem…
When music producers, mastering engineers or anyone working with audio, they usually work completely in the digital domain. When checking levels, they rely on the digital level meter, that shows how close to clipping they are. The problem is that they usually doesn’t have a clue on how an DAC actually works when converting the digital recorded signal back to the analog domain. You can have a recording which never reaches the “red” level on the meter, but when a couple of samples are at or near 0 dB, the DAC interpolates the signal, making the resulting signal higher and dutifully clips the signal.
If you have a quality DAC, like the Benchmark DAC2 HGC, they designed the DAC to have a +3 dB “headroom”, and a lot of other quality DACs can handle inter-sample peaks, but the cheaper ones found in phones and computers? Not so much. And paying $2000 to fix a problem that shouldn’t be there in the first place? Not the best reason for choosing a DAC. Although, If I had $2000 I would happily buy a Benchmark DAC, but for the audio quality alone.
So how do Inter-Sample Peaks look like? Here you go. This is “Felt” by Garbage. Fitting name…
Check the green (now mostly red) line, showing when inter-sample peaks rears its ugly head.
Why is this a problem, I can’t hear it?
Take a song with a lot of inter-sample peaks, and run it through an MP3 or AAC converter and look at the result. It’s even worse. Here, let me do it for you…
Look Ma! I made it even worse! Maybe I have the chops to be in the music business after all!
By the way, if you take appart Apples script for testing audio to be labelled “Mastered for iTunes”, this is what it checks for. How hard can it be?
Yeah, I know that you can use it as an artistic expression (I like Iggy Pop’s mix of Raw Power better than David Bowie’s, even though everything is cranked to 11), but for god sake, get a grip!
I’ve been working on a high-end headphone amplifier for the last eight month and it sounds fantastic. Many of my friends tease me, saying “why the gear when you only listen to old punk records anyway?.” Funny thing is, the old punk records have more dynamic range than most of the modern pop records.
The EU Volume Cap
“In October 2008, the European Commission warned that listening to personal music players at a high volume over a sustained period could lead to permanent hearing damage. As a result, the European Committee for Electrotechnical Standardisation (CENELEC) amended its safety standard for personal music players.
Now all personal music players sold in the EU after February 2013 are expected to have a default sound limit of 85dB.
The user can choose to override the limit so that the sound level can be increased up to maximum 100dB. If the user overrides the limit, warnings about the risks must be repeated every 20 hours of listening time.”
When I’m on a roll ranting, let’s take up another pet peeve I have. EU has put a volume cap on all portable music players, like phones or DAP’s so they aren’t allowed to output more than 85dB. So an iPhone sold outside of EU has more amplification. The snag? The measurements are made with the crappy headphones that comes with the unit. There headphones are usually made for maximum efficiency, not audio quality. That means that when you connect headphones that needs more power, the volume is too low. So you need to invest in an external amplifier or DAC/amplifier combination to be able to use better headphones.
This is insane! I’m pretty sure that bad headphones do more damage than high volume due to distortion. So If you’re planning to buy an iPhone, do it in the US, where you get more output. There used to be a patch for the iPhone firmware to remove the volume cap, but it doesn’t work anymore. Total insanity.