(This is a follow-up to a widely read piece we published last month about The Loudness War by guest writer Alex (the co-founder and chief editor of Metal-Fi and an audiophile who has been listening to metal for more than 20 years). This article is also being cross-posted at the Angry Metal Guy blog as well as at Metal-Fi.)
“But I like my metal loud. It just sounds better to me.”
This is bar none the number one reaction I get from fellow headbangers, who after they read one of our articles, go off in a frenzy and measure all their record’s dynamic range only to discover they almost always prefer the hyper-compressed albums over the dynamic ones.
Not surprising. In fact, most of the time their results only reinforce why the Loudness War exists in the first place. Let me explain.
From a strictly Darwinian standpoint, the ability to hear is critical to your survival. It allows you to perceive the natural sounds in your environment, find food (as well as avoid becoming it), keep your balance, and most importantly in this day and age, communicate with your fellow headbanger more effectively. Consequently, your ears have developed into a highly sensitive instrument. You have the ability to hear a whisper in a sound-proofed room as well as detect the loudest scream at a Pig Destroyer show. Putting that into numbers, the human ear has about 140dB of dynamic range (the CD is only 96dB) and can hear up to three orders of magnitude in frequency (20Hz – 20kHz).
However, the perceived response of human hearing is not linear with respect to frequency. In English, the apparent loudness of a sound depends on its frequency and intensity. As I stated in my previous article, the study of how our ears perceive sound with respect to volume was first researched by Fletcher and Munson back in 1933. They came up with the first Equal-loudness contour, a plot that shows how your hearing changes depending on a sound’s intensity (sound pressure level or SPL). The curves are measured in Phons, which is the value of SPL that has constant apparent loudness for average human hearing.
Looking at the graph above, imagine you are listening to Bolt Thrower at 80dB. Your head is moving back and forth as the British-made tank steamrolls over your eardrums. Life is good. Then you turn the volume knob down and play the same song at 10dB. Take a look at the frequency disparity between the two. At 10dB, anything below 500Hz is pretty much a wash (sorry Jo). Also note how bass (less than 250Hz) and treble (greater than 2kHz) dramatically change in general depending on SPL. That’s why at 80dB, Bolt Thrower sounds richer and fuller than it does at 10dB. Your ears literally hear more of the frequency spectrum. This graph also explains why, historically speaking, most engineers tend to stay around 85dB when they perform their magic, since that’s where the Equal-loudness contour tends to be the flattest.
Armed with this knowledge in mind, you understand now why louder generally sounds better. That’s why it is absolutely imperative that when you decide to sit down and compare two pieces of music (say your original Septicflesh records versus the recently released remasters) you level match them, i.e., you compare both pieces of music at the same volume. Otherwise your ears will be fooled into thinking one sounds better than the other when in reality it’s actually the reverse.
The two most widespread technologies that allow you to level match your music are ReplayGain and Sound Check. ReplayGain is the most prevalent of the two since it is supported by a myriad of players on Windows, Linux, and Mac. Sound Check is the name Apple uses for the equivalent technology built into their venerable iTunes client. Both basically do the same thing and work almost identically.
Both technologies analyze a track or entire album to calculate the material’s peak levels and perceived volume. Then based on a configurable target value (how loud or soft you want the music to sound) another calculation is performed to determine how much gain or attenuation should be applied to the music to hit that target value during playback.
Please note, no samples were harmed in the making of either the ReplayGain or Sound Check values. Neither of these technologies degrade or modify the source material in any way. The calculated gain value is typically stored as metadata (usually an ID3 tag like the ones you use for album, artist, track number, etc.), so a supported client can read back the number and apply normalization upon playback. Think of ReplayGain and Sound Check as an invisible hand around your volume knob, rotating it up and down to keep your perceived volume consistent across different source material. If you are interested in finding out how to use them, please check out my detailed article here or Google around — they are both fairly straightforward to use.
At this point you may be asking yourself, why am I so against the Loudness War? Doesn’t dynamic range compression and brickwall limiting make the music louder, which means it will sound better? Right?
Wrong. There is a major difference between manipulating the recorded volume versus the playback one. When a mastering engineer artificially pushes the volume higher by applying DRC, he or she is changing the recorded volume by squashing the high and low ends of the frequency spectrum. This process has the nasty byproduct of causing transients and imaging to substantially degrade, making the music sound lifeless and dull. Take a look at the same track level-matched with itself taken from two different masters off the same album (by the same mastering engineer no less):
See all those peaks on the bottom vinyl mix? Those are kick drums, cymbal crashes, transients, etc. They have been completely chopped off on the top digital mix in order to maintain volume homogeneity. There is just no longer any room left for them to exist. Simply put, hyper-compression sacrifices fidelity at the expense of volume and irrevocably damages the original recording. Read Dave’s “Take the Swanö Challenge” article for the complete lowdown (even better, take the challenge yourself).
However, turning the volume knob up or down, as in our Bolt Thrower example above, doesn’t affect the music’s dynamic range or fidelity at all. The original recording is preserved since you are only changing the playback volume. As a result, you can feel confident that no matter at what volume you ultimately decide to enjoy the music, you’re not damaging it in the process.
Do you think the folks I mentioned at the beginning of this article level-matched their hyper-compressed metal records with their highly dynamic counterparts? Almost always the answer is no. They are not alone though. The overwhelming majority of reviews I read online rarely ever talk about level matching when discussing a remaster or remix. So how can you trust a review that is touting how amazing a remaster sounds if the author did not level match it with the original? Simple, you can’t.
Come to NCS, learn about audio engineering. Always interesting this site.
“doesn’t effect the music’s dynamic range or fidelity at all”
Good article, though! Interesting stuff.
Good catch — now corrected.
i made a small change to my recording/mixing/conversion process after loading my songs in the TT DR Offline Meter. i’ve stopped raising the volume in a wav editor and instead just leave the wavs raw as they are. unfortunately this didn’t improve my DR number (6), but it did change the way the songs actually sound; the soft/hard dynamics are much better and there’s an overall improvement in the clarity of the tracks. i’m going to keep tinkering with it on future tracks and see if i can improve my “score”.
Don’t focus on the score, focus on the sound! The score will follow. But be sure you level match everything when you compare so you aren’t fooled by the volume!
I scored an “8” on my own music, what was interesting is that the MP3 and CD versions were actually scoring less, 7 and on occasion a 6. I am still happy with how what I did sounds, but it was an enlightening experience. I know, for the most part, how to get the sound I want with what I have to reasonable degrees, tools like TTDRM help me kind of match how it will sound across all mediums a lot better, with some measure of predictability. After making some small adjustments to my newest mix and pushing certain areas more or less as needed and using less and less limiting, I actually got a 12.
I noticed immediately that on my mixing monitors, normal headphones, and car audio the audio is not just better but far more consistent. Thanks very much for posting this article, it has given me some new tools and ideas to work towards.
No thank you! 🙂
I can’t claim to be an audio expert like you, but I do know that any mix that looks like a sausage is gross and undesirable.
As always, a good read from the folks at Metal-Fi. Just so happens I’m in the middle of a remaster project and this has served as an additional reminder to *truly* compare the results vs the original with links to tools to boot. Love it!