Bob Katz talks about loudness: sounds like crap, only worse
Mastering engineer Bob Katz has an interesting rant about unlistenably loud CDs. He asserts that contemporary CDs are ten decibels louder than those made 15 years ago.
Remember that decibels are measured on a logarithmic scale. A ten-decibel increase means that contemporary CDs are nearly four times as loud as older discs. (See this handy article on ProRec.com to learn more about why you’re going deaf). But the volume comes at a cost — There are a limited number of bits in each sample. Once all those zeros are flipped to ones, there’s nowhere else to put the data, and things start sounding bad. Really bad.
The war for loudness has only casualties and losers. Some CDs made in the year 2001 are 10 dB hotter than those made in 1990! But the system can’t take it…this is only obtained with horrendous amounts of compression and limiting. Can you take one of these CDs for more than 5 minutes? They sound fatiguing, overmodulated…..
Fact: Your CDs are hotter than anything made in 1990. They have reached the maximum level that they can and still maintain the sound quality. CDs cannot escalate because there is a limit. The waveforms of the top of the charts “hits” are shaped like 2 x 4s, sound fatiguing and unrelenting and have no relationship with the sound of a good album. No one is happy—-not the artists, not the producers… And on the radio? Still sound like crap, only worse.
Read the full post at DigitalDomain.
I may do a little post about binary mathematics and explain why it sounds like ass when CDs clip. I’ve become obsessed with this recently because I’ve been ripping a lot of my old vinyl LPs onto my computer. I watch the wave forms as they scroll by on the computer, and marvel at all the musical detail we’ve lost.
RSS blog posts
Rolling Stone’s site of all places had a decent discussion on this subject. http://www.rollingstone.com/news/story/17777619/the_death_of_high_fidelity
Comment by Steve — 12/29/2007 @ 2:20 pm
Loudness in the digital domain is a big kettle of fish. Some of the researchers at TC Electronics, otherwise known for nice but expensive rack and floor effects, have done some investigations into this question of digital mastering into the red. The best read is Thomas Lund’s cleverly titled, “Distortion to the people”, which covers it from mostly a practical perspective.
The remarks by Daniel Levitin in that Rolling Stone article (and in his book) are, at the least, contentious (or in less polite terms, horribly wrong). There is a big difference in the effects of acoustic loudness - the sound pressure level - and perceptual loudness - the perception of that level, that he seems to (often) confuse. There have been a few studies in defense of loud sound, but that is acoustically loud sound, that leads to bone as well as air conduction (e.g., dance music that you literally feel). The compression of the inner ear allows us to have an incredible dynamic range, not to be attentive to loud sounds. The middle-ear stapedius reflex, which protects the inner ear from the duration of loud sounds, is not compressive like the inner-ear cochlea, but a dampener (hence the ‘my head is in a fishbowl’ effect of remaining sound be transmitted mostly through bone). But this comment begins to widely diverge from the topic of digital loudness.
There seems to be multiple issues to the state of fidelity that are not converging at all:
(1) Red Book Audio (16-bit, 44.1-kHz sampling rate audio) has some clear issues because it is based wholly on the math of sampling error, that is, analog-to-digital conversion, and not digital-to-analog conversion. There are so many moments for fidelity loss in the chain from A/D to D/A (I’ll take input and output limiters for two, Alex), that increasing the standard seems an easy win. I’m not sure what the rationale behind 16-bit depth was, and still is.
(2) The new (several years old, but I just can’t keep up with the kids) standard is 128-kbps MPEG-III (mp3). Based on some very simple psychoacoustics (e.g., upward spread of masking, forward masking, most adults don’t hear much above 16 kHz), this form of data compression, and its subtle improvements (e.g., AAC, OOG) is all based on the faulty logic of processing data based on a model of a given system to then be processed by that system.
When you couple this “standard” format with the current practice of digital loudness in mastering recordings, you truly create a music that is only a few degrees separated from the square wave.
Comment by William Whitmer — 2/14/2008 @ 3:01 pm