On Tue, May 04, 2004 at 10:57:21AM -0700, Mark Knecht wrote: The kind of brick-wall limiting and gain-stage-hammering that people do to get "loud" mixes introduces significant harmonic distortion. That's "where". I can't tell you "what", I don't know the characteristics of the distortion technically. As for "why"...

Wrong choice of instrument ;-) Classical engineers tend to lean toward neutrality and accuracy in their recordings.

I suppose you use only measurement microphones too ;-)

Really it's nothing new. Plenty of "old school" pop/rock/blues/etc engineers pushed their analog tape and analog consoles hard to get some soft clipping, which both introduces harmonics and acts as a sort of peak limiter. Now we're still doing much the same thing, only we're using high-tech brick-wall limiters.

I suspect those old-school pop engineers discovered tape saturation and related techniques somewhat by accident. Nobody thought it out ahead of time, they just noticed it made some records stand out from the crowd, maybe carry a bit better on AM radio, and the next thing you know your kids are watching "duck-and-cover" films in school. Oh wait, that was a different arms race ;-)

On Tue, May 04, 2004 at 12:08:23PM -0700, Mark Knecht wrote: Yeah, I got sloppy there. And here I was just listening to some bluegrass yesterday... All I meant is, it's a pretty unusual candidate for deliberate harmonic distortion (velvet underground notwithstanding).

On Tue, May 04, 2004 at 05:20:57PM -0400, Eric Dantan Rzewnicki wrote: Yes, courtesy of the great John Cale.

I've followed this thread with some amateur interest, as I am not and have never even pretended to be an audio professional. I will say this however... I think there are a couple of levels of discrimination involved with this "louder is better" thing, which is a bit similar to the similiarly greatly annoying trend of television manufacturers' rising colour temperatures. There's the immediate impression... "OOOH! Louder! I can "hear everything" better" or whatever it really is. And then there's the "Wait a moment, where are the dynamics? A piccolo doesn't really sound like it has a 200HP air compressor hose attached to it." Let's not even venture into the shark-filled waters of how you're reducing effective detail by "raising the floor" to such a degree, or the clearly obvious clipping and distortion you're causing. Incidentally, the "additional harmonics" argument is a little weak given the fact that the digital audio chain is not perfect, not by a longshot. For one thing, anything above 22KHz is aliasing noise and should be filtered out by your decoding sections. For another, you have ZERO TOLERANCE for power levels beyond +/- 32767. Zero. That nearly 90 degree shape in the truncated waveforms means you've overlayed ~ 22Khz crap onto your 100-8Khz sound. That's NOISE, and what's worse it's at nearly maximum signal level too. Even as an amateur audio consumer, I find FM radio annoyingly companded. It's like listened to a used car commercial. To get that kind of audio out of a CD would be completely unacceptable. Alas, I am not a fan of rock music so perhaps I'm just out of touch with its sonic priorities. I have a friend who is a recording pro (who invariably sniggers and laughs at my commentary on audio matters) and his remarks on alot of this "digital misery" goes like the following.... "I am a firm believer in this studio philosophy: if you can't hear it, it doesn't exist. Ignore what the machine shows you and LISTEN. That goes for VU meters as well as digital waveform displays." "Recording engineers need to LISTEN again instead of LOOKING." "Values above the threshold are spat out at 32767. Ergo, you have a shitty sounding square wave brought on by a system with zero headroom." "On tape, zero was an absolute reference which could be measured exactly. You then decided how far above zero to allow your recording. If you liked the result, you went with it. And tape being what it is, it is very forgiving and will naturally compress hot signals." "In digital, zero means NOTHING. It means 'don't go past here because this is how we built our system, and if you do you're in square wave land.'" "This is a problem. A generation of engineers is growing up with no basic tracking skills because of this. They keep using their eyes instead of their ears." "The engineers who design this shit are even more guilty of that." [Incidentally, he is a die-hard "Old School" (not host-based) Pro Tools Snob, though you might be tempted think he's some analogue dinosaur.] Back to Malcolm here. There's NOTHING that says digital audio devices can't have headroom, but it means calibrating your levels to avoid peaking above -3dB or whatever the headroom you want to reserve. There's the human "greed" instinct of course which resents this sacrificing of come resolution. [in a digital sample, this is about 1/2 the values!] It's this very greed which allowed The Dark Lord to ensnare the Kings of Men with The Nine to start with. You can be sure the Witch King of Angmar liked maximum squeeze on his mastering projects. Look what it got him in the end. I think the best compromise of this is to provide dbx companders in a button on consumer CD players, so that CD audio can be mastered properly with full dynamics and unclipped sounds. This way, when Joe Jughead wants his compressed and clipped noise, it's just a button away. Incidentally, I wish I can say I claimed credit for this foul idea, but some college students beat me to that. Even The Dark Lord's mastering engineers didn't believe in clipped waveforms, they preferred Quality,

=MB=

Inexperienced listeners often report that they like one mix over another when in fact the two mixes are identical and the preferred is 1 or 2db louder, at least over certain listening ranges. There is absolutely no significant harmonic anything going on when that happens. There are tricks the ear is playing on them when they listen at low levels vs. louder levels.