mixing/mastering and high end audio

The best way to understand how accurate something is is to find an acoustic performance you know well and find this recording on a small label that does work faithful to the performance. That way you know what is true.

Once you know that your system is accurate, you will hear what is on the recording. Some bad recordings will sound bad but good ones will be sublime.

matadorpoeta,

You've pretty much listed the basic problems. There is really no standard for accurate playback today, and recordings do sound markedly different on different systems. And good systems can sound very different from each other as well, so even if you decide to optimize for only audiophile setups, there's still no guarantee of accurately transmitting the artist's technical intent (eg. bass line too loud).

That's why audio quality has to play second fiddle to performance and artistic quality. Artistry will always come through even the crappiest channels: listen to some old Hendrix recordings or the old Caruso recordings. Both sound horrible, but it doesn't matter because the artistry overrides everything else.

I think that's a huge problem today: performance values are non-existent, and sound quality is total crap with the loudness wars where everyone tries to make their recording sound louder than everyone else's. It's no wonder that people pirate music like it's worth nothing, because most of it is worthless.

This is not to excuse poor technical quality: obviously you try to do the best job possible so the technique doesn't distract from the artistry, but if you make a pristine recording of a bad performance, it's still a bad performance, and that's what everyone notices.

--Andre

Hi, I can try to be of help, since I've been involved on projects recorded in top studios. The recordings I've participated in have been featured on network TV, movies, and my LP project was actually used to demo super high end equipment at CES. I am player, not a recording engineer, but I been lucky enough to watch many skilled pros in action. As I understand it, mixing and mastering are two different things. Mixing is the process of combining all the sounds that have been recorded in the studio into "tracks" or individual songs. The mixes are stored as a "final mix". From the late 1940's to about the 1990's, this was usually 1/4" or 1/2" analog tape. More recently, final mixes are usually stored on digital formats, which have included DAT, CD, and most recently the hard drives of computers. During the mix, effects can be added, such as reverb, echo, compression. Sounds can be faded in and out as desired. The basic goal is to try to represent the intention of the artist. Some artists will try to mix their music for an intended format. A classic example would be the Motown recordings of the 60's which were mixed so they would sound best on AM car radio speakers. Some artists will mix with the idea of getting the purest sound(which can mean many things) they can get, and they may base their decisions on the sound they get from the equipment in the studio they are using. Both approaches are valid and have resulted in widely admired recordings. Mastering is the process of putting the tracks in the appropriate sequence, and preparing them for the intended format on which they will be duplicated. A mastering engineer will play the final mixes on whatever format(s) they are on and transfer them to a "master". To master a LP, the engineer will play the final mixes (in order), and "cut" each side of the record in real time (or half speed, a much more esoteric process). While he does this, he may add compression and equalization, so that the recording will sound "good" on vinyl. He will also adjust the volume of the tracks, so they are of uniform volume. He makes sure the grooves are the right size, and don't take up too little, or too much of the record's surface. He will try to make sure the needle won't skip during a loud passage. The record he cuts will be sent to the LP pressing plant and a mold will be made for duplication. To master a CD, the engineer will play the final mixes and combine them on a format that the CD pressing plant can accept. These formats have been evolving, and honestly I have not been able to keep up with all of them, but basically it will be a digital file (the file can even be stored on a recordable CD). While he does this, he may add compression and equalization, so that the recording will sound "good" on CD. The style of compression and EQ will be different for CD than LP, because of the characteristics of the different formats. The mastering engineer will also adjust the volume of the tracks, so they are of uniform volume. The file he writes will be sent to the CD pressing plant and a mold will be made for duplication. This a drastic simplification of the process, but maybe it can at least help explain how the systems work.

That's great you had the experience of listening to your music on all different systems while mixing. Many pros like to do that. Actually, I am a big fan of the 60's Motown sound and have spent much time looking around the museum, which is the studio where most of the hit songs were recorded. When I listen to 60's Motown on a high end system I don't really think it sounds that great. It doesn't matter what the source is--CD, 45 or LP, although I favor the Mono mixes on the old LP's. Because the music was mixed to sound good on a car radio (which it does), it has a boxy, midrangy sound with certain elements such as the vocals, bass and tamborine boosted up unusually high. It's fun, cool music, but it resists being high fidelity. Yes, the current trend is compress, compress, compress. Make that music as consistently loud as possible! Actually, you can trace the heavy compression trend back to the early 1960's when pop music LP's and 45's were cut really hot, usually to distortion. In modern times there is a war to see who can make the loudest CD. http://en.wikipedia.org/wiki/Loudness_war The heavy compression causes the music to pop out of computer speakers, but it will sound dull and lacking dynamics on a high resolution system. Stan Ricker* http://rickermaster.com/ told me a striking story of a rare example of an artist involved in the mastering process. Jeff Lynne showed up at Stan's mastering lab in 1977 with all the members of ELO. Lynne had just come from Munich and hand delivered the master tapes for Out of the Blue. (The original LP pressings of Out of the Blue are half-speed mastered.) Stan told me that was very unusual for an artist to care that much about the sound, and rare that the artist was allowed by the record company to care about the sound. *Along with Bernie Grundman and Bob Ludwig, one of the greatest mastering engineers of all time.

Here's the point I'm making by way of analogy. In science, you can't calibrate gear with an inaccurate measuring device. You need a baseline or benchmark. Without that you are guessing as to what will sound good.

Yes, many modern recordings sound poor and are compressed. But how do you design or assemble a stereo system to handle how some producers are mixing for radio or loud CDs? You really can't. You need an "absolute sound" approach which sets a baseline and allows you to enjoy good recordings.

I would possibly go as far as saying you probably can't make even a cheap system sound great with a poor compressed modern recording. It's akin to polishing a turd.

When an artist's style is as raw and abrasive as someone like Hendrix, I think that 'poor' audio quality adds to the piece. See Guided by Voices.

I disagree. I don't think Hendrix was trying for poor sound quality. He was looking for a new, dynamic sound quality. The goal of music isn't only to sound "good". That would be incredibly shallow. The goal of music is to communicate a message, tell a story.

I know guys who have spent many thousands of dollars on equipment trying to reproduce the Jimi Hendrix sound with their guitars.

Also, regarding Hendrix, there's an audiophile LP reissue of Axis done by Classic Records. A friend and I were listening to it last week, and were blown away by the sound quality. On the record I started to notice how much Mitch Mitchell used brushes instead of sticks. On mp3 it would be hard to tell what was going on.

Who knows what Hendrix was going after? That's not really important to me anyways. I only meant that my listening enjoyment of the raw, gritty quality of some of his songs is enhanced by the recording style of yesteryear. I certainly don't want to listen Hendrix as recorded in the style of Nickelback.

I'm not so sure. Hendrix had a raw sound and good fidelity on the Classic Records LP reissues has let my ears hear more of what that raw quality was.

Edit: Looks like tlmusic and I are on the same page.

I was thinking about this and came to the conclusion that one of the reasons I just frankly don't give a shit about having perfect or high-end audio equipment is because no matter what system you have, you can guarantee that the sound wasn't designed specifically for it. In an art gallery, the artist setting up the exhibition decides where to place everything, how things will hang, what color the walls will be, where the text will be and what font it is, how far the viewers will be from the piece, etc. In other words, they have complete control over the surroundings by which the art is viewed. The converse is true for music, which because it is largely based upon consumerism, the consumer is the one who places the speakers and sets up the equipment. Therefore no one experience is the "correct" experience, since no one experience is established by the artist (and no, sound engineers are not the artists, they're skilled tradesmen). Ultimately, you can only judge by how you yourself wish to listen to the music. I've listened to music in a $200,000 media room, and guess what? It still sounds like music coming from speakers, albeit with good resonance and pitch representation. Granted, using stone speaker boxes and having everything placed and sized perfectly might be the kind of sound that gets you off - but it's by no means accurate, because the music still wasn't designed for it. My dad has a $10k audio set-up and there are some CDs I just like more when listening to them in my car on my $500 speakers because I like the tweeters accenting the picking of the guitars. There are plenty of albums I like better on my dad's system because I can better hear the different ranges present. If you want your music to sound like you're listening to it in a smokey bar, you're going to have to listen to it in a smokey bar. If you have a super high-end audio set-up, it's going to sound just like a super high-end audio set up. What's the point, anyway? "Like a Rolling Stone," was only played perfectly once and no matter how many thousands of dollars you spend, it's never going to be accurate to the original stage.

This is not, imho, analogous to music playback. Unless you just listen to a lot of poorly recorded pop music, the better analogy is that of a photographic lens. A high end system is like a Leica lens. It let's you see more of what is there, good or bad. It lets you hear more of what was done in the studio or at the live performance.

I listen to a lot of acoustic, jazz and classical where there are many great recordings, even on CD. The Verve Masters series is one example. The Living Stereo, Rolling Stones, and Bob Dylan SACD series are another. A good SACD or CD player can present a wealth of musical information.

A mediocre stereo will typically have too much bass and very veiled midrange. A bit more money spent at whatever budget you have can lead to a good return on your investment in terms of sonics and musical enjoyment.

When the sonics are good I find I focus on the performance and get drawn into the music more.

You ever listened to the old rachmaninov recordings? Talk about line noise.... aieeeeee. It's interesting to listen to him play and compare him to others playing his works, but it's not something I can do very often. The poor recording quality is actually painful to my ears at times. Of course, they were recorded mostly in the 20s and 30s, so it's understandable.

Yeah, it's bad, and a lot of the nuance gets lost in the noise, but it's kind of a unique artifact of one of greatest pianists ever. Have you heard the Telarc recording of a player-piano that played back some of Rach's recorded rolls? It's interesting, but I'm not sure it's an accurate reflection of his playing.

http://www.arkivmusic.com/classical/...p?album_id=650


I think this is true for some aspects of playback, but not for others. The biggest thing you lose with recording and playback (especially 2-channel) are the spatial cues: the things that give you a sense of the place, where people are, etc. This is why mic placement is such a black art: each recording engineer has their own system to get a certain rendering of the space they're in, and no two are alike in that respect.

However, you can still deliver what each mic heard with as little distortion and change as possible: the mic preamp, recording system, delivery medium, and playback equipment should have very low distortion and flat frequency response.

This gets tricky again when you get to the speakers. The speakers not only cannot reproduce the original acoustic space, but their own tonal response is altered by their design and the room they're in. You can make speakers that are relatively low distortion, but you're going to have a hard time getting them accurate for frequency response and spatial reproduction.

There are many design philosophies, some backed up by impressive data (eg. the Harman group which produces JBL, Infinity, and Revel speakers), and some half-baked about trying to get good frequency response for speakers, but no one seems to know what to do about the spatial aspects yet.

A technical way to look at the system is that the hard parts involve multidimensional vector fields: miking the original soundfield and speakers playing back in a room. Both things are trying to effect a 4D vector field from basically 1D scalar information.

1D information is all the electrical stuff: the signals from the mic through the recording medium to your playback system up to the amp's outputs. In 2-channel you get two of these 1D signals: left and right mappings of amplitude vs. time.

The 1D stuff is straightforward to optimize. The 4D stuff isn't because it is essentially throwing information away to compress it into a 1D signal. An interesting way to think about it is like a spatial version of perceptual coding (eg. MP3, AAC, etc.). The recording engineer decides what's important to record in a particular soundfield, and tries to deliver that with the limited means at his disposal and throws everything else away.

Your speakers then have the job of essentially trying to guess what the recording engineer wanted to do, and expanding its 1D input signal into a 4D output field. Sometimes speakers get help from algorithms like Meridian's Trifield, Lexicon's Logic7, or Dolby's DPLII.

--Andre