"IT SOUNDS GOOD" : DOES IT MEAN ANYTHING ?

In other words, is it possible, and if it is, how is it possible to properly define audio quality ?

 

 

It's not so easy to answer : first of all, it depends on the kind of music we speak about. In classical music, engineers have turned out with several "objective criteria" to judge a recording. But above all, one might say a classical music recording sounds "good" when it gives an impression of realism.     Either the so called "illusion of stage" accurately reproduces what originally happened in the room, or the illusion "lies" but do it well : the illusion is realistic, possibly more interesting than the original reality.

 

 

When dealing with other kinds of recording, the issue is quite different. For instance, a more modern kind of recording, as it is commonly done in "instrumental based rock music" :     Classical music criteria might be valid here, to some extent. For instance, it's nice if we still have width and depth, but it's not mandatory : it can be the producer 's will to have the whole recording in mono if he thinks it's the right thing to do. Or he can decide that the recording can be a bit blurred : lack of transparency and definition, which can be an interesting choice sometimes.   In this case, the perspective is changed : the meaning of the criteria axes are changed :   Each "objective quality criterium" can become a musical parameter. In this case, there is no "good" or "bad" anymore.

 

 

The more "non acoustic" the music, the more difficult the issue. Criterias tend to vary a lot, depending on each recording and/or style. Here are a few anyway : - Spectral steadiness In this kind of music, it's often important to get a steady spectrum. Basically, it's the mastering engineer job to get such a result, but it's a good idea to already have a good spectral steadiness after the mix is done. Basically, it means that for instance, the bass level is always more or less the same throughout the song. Same thing for the hi medium level etc. - Clarity/differentiation A good mix is often clear - it means that whatever the number of parts to be mixed, every single part remains clearly understandable. Once again, it's possible to make "meddled" mixes on purpose, but it's quite rare. Sometimes, musicians/producers/mixers can play with differentiation to display a great deal of virtuosity - for instance, the NIN guys are very good at that game, in the Perfect Drug remixes for instance. - Punch/accuracy Very often, even if the music itself is quite soft, the parts need to be processed so they are accurate and punchy. This is especially true with the drum parts, but every other instrument often needs to be processed this way as well. To get to this result, one needs to deal carefully with EQing and compressing. It's a bit like dealing with levels, color balance and contrast on pictures so they get reasonably flashy

 

 

Afterwards, it gets a bit more difficult and, above all, style specific. You've got all the criteria which deals with one instrument at a time : what's a good snare sound, a good bass guitar sound etc. And once a so called "good" instrument sound is defined, it becomes a reference not for being good in itself, but for becoming a possible position on an axis, like we've seen a few lines above. With electroacoustic oriented pop music, a lot of parts are more abstract than instrumental, thus the difficulty of saying "it sounds good" or not. You can always judge a sound in terms of transparency, definition, acoustic surrounding etc. though, but those criteria are not sufficient to properly appreciate the sound in its context.

 

 

MICROSCOPIC ISSUES / ANALOG VS DIGITAL ISSUES   Another important aspect, which can be called the "microscopic" or "texture" aspect, and which is related, to a certain extent, to the "transparency" criterium. An interesting way to introduce this subject is to consider the "analog vs digital" issue. Quite a bit of time ago, there were no digital techniques - everything was recorded on analog systems, with their usual drawbacks : a relative lack of transparency, jitter, background noise etc. With the coming of digital techniques, engineers and musicians saw an opportunity to get rid of all the analog related problems : transparency could at last be easy to get, jitter was a bad memory, and the background noise would be so low that it wouldn't matter anymore. Unexpectedly, other trouble arose : an "all digital" recording is often dry, cold and disincarnated. The more digital one goes - eg mikes with integrated ADCs and AES/EBU outputs, going directly into an all digital system, the more painfully cold the result gets. After all, it seems that the "microscopic" artifacts brought by analog machines are very useful. Nowadays a lot of recordings are made using a mixt setup : basically, a digital DtD system with analog preamps or peripherals.

 

 

This "analog vs digital" issue leads to a wider issue, related to sound design : the "texture color", "microscopic" aspect of sound. It can be noticed that : - analog machines bring a different color than digital machines - different analog machines bring different colors - different digital machines also bring different colors The first point is obvious, and if the second point also quite obvious, the third point is a bit less so.   Back to main index

 

 

See another approach of the matter here