Why do mass marketed CD's sound so crappy?

Hammergjh: Yeah, they heard the difference; they could also easily hear the difference between the 24/96 master tape we got to hear in the control room and the 16/44.1 CD they mixed down from that tape. But they're really not the record producers in this case, they produced a demo CD for this artist in order to sell the artist to a major recording label, so they did what they had to do to get the attention of the major labels. No shame there. And I guess the major labels want what sells, what will sound good on a car radio, not what will sound good on a high end system because we're unfortunately not the market they're worried about. I'd love it if the companies would release two versions of a disc, a compressed one for the radio stations to play and an uncompressed one we could buy, but there I go dreaming again!

There was an article in Bass Player magazine about this. Bela Fleck and a famous bass player (I forget his name) did a recording. The article talked about how they struggled with the final mix. The bassist came from a classical background, and wanted no compression, and the microphone far away from his acoustic bass. Bela, being from a more pop/jazz background, wanted lots of compression so it would sound good on car radios.
Why don't we have DECOMPRESSORS anymore? When I was just starting out, I remember seeing in the HighEnd rooms DBX gear that was designed to restore the dynamics original to the music. Now that we're in the digital age, it should be easier to undo than ever.

The use of compression is an extremely MAJOR problem. This not only alters the dynamics of the recording, but can also introduce major amounts of smearing with a lack of definition.

Another problem is the fact that some studios are using speakers with very poor frequency response linearity as their point of reference. When mixing down, the engineers have to factor in what they hear in terms of the sonic presentation that they are trying to achieve. The non-linear tonal balance of the speakers definitely alters their perception of the original recording, affecting the end results and presentation of what we hear in our systems. Recordings that were made and mixed down using reference speakers that lack bass end up getting too much bass to compensate for what they heard in the studio. Recordings that were made and mixed down using reference speakers that are bloated end up getting thinned out and lack impact and warmth. Same goes for speakers that are excessively bright or dull, etc...

As far as recordings sounding good on a car stereo, that is the method that Ric Ocasek of the Cars used when mixing. He would listen to the recording through various "beat to death" car speakers and adjust the sound accordingly to what he heard there. Anyone that has heard Ric's work knows that it is actually better than most of what is being reproduced today ( or even back then ), so we can't blame the bad sound on even the low quality of most car speakers. Sean
>

Sean's final point brings to mind the story, I believe true, of a record producer from the 60s who reputedly did the final mix of the records he produced over the telephone, because he was listening through the same speaker you would have in the then prevalent transistor radio. He figured it had to sound good over that medium to sell, and he was right.

Sean...Your rant about tonal balance of recordings is exactly the reason why you need at least tone controls and perhaps an equalizer in your playback system.

Compression is not always as bad as you say. The most benign form of compression is "gain riding" (manual gain control). This causes no distortion or "smearing". The objections you raise would apply to "fast attack" compression and/or peak limiting.

I once had a dbx expander/compressor, the main use of which was to further compress recordings when I put them on tape for use in a car. Using this device I did make an interesting discovery: compression, and at the end of a recording, the fadeout, is evidently based on the common mode (A+B) signal, and this kills natural ambience which is generally differential (A-B) in the recorded signal. By compressing based on the A-B signal quite astonishing results were obtained with a matrix multichannel system. This seemed to be a bit different from the gain control logic used by the better decoders.