recording method vs sound reproduction

[esldude]

I have seen a couple different research papers where recordings were made at the same time with several different miking techniques. Then judged by a panel as to which were most accurate upon playback over good systems. In one of those papers they used a very good system indeed. In one case the listening panel was shown positions of what was recording, and in the other they were present to hear it live. IRC, both chose Blumlein crossed figure 8's as most accurate. These both used playback over speakers. Speakers were used at an angle of 60 degrees for all playback. Strictly speaking accuracy during playback with crossed figure 8's would be improved somewhat if the speakers were angled at 90 degrees.

 

I have done a little recording, and find crossed figure 8 mikes or the mid/side version of that to perform most accurately most often. Other versions of paired mikes can also work very well if angle of recording and playback are taken into account. The better you deal with your room and the speakers within it the better it can be, but usually the recorded room masks your room pretty well spatially, but not in terms of frequency response. Sometimes those response errors can manifest as perceived spatial anomalies too.

 

Nevertheless, there are limits to true playback accuracy possible from using only 2 channels over speakers. More channels can create some more accuracy than 2 channels over speakers can. But for the mainly upfront portion of it stereo can be very good, and rarely is commercial recording that well done.

 

Now theoretically binaural could be better if you used a dummy head with the right shape and the correct outer ear shape. Some people find it works while in my experience binaural never quite delivers. And how much it delivers varies greatly from headphone to headphone. It also punctures the naturalness to me when minor movements of your head have the soundfield moving with you. In reality moving your head should occur within a static soundfield. That is something speakers get right better than binaural recordings over headphones. There is equipment which can process the sound so head movement allows the soundfield to remain fixed with headphones. I haven't had the pleasure of using such equipment, but many who have speak highly of it.

 

Now there are several variations of two microphone recording than can work quite well. They produce very natural sound handled properly. Such recordings commercially are extremely rare. Even many of the outfits with a rep for minimal miking and processing use three or four or several microphones for various reasons. Similar to compression and the loudness wars. A little compression sounds better, sounds louder and within reason we always perceive louder as better. Until of course it becomes a ridiculous amount. Blend a couple extra mics for better low end or a bit more room sound being evident or whatever. In those ways it sounds a touch better until it becomes ridiculous.

 

As much as I wish more recordings were done simply, and the good sound comes from the good musicians in a good space that is not happening. The genie will not be put back into the bottle except rarely.

 

Now for an example. Several of the Enya tracks on her early CDs were made with extreme layering on her voice or so I have seen reported. She sings the vocals once, then a second time, and then a third time and so on and so forth. These are all mixed together. With of course slight variations in her voice, her pitch or timing. The extreme layering involved her doing this over 100 times for a few of the songs. So she sung those lyrics 100+ times and they were mixed together. A few too far out of timing were tossed. You get this ghostly, spacely unique sound to her voice. It isn't and could never be natural. In this case, in my opinion, it actually was in good taste and makes for a pleasing and interesting sound. There is no natural version of it however.

[esldude]

I couldn't agree more.

 

It completely baffles me how anyone could possibly think otherwise.

 

I agree with that. I believe Barry D has the idea one mike per channel of playback. I think that is a good way to go.