Originally posted by HighlandDougie
View Post
I've been working my way through the Mercury box, and I like some recordings quite a lot, others less so. The label seemed to acquire a "legendary" status, which might now be questioned, or at least re-evaluated. Probably the first I knew about Mercury was the recording of the 1812 overture, which a few of my family bought. I think one challenge in those days was to keep the pickup in the grooves. That was slightly before the "hi-fi" thing took over, so high playing weights and auto-changers were order of the day.
Stereo was just beginning to take over from mono around the late 50s, so that was a novelty. Different companies seemed to adopt different policies, perhaps based on ideology, as to how to record in stereo. With mono they could use one or more microphones, and then "simply" mix them all together to form a sort of blend. I'm sure it was more tricky than that, and there could well have been some phasing and level issues, but since there wasn't any need to maintain spatial cohesion there wasn't perhaps such a need for accurate phase/timing and levels which are needed in stereo or multi-channel recordings.
With stereo there were several approaches, all with some theoretical and experimental justification. These would include:
1. Widely spaced microphones.
This wouldn't work well with only 2, but with 3 or 4 could give passable results. Possibly Mercury and RCA went this way.
2. Dummy head approach.
This was tried, but never really worked for loudspeaker listening. However some companies went partially this way, using microphones not too far apart, and separated by a baffle board. Spaced microphones will give time delays between the channels (also phase shifts) which can be used to locate sounds spatially.
3. Co-incident microphone approach
Blumlein at EMI had described an approach using coincident microphones, and in this method any phase information which could be used for location was essentially lost. However, the theory was that on playback, the spatial location of frontal sounds would be good, as the (perhaps subtle) accurately recorded level differences between the left and right channels would give good spatial localisation.
4. Various modifications of the above
While the general ambience could be captured using one or more of the stereo techniques, sometimes engineers wanted to spotlight a particular instrument, or group of instruments. They could do this using extra microphones. Quite possibly they would have used mono microphones in the early days, and mixed this in with the general sound. In some cases they may even have had to do this in real time, since in the very early days they might not have had multi-channel recorders. Later they would have recorded the spot mics on separate tracks, and done the mixing in the later stages of production. In order to keep spatial locations fairly accurate this mixing process required some adjustment in the mix to the left and right stereo channels. The simplest form was called pan-potting, and simply used slightly different levels of the mono microphone output to feed to the left and right channels in the recording. If the spot instruments were on the left, then the strength of the signal fed to the left channel would be stronger, and vice versa. Some companies may also have put timing/phase delays in as these can also be used to enhance the spatial localisation.
Experiments done on human hearing showed that spatial location could be mostly determined by two parameters - the relative amplitude of the inputs to the left and right ears, and the relative phase of the inputs. These basic parameters have different effects at different frequencies in the audible spectrum.
As recording engineers became more knowledgeable, some paid more attention to phase and timing delays, and used equipment to ensure that different microphones in the recording studio/hall would not produce cancellation or other phase effects.
Decca decided on a fairly standard recording arrangement, at least for orchestral recordings, which they called the Decca Tree.
Besides all this, different microphones sound different, and different companies used different kit.
On top of this all, they were producing recordings for a new mass market of LP buyers, and the engineers were aware of limitations of the recording equipment, the media and the equipment used to reproduce the sounds. Mostly they did not want to have a very wide dynamic range as this would cause distortion in various parts of the system - the disc cutters, the playback devices (domestic record players), and lastly having a very wide dynamic range would cause problems with the amplifiers and loudspeakers, and would also tend to be considered anti-social!
Some of these concerns still apply today, which is perhaps why some companies do produce recordings with a much wider dynamic range than others.
There is a track on the Tchaikovsky 1812 CD (Mercury) which gives an explanation of how the recording of the cannons for the 1812 were made - CD9 in the box. This might at least show some of the problems which recording engineers faced, and no doubt continue to face.
Comment