[78-L] Cambridge Process

[Chris Zwarg]

At 16:01 16.12.2008, you wrote:
>Erwin Kluwer wrote:
>> I never really understand that whole filtering thing.... except for really
>> badly worn records (and no better copies can be located). To my ears it
>> almost always sounds less....

If you talk about constant-noise (hiss) reduction you are quite right - every attempt to suppress this more than a tiny little bit will adversely affect the fidelity. Unfortunately, the problem is worst with very noisy recordings that really would benefit most from less noise (e.g. zinc-etched Berliners, faintly recorded cylinders, early sound-on-film). 

OTOH, short transient noises (clicks, crackles) can be dealt with very effectively, and this "smoothing" of the background will bring out the music much more clearly if done properly. Longer intermittent defects like the swish inherent in many early 78rpm masters or the modulation noise (aka "blasting") associated with worn and/or over-recorded spots - think soprano top notes on acoustic recordings - require much skill and time for manual adjustments but can be much improved as well. If such corrections are done properly you will not even notice *anything* has been done with the source, but simply get the impression you are hearing a record in much better condition than was actually available (often better than the sound of a freshly struck vinyl pressing). This is what you often hear on well-made CDs, that seems to sound pretty much unprocessed and "just like the original", but if you actually compare the two you notice the original is much cracklier.

Also, careful re-equalization, to remedy the "colouring" of the sound caused by spurious resonances and non-linearities in the original recording process, can startlingly improve the vividness and naturalness of the reproduction - details such as room ambience obscured by an inadequate recording process can become audible, vocal and instrumental timbres can get much closer to the original, which in my experience greatly helps in making early recordings enjoyable to a non-specialist audience. I have often found it is not so much the inevitable noise that puts many non-collectors off, but rather a constricted or dull sound that makes e.g. a singer's words hard to understand, or completely alters an instrument's familiar timbre (like e.g. most acoustic piano recordings).

I sometimes find it useful to think in terms of photo processing as an analogy: While it is all but impossible to remove film grain or paper structure (constant noise) from digitized photos without making them look very unnatural, you can effectively retouch spots and scratches (transient noises) and also in most cases restore contrast, colours, and to a degree sharpness to faded or under-exposed originals (equalization), and end up with a result that still doesn't look like a modern digital image, and yet much better than your average decades-old photo, without any obvious defects and irritating colour cast.

Chris Zwarg