CEDAR Single-Ended Noise Reduction System (part 6)
Up until recently CEDAR has been operated purely as a bureau service, and this facility is still available from Cambridge Sound Restoration. A potential user of this service may either send up a DAT cassette which Cambridge Sound Restoration will descratch using either the 1-pass or 2-pass process. If the DAT is accompanied by an engineer, the Dehiss and EQ, which need subjective decisions, can be done also.
The environment in Cambridge Sound Restorations headquarters is far removed from what a studio engineer might expect, appearing more like a cross between a computer dealer and a hifi demonstration room. For the standard unassisted bureau service the operator works with headphone monitoring which, although good for hearing clicks, strikes me as not being entirely adequate for evaluation of audio quality. In the more normal au dio domain it is well known that some things, such as pitch-related phenomena, while clearly audible on speakers are difficult to hear on headphones. Both the 1-pass and 2-pass scratch removal processes can have a damaging effect on the music if taken too far, and correct judgement is essential.
Dehissing and EQ can be done using Cambridge Sound Restorations monitoring system which offers a choice of small Rogers loudspeakers or (original) Quad electrostatics in an as yet acoustically untreated space. I was disturbed that the left and right channels of one of my test pieces were treated with casually set different parameters, even though I was assured that this would only result in a slight imbalance in the distortion in the two channels.
But despite these points, the proof of the pudding is in the eating, and I was able to take home the CEDAR demo DAT and another DAT of some tracks which were processed during my visit. One of the most remarkable examples on the demo is a jazz 78 which had been broken and glued together - glued together amazingly well but still with some horrendous scratches. On this example, processing has wholly beneficial effects, as it does on other examples such as one which had scratches which, so I am told, occurred at a rate of around 2500 per second. There is no doubt that CEDAR can produce very listenable results from previously unlistenable discs.
When the scratches on a disc are less major, some possible doubts about CEDARisation might set in, such as that there might seem to be less top end, and there might appear to be some slight compression. This is very difficult to judge, one has to be very aware that the ear is so easily fooled into thinking that artifacts such as scratches and crackles are part of the music that taking them away inevitably results in a sense of loss. On balance, I think that for declicking 78 rpm records the system works well.
When it comes to Dehissing I was less happy. There are two very illuminating examples on the demo DAT. One is a fairly noisy recording of Schumanns Piano Concerto, probably 1950s, which after Dehissing is as silent as a modern digital recording, apart from a little modulation noise. Perhaps its a matter of aural education but there is such a contradiction between the restricted frequency range of the recording combined with high distortion levels and the almost digital silence, that it somehow just doesnt sound right. Maybe it has been overprocessed and with a little noise left in it would have been OK. The other example is a stereo recording of Prokofievs Classical Symphony which in its original form is very present and vivid, but loses sparkle along with the noise in the processing. An example which actually is very successful is a live recording that was made unusable by dimmer noise. CEDAR was able to eliminate this problem c ompletely.
The last track on the demo DAT is a recording of Noel Coward which has been Descratched, Dehissed and EQd. This gives a good opportunity to hear the artifacts of overprocessing. On the CEDAR version there is an effect very similar to the phasing of a poorly aligned NAB cartridge which is definitely not on the original. If you get a chance to hear the demo its very clear just after the words What avails the sceptred race. (No blame is attached to Cambridge Sound Restorations judgement on the degree of processing applied as the customers engineer was present during the restoration).
My tests on tracks I had taken along to Cambridge Sound Restoration confirmed that the demonstration tracks were typical examples of CEDAR at work and not tracks which had happened by chance to work particularly well. My personal conclusions are that Descratching is excellent and has a lot of potential. Dehissing definitely works, but it is not without side effects and needs to be carried out with extreme care.