|
|
Toronto AES BulletinFebruary 97
|
Peter Cook (Editor, CBC Records) introduced tonight's topic. CBC Records is dedicated to Canadian talent and has three labels: SMCD for large ensembles, MVCD for soloists and chamber groups, and PSCD for historical recordings. The 6th Glenn Gould CD is due out soon on the PSCD label.
Glenn Gould was recognized early as a major talent by the CBC and made a number of broadcasts in the early 50's. Most of these were recorded by the CBC as "air checks" on 16" acetate disks. When it came time for the CBC to release these recordings on CD most of the original discs were missing. Fortunately Gould had archival copies made for his own use first on 78's, then 33.3's. These copies are in the Glenn Gould Collection at the National Library of Canada.
Gilles St-Laurent (National Library, Music Division) next talked about the National Library's collection. The library's role is to acquire, preserve and promote the published heritage of Canada. As a result it has the most comprehensive collection of Canadiana in the world (over 14 million items), consisting of books, periodicals, sound recordings and other materials.
Two copies of everything published in Canada (and one copy of all music recordings) are deposited with the National Library's Music Division. Also in the collection are sheetmusic, demos, original records, etc. starting with Emil Berliner's 5" flat records of the 1880s.
Among the well known music industry personalities whose collections are held at the library are Randy Bachman, Oscar Peterson, Jim Vallance, Bryan Adams, and Bruce Allen.
But the busiest collection by far is the Glenn Gould collection. Researchers (and fans) from all over the world visit the library to listen to rare recordings, read Gould's writings, examine his scores and personal effects, and wallow in all things Gould. Gilles showed a 'score' for the prologue from the radio documentary "The Idea Of North" which clearly displayed the theme and variations form of the speaking voices.
Gilles' studio is well equipped to handle the archiving and restoration of all formats of sound recordings. He has probably the only laser turntable in Canada. At a cost of $30k the payback comes from being able to play irreplaceable records without causing any wear. He also has a standard turntable with 25 styli of different sizes and shapes. Gilles played a dramatic demo of various styli on a damaged record which illustrated the importance of matching the stylus to the record being played. This is very important for disks recorded prior to the microgroove standard existed.
For restoration he has the Declicker, Decrackler and Dehisser from CEDAR (Computer Enhanced Digital Audio Restoration). However Gilles made the important point that archive recordings are copied (to CD and analogue reel-to-reel) warts and all via the best possible playback, but not processed, to avoid altering the material in any way. Future technology will allow for improved restoration of original materials.
Other equipment includes the new Pyramix digital audio workstation from Merging Technologies, Apogee ADC and DAC, Studer analogue reel-to-reels, CD-R's, CD analyzer, Quested loudspeakers, etc.
Peter returned to outline the details of his restoration process. Acetate records are made of soft thin plastic on a metal backing; the plastic shrinks over time, and this causes cracking and a particular crackly surface noise. Acetates are only 15-25 minutes per side, and break off wherever they run out of space so that editing must include the characteristics of two different records within a musical phrase, often with no overlap.
Next, Peter explained that personal concert copies were often done off-air by recording studios in the early '50s and the Gould recordings found in the National Library's collection were of uncertain parentage. He began the restoration by a transfer with no EQ (velocity output), a stereo cartridge, the most suitable stylus, and a super-cleaned disk. The output went via a 20-bit A-D to a hard disk, for all-digital processing. He used the Sonic Solutions denoising equipment at the CBC and the CEDAR equipment at the National Library.
The next step was CEDAR declick with a medium to large click setting. This consists of removal of all sound during the click pulse, followed by interpolation of adjacent sound and seamless substitution in the empty spot. This was followed by a small declick pass on the CEDAR declicker.
Next is the decrackler (if required) and the dehisser. One must be wary of processing artifacts in heavy decrackling. Broadband denoising is potentially very hazardous since it can affect the sound timbre, so it needs to be applied judiciously.
The Sonic can do manual declicking to catch the clicks missed by the CEDAR. It can take 4 hours to manually fix the clicks in a minute of music. The Sonic also has powerful multifilter processing capabilities (up to 256 filters may be applied at once with Q factors up to 100). And of course it is a powerful editor. With the stereo cartridge, one can choose the better groove wall (left or right) of mono recordings and with a digital editor edit back and forth between the groove walls as necessary. Peter suggested that for the most serious work, both Sonic and CEDAR systems are needed!
The CEDAR Azimuth Correction is very useful before mixing to mono, and can constantly adjust for slight left-right timing errors. The next step is summing left and right to mono, to cancel turntable rumble, other low frequency noises, and to avoid left-right artifacts or leftover noises from becoming noticeable as stereo.
At some point in the restoration process there is usually groove wall selection of the best one, pitch correction to correct for incorrect turntable record speed (resample), complex filtering such as 60Hz and harmonics (and under 30Hz is cut off), followed by final cleanup and final EQ to make the record sound "right."
Peter played some material flat from the disc, then EQ'd (less highend noise but more hum), followed by one pass of decrackle and declick which left merely some hiss and some swishes in the music. The final version had no hum and only occasional surface noise.
Gould's recordings did not have many edits by today's standards, notwithstanding his legendary fussiness. And these live to air broadcasts are proof (if it needed proving) that he did not rely on editing to correct for technique.
Many questions were answered throughout the presentation. Lorne Tulk (Gould's CBC recording engineer) and Vern Edquist (Gould's piano tuner) were present in the audience and very helpful with the historical questions that no one else could answer.
Thank you Peter and Gilles, for an excellent recording and a very interesting evening.
Reviewed by John Fourdraine.
| Preview of this Meeting | Back to the Mar 97 Bulletin |
| Toronto AES Home | Current Bulletin | Future Meetings | Back Issues | Search |
This page was created Mon, 17 Mar,
1997 at 23:07
Please contact TorontoAES@vex.net to report any problems.