background image
DigiCULT
.
Info
28
for instance, we have used Sonic Solutions
NoNoise with good results for several years.
T
here is always a subjective element
inherent in restoration. One has to
wage a constant battle between removing
as much noise as possible and changing the
original signal as little as possible. For this
reason, we treat the restored sound file as
an alternative to the original sound rather
than as its replacement. In practical terms
this means that, when we decide to digitise
and restore an old recording, we store both
the untreated sound with the noise inher-
ent in it, and the restored files. Our cus-
tomers can hear both and decide which
one they will use.
W
e also know that in ten years' time
restoration methods will probably
improve considerably.We know from expe-
rience that it is generally useless to attempt
further restoration work on a recording
that has already been treated. In ten years'
time we might also have better equipment
to play our old discs and tapes, so we
intend to keep them as long as possible
after they have been digitised.
THE FUTURE
W
e have learned a lot from our
experiences digitising this archive
and now need to use this experience when
we look to future digitisation efforts. It is
of utmost importance that the YLE Digital
Sound Archive continues to provide users
with a means of quickly finding and
accessing high-quality material if they are
to fully exploit this rich and varied resource.
Listen to two examples of the digital
sound archive: a German language inter-
view with the composer Sibelius and a
rendition of The Magic Flute:
Sibelius
The Magic Flute
Sounds YLE Digital Sound Archive,
2003, http://www.yle.fi
the process is fairly automatic.There is no
reason to alter or `remix' the recordings,
and if a future producer wishes to change,
say, a concert recording made in the1980s,
we see this as post-production which may
also be archived one day but will not
replace the original.
H
owever, when it comes to vinyl or
shellac records or magnetic tapes
from the 1950s or 1960s, there are fre-
quently problems which limit their
usefulness in current production: clicks,
dropouts, hiss, mains hum and so on.
There are many efficient methods of han-
dling such problems in the digital domain;
F
rom a technical viewpoint, the materi-
al most in need of digitisation consists
of open-reel analogue tapes of various
sizes. So far our efforts have concentrated
on finding a method for digitising such
recordings as effectively as possible. In
2001, we started using a Quadriga work-
station for this purpose. As the archival sys-
tem was at that time not yet ready, tapes
were temporarily copied on AIT (Advanced
Intelligent Tape) tapes, and are now being
transferred to their final destination.
A
fter analogue tapes, we will have to
consider DATs and CDRs. Although
they are considerably younger, our digital
carriers already show all kinds of problems:
tapes that will not start, CDs that have
BLER rates (BLock Error Rates) far
beyond acceptable standards and so on. For
CDRs, we can use the same digitisation
process that we are using for commercial
CDs. For DATs, we have used an outside
supplier to transfer 700 DAT tapes to
sound files on AIT tapes, which we have
then moved to the archival system our-
selves.These tapes were already catalogued
in our database, and each tape had a bar
code with the tape's UID.Technically, the
experiment was successful, but we still have
to evaluate the economic aspect and deter-
mine which procedure is more effective
outsourcing or doing it ourselves.
RESTORATION
A
s a general principle, a sound archive
aims to preserve the recordings in its
collection in a condition which is as close
as possible to the original. However, our
main focus in preservation is on the signal
rather than the carrier. All magnetic tapes
will crumble with the passing of time, but
we can hope to preserve the sound that
has been recorded on the tapes.
W
hen professional-quality tapes from
the past three decades are digitised,
LINK
LINK
Images of The Hague
Salzb
urg Researc
h