Was: arsclist reel to reel player/recorder

["George Brock-Nannestad" <pattac@xxxxxxxx>]

As is common in web discussions, this one has branched out. 
Both lines are extremely relevant.

The present theme is "Digital Storage vs. Analogue Storage", and 
responses (if any) should change the subject line.

I believe that the idea of transferring sonic content to another 
medium should be seen with a perspective. The Technical 
Committee of IASA is currently revising the technical 
recommendation TC-03 entitled "The Safeguarding of the Audio 
Heritage: Ethics, Principles and Preservation Strategy". In my 
contribution to the work I have stressed that we both have the 
intended content, the primary information, and ancillary content, 
the secondary information. The primary information is what we 
colloquially call "the recording", and obviously that is what we must 
strive to reproduce as correctly as possible to obtain a signal which 
may then be stored by digital means, with all the error correction 
and cloning of signal that is possible in that domain. However, we 
frequently cannot determine the primary information precisely 
without having access to the secondary information and being 
aware of it. The secondary information may be background sounds, 
hum, certain distortions, rumble as well as separate materials 
surrounding the primary information, such as tape boxes with 
handwritten information, writing on lead-in tapes, colour codes, etc.

 The primary information is quite naked, and source criticism as to 
its proper use will have a hard time, briefly put, you may compare 
performances, but you may find it difficult to establish the 
circumstances that created the difference between two recordings. 
If transparency in the recording-reproduction chain has been 
obtained, then there is only description ("metadata") to provide a 
check on contents. And since the description is human-generated, 
it may also contain lies. Hence, authenticity of the original source 
must be relied on, rather than determined by source criticism. If, on 
the other hand, there is a lot of additional noises, then a lot of 
ancillary, secondary information is present which has not been as 
easy to tamper with, and responsible analysis may proceed.

Now, dividing the total signal content from a carrier into these two 
completely different categories is in itself an artefact, it is a way to 
look at things. The reason that I use it (it is an application of 
Operational Conservation Theory) is that it forces an awareness of 
what we are handling. In principle we destroy information 
(secondary information) when we make a transfer to a digital 
medium in which domain there is potential for eternal life (still-life). 
There are many things which we cannot find out afterwards. If we 
filter and noise-reduce, then those offending noises will not only 
have gone (for immediate gratification), but long-term we will never 
be able to use the qualities of these noises to increase our 
knowledge about the primary signal. 

Let us have a little cry over the fact that we cannot do the ideal 
thing, and then move on!

In the real world we have three possibilities: to collect everything 
and let wear, tear and environment do the selection for us, to 
preserve as long as possible the medium and the machines for it 
(the system), to transfer to a more durable system. If we are poor 
we must make a selection ourselves. A number of years back, the 
Danish Radio would preserve 7% of the output, selected with 
breadth in mind. However, we cannot be sure if the future will have 
the same concept of breadth as they had when the selection was 
made. These 7% were (and are, I expect) guarded and protected 
by re-copying programmes. 

I will now express a heretic view which obviously cannot go into a 
recommendation from an international organisation: we have to 
indicate what must be regarded as best archiving practices, 
something that activities in planning and carrying out daily archival 
activities can be measured against. If these recommendations are 
followed, we can be sure that the material thus handled will be 
available to future users of all kinds. However - and this is the 
heresy - the future users will in many cases desire a breadth which 
widely surpasses that which has been selected out. I believe that 
they will sacrifice quality for variation and breadth. In other words, 
rather than spending time and effort on doing the best final transfer 
before eternal life (analog generation of a signal which will be A/D-
converted), perhaps we should do mass digitisation without much 
calibration, but with good indexing for later retrieval. The future user 
will in a number of cases find distorted or ill-sounding sounds, but 
they are still better than nothing! And (a worse heresy coming up!): 
a data reduced (physiological data reduction) signal which the ear 
perceives as sound, but which has had all its statistics destroyed 
*is still better than nothing at all*. Not all the future queries to the 
material are of the type "can we determine if a female voice 
belongs to a nervous person by means of the power spectrum of 
her utterance "ouch"".

So, to sum up, the heretical view is "bad sound is better than no 
sound at all".

Now for a visualisation of what 16 bit, 44.1 kHz really means:

"This means that the amplitude can linearly assume + or - 32565 
different sizes (sic!) If we suppose that our eyes are extremely 
sharp, we could draw lines of a hair's breadth, namely 0.01 mm. 
With the same graphic resolution, a signal of 1 sec duration could 
be drawin along a time axis with a length of about 45 cm. Similarly, 
the range of amplitude covers a length of approximately 70 cm. A 
graphic presentation like this requires a sheet of A2 size [GBN 
note: this is similar to 4 sheets of letter-sized paper]. This will 
suffice to show that even the representation of the processed date 
is highly problematic, let alone their survey."

The above is a quote from: Istvan Pintér: "Sound Microscopy and 
Music in the 20th Century. A Survey with Special Reference to 
Hungary", in "I sing the Body Electric", Ed. Hans-Joachim Braun, 
Wolke, Hofheim 2000 (pp. 135-147). This is a book that I have 
reviewed for the IASA Journal.

The above precision shows what Art Shifrin is aiming for when he 
draws waveshapes.

Archiving philosophies are good, not because they get the job 
done, but because they create awareness about our goals and 
responsibilities.

Kind regards,


George Brock-Nannestad
Preservation Tactics