Re: [ARSCLIST] Two other N.Y. Times article on a different type of digitizing

[Karl Miller <karl.miller@xxxxxxxxxxxxx>]

Based upon the information Zenph sent me, what Tom describes below is exactly what they have done. There are, of course, some curious limitations. When it comes to acoustic recordings, some of the bass notes are below the range of reproducable frequencies, so one has to interpolate what the fundamental and its relative amplitude would be by examining the overtones.
   
  One of the examples Zenph sent me, and I may be repeating a previous post on this subject, was a Cortot performance of a Chopin Prelude. A few notes seemed to be oddly louder in the Zenph playback, versus the original disc recording. There is also the same potential for the playback of a reproducing piano roll, however, in the case of the roll, we wouldn't know if there was a discrepancy since we wouldn't have anything for comparison. The reproducing piano roll is not unlike a midi file. Companies like Welte, made every effort to have the action of the piano even at the time of the recording. Similarly, it is essential that the action of the playback piano be adjusted similarly. In the Cortot example I heard from Zenph, it seemed likely to my ears that there was some discrepancy between the alignment of the action of the keyboard of the original instrument used in the recording versus the playback instrument. Such differences could explain those amplitude differences I
 "heard." I use quotation marks since I did not check for any relative amplitude differences, it just sounded that way to me. I should add that I only noticed it on only a few notes. In all other respects I found Zenph's work to be remarkable. 
   
  There is also the potential for a differentiation between the acoustics of the original recording and playback venues, a problem that is also present in roll transfers. The Zenph technique has the advantage of being able to hear the acoustic of the room (by listening to the original recording) and trying to make the appropriate adjustment for playback...assuming that that was not any electric manipulation of the acoustic for, in the case of the post acoustic recording era,  the recording.  In making roll transfers we try (for the releases on my label) to estimate the acoustic of the original recording venue. We know the approximate size of the Welte recording studios (which is only a small part of the equation), but in the case of "location" recordings, we try to estimate the acoustic based upon the way in which the melodic line is sustained. Of course, it is not just a question of the size of the room, but the sound of the actual piano used for playback and the materials
 used in the construction of the room, etc.
   
  The real problem with what Zenph does comes with looking at acoustic recordings. As I mentioned above, there are the frequency limitations which would likely require some interpolation to derive amplitudes for the bass freqencies below those recorded by the acoustic process. For me, the most curious potential problem is the dynamic range limitations of the acoustic recordings and more importantly, the mechanically created equalizations that were done. As we know, during the acoustic era, pianists were instructed to adjust the amplitudes they used in their playing to compensate for the nature of reproducers. It would seem that taking an acoustic recording and playing it back flat, and then deriving a midi file could produce some curious results. While it could have the potential of being an accurate reproduction of what was played, (with the aforementioned problems of the acoustic of the respective recording and playback environments, and the need to recreate the
 fundamentals of those frequencies below the available range of the acoustic recording) an accurate reproduction might be just that, an accurate reproduction of a pianist trying to compensate for the limitations of the recording technology of the time, and not much of a representation of what they intended to project. Of course one can make "appropriate" adjustments and guesstimates.
   
  As to the question of what does one preserve and how does one "catalog..." Obviously if one creates an "object" like a CD which is either a realization of the file or the file itself, it is not much of a question. Of course playback is subject to hardware and software, even with the red book audio CD. Interesting that cataloging specifications do not include the code for the software to playback audio CDs...well I suppose we don't have a rosetta stone or a dictionary included in the description of every book we catalog.
   
  As to the article on video games...I loved the part about the need for emulation software and that the access to such software could be a copyright violation...
   
  No doubt I am in the minority, but I find the notion of a "digital object" to be an oxymoron and a wonderful example of those who can't seem to let go of the old ways of thinking. At least I know I am not alone, especially on this list, when it come to thinking that our copyright laws are absurd, with the video game emulation being an excellent example. Again, it seems to me that we have too many copyright laws and methodologies predicated on those old Euclidian notions of linearity...thank you Marshall McLuhan (my source for that phrase "Euclidian notions of linearity").
   
  Kinda makes me wonder...are libraries going to circulate ebooks and iPods loaded with information? So, will we have a cataloging record for the file, then one for the object on to which we have loaded that file. I also wonder if by the time we get an AACR III we won't have as many objects being published that we can describe...
   
  Another thought on the wacky world of copyrights...I assume that any midi file derived from an old recording would be subject to the copyright of the recording used as the source for that midi file!!!!! I can hardly wait to read the court cases on those legal battles.
   
  Karl
   
   
   
   
  
Tom Fine <tflists@xxxxxxxxxxxxxxxxxxx> wrote:
  One thing related to this I've always wondered. For instance, an old piano recording, a great 
performance but a crapola 78 recording. Why couldn't modern MIDI software recreate all the subtle 
attack, decay, rhythmic eccentricities, etc that make the performance unique and then play it back 
on a good if not fantastic sounding MIDI Yamaha grand piano, for example? Not sure if this is doable 
to the level of precision I'd want, but it's an interesting thing. Perhaps one day, all low-fidelity 
recordings of great musical merit can be recreated in high fidelity. Then again, perhaps not?

-- Tom Fine


----- Original Message ----- 
From: "Marcos Sueiro" 
To: 
Sent: Tuesday, March 13, 2007 5:40 PM
Subject: [ARSCLIST] Two other N.Y. Times article on a different type of digitizing


> The first is about the new recordings of Zenph's "recreations" of performances in old recordings 
> (How are we going to note these in the metadata?)
> http://www.nytimes.com/2007/03/12/arts/music/12conn.html
>
> This one is about preserving videogames (which, of course, include sound). You may think it is 
> challenging to safely point a digital file of audio to, say, the corresponding LP cover. Imagine 
> keeping the code and machines necessary to "preserve" these:
> http://www.nytimes.com/2007/03/12/arts/design/12vide.html
>
> Marcos
> CU Libraries
>