Pristine Audio XR - doubling the frequency range of 78s
Andrew Rose
Extended Range (XR) recordings which, using a combination of our
existing Natural Sound techniques and new targeted audio extraction
methods, has succeeded for perhaps the first time in restoring the 'lost
frequencies' from pre-ffrr electric 78s between approx. 6kHz and ~12kHz
or higher.
Despite the huge selective amplification required to achieve this, the
results are nothing short of phenomenal. We currently have six examples
posted on our homepage, alongside further details about the technique.
Later this week we will be adding the first five of a series of
recordings originally released by _Music and Arts of America_ (and now
OOP) - ex-acetate 30's and 40's recordings made by Szigeti, Furtwangler
and Toscanini, playing music by Bach, Beethoven, Brahms, Mozart and R.
Strauss. These recordings have all been restored using our new XR
technology.
I invite you to take a listen at http://www.pristineclassical.com
--
Andrew Rose - Pristine Classical
The online home of Classical Music: www.pristineclassical.com
Otterhouse Rolf
> I'm pleased today to announce the launch of the first Pristine Audio
> Extended Range (XR) recordings which, using a combination of our
> existing Natural Sound techniques and new targeted audio extraction
> methods, has succeeded for perhaps the first time in restoring the 'lost
> frequencies' from pre-ffrr electric 78s between approx. 6kHz and ~12kHz
> or higher.
>
> Despite the huge selective amplification required to achieve this, the
> results are nothing short of phenomenal. We currently have six examples
> posted on our homepage, alongside further details about the technique.
>
> Later this week we will be adding the first five of a series of
> recordings originally released by _Music and Arts of America_ (and now
> OOP) - ex-acetate 30's and 40's recordings made by Szigeti, Furtwangler
> and Toscanini, playing music by Bach, Beethoven, Brahms, Mozart and R.
> Strauss. These recordings have all been restored using our new XR
> technology.
>
>
> --
> Andrew Rose - Pristine Classical
>
> The online home of Classical Music:www.pristineclassical.com
And now this technique with acoustical records..... :)
Do you extract the interference of higher (non-recorded) overtones to
reconstruct them? When I worked at Philips, we used this on small
scale speakers.....
Rolf
Lionel Tacchini
> I'm pleased today to announce the launch of the first Pristine Audio
> Extended Range (XR) recordings which, using a combination of our
> existing Natural Sound techniques and new targeted audio extraction
> methods, has succeeded for perhaps the first time in restoring the 'lost
> frequencies' from pre-ffrr electric 78s between approx. 6kHz and ~12kHz
> or higher.
How does it work ?
Lionel Tacchini
William Sommerwerck
Ever heard of an Aphex Aural Exciter? It generates musically related
harmonics.
Andrew Rose
The frequencies are there - they always were, usually so buried in the
background noise as to be useless. But if you know where to look, how to
look, and how to extract them, they can be restored alongside the rest
of the 'regular' audio.
If that appears a little thin on the detail, forgive me - I don't wish
to give away too many secrets for the time being. Suffice to say we are
not using any method of artificial synthesis or regeneration to create
anything here - what you hear on the examples posted is what's come off
the shellac discs themselves, carefully amplified back to appropriate
levels and largely clean of hiss and noise.
Of the examples posted, the Schnabel Emperor gave the most trouble and
is perhaps less successful than the rest, though this has more to do
with overall distortion and compression at certain frequencies in the
original. There are also a couple of sides where I had to revert to UK
rather than quieter US pressings, which leaves higher noise levels.
Meanwhile the Vaughan Williams is posted as an example of using aspects
of the technique on a post-ffrr 78rpm recording - clearly there's much
mileage to be had here as well.
As for the rest - well the Kreutzer is a personal favourite. One person
who's already heard it asked me about microphone movement between
movements or takes, not realising that it had been recorded in two
locations on different days. That's the kind of detail I don't think
many have been able to hear before.
Die Schone Mullerin I've chosen to present as individual songs, with
fades between each, rather than risk compromising the excellent sound by
matching hiss at the start and end of each side to create continuity.
The Weingartner Eroica I was particularly impressed with - it was the
first one that we succeeded in doing this with, and left me speechless
in amazement at what I was hearing.
Finally the Long Fauré is taken from Decca's own 33rpm transcriptions of
her 78s, and as such probably pre-dates as well as post-dates the
introduction of ffrr, not that you'd necessarily guess the recordings
dates now...
Andrew Rose
The frequencies are there - and there's no aural exciter being used. The
problem with that idea (which has been suggested many times) is that it
creates more problems than it solves, producing hissy and entirely
unrealistic results. It is also disastrous on non-pop vocals, where for
much of the time you really don't want too much other than sibilances
above about 5kHz...
FWIW if you look closely at Mark Obert-Thorn's restoration of the
Cortot-Thibaud Kreuzter, which he did from vinyl test pressings, you'll
detect the same frequencies I've been able to isolate from the shellac.
As I say - they're there, you just need to know where to look and how to
get them out without turning everything into a hissy mess.
Anyway - have a listen and see what you think.
Andrew Rose
>
> And now this technique with acoustical records..... :)
I'd love to, but there really doesn't seem to be anything there to be
had at all - certainly I've not managed to track it down and extract it
using these methods, beyond a very slight possible extension that I'm
not really convinced is anything new.
I realise that some people are going to take a while to believe this is
for real - unless they take a good listen to the results - but all I've
done is figure out how to find and amplify frequencies that were
recorded. Getting at them is extremely difficult, which is one very good
reason why it's not been successfully achieved before - I spent 6 months
on this a couple of years ago before giving up and deciding it was a
waste of time. But technology moves on...
Lionel Tacchini
>
> The Weingartner Eroica I was particularly impressed with - it was the
> first one that we succeeded in doing this with, and left me speechless
> in amazement at what I was hearing.
There is a significant difference in sound with the Naxos transfer,
although this was already very good.
I wish we would get new transfers of the early Bruckner recordings,
especials Fried's and Horenstein's 7th.
Lionel Tacchini
William Sommerwerck
potential for thoroughly removing the noise...
Andrew Rose
> If you can extract frequencies that are actually on the disk, there is a
> potential for thoroughly removing the noise...
Up to a point that's already happening with the Natural Sound technique
as well as XR - indeed it's what helps make this possible. However there
is of course noise at all frequencies, including the ones you do want,
and at these higher frequencies it's much more pronounced.
Once you add in pitch fluctuations due to vibrato, portamento, poor
tuning (there's a very nasty trombone fluff towards the end of the 1st
movement of Schnabel's Emperor!) and inevitable wow and flutter in
analogue systems, all of this becomes a lot less straightforward than it
might initially appear.
Rich S.
If so maybe it can be applied to rmcr...
Andrew Rose
commercially available transfers here:
http://www.pristineclassical.com/More/NaturalSound.html
dk
> On Mon, 9 Apr 2007 05:24:03 -0700, "William Sommerwerck"
>
> >If you can extract frequencies that are actually on the disk, there is a
> >potential for thoroughly removing the noise...
>
> If so maybe it can be applied to rmcr...
You gotta be kidding! ;-)
dk
Andrew Rose
Technology has its limits... ;-)
Andrew Rose
> Do you extract the interference of higher (non-recorded) overtones to
> reconstruct them?
I just wanted to pick up on this point - the 'non-recorded' upper
frequency assumption is actually totally invalid, and it's a dangerous
assumption to make, because it means you don't bother working with the
frequencies which really are there.
For pre-ffrr electrical 78s, engineers used (analogue!) bandpass filters
prior to recording between about 50Hz and 6000Hz. Now it's safe to
assume that these were not brick-wall filters, but that they rolled off
the upper frequencies smoothly up the range, with those upper harmonics
gradually disappearing into the background noise.
(This incidentally explains why there remain some people who'd rather
hear original 78s over any restorations they've ever heard - they're
hearing those upper harmonics which restoration traditionally ignores or
filters out.)
If you use software such as Adobe Audition, which has a spectral as well
as a waveform view, to look at straight transfers of 78s, you'll be
surprised to see just how high the harmonics go - it's just that they're
diminished and effectively buried by the surrounding noise. The signal
to noise ratio is simply too poor to allow them to be exposed using
regular noise reduction techniques.
What XR processing does is help to improve that S/N ratio prior to noise
reduction - if you can get enough separation between the harmonics and
the background noise you can use that audio and restore it to its proper
balance.
I've not managed, in the case of the Kreutzer, for example, to bring out
the very highest recorded frequencies. Spectral analysis of the final
movement, for example, shows harmonics up to about 16kHz, whereas I've
only extended up to about 12kHz. I have a later recording I'm currently
working on (from the 1940's) which indicates a sharper roll-off curve
was used than in that 1929 Kreuzter recording, and I'm only seeing
harmonics up to about 12.5kHz.
But they are most definitely there, and certainly can be extracted, as
the XR-processed results demonstrate.