>RE - But in the real
>world, I'm not sure if this has proven true, perhaps Jamie or David Glasser will comment -- is
>actual bias frequency usually or often about or exactly what the manufacturer specified? Even in
Not really no -- usually within 10% but never exact and bias is often oddly "wrong" - case of a
machine that was aligned via a homemade umbilical card extender and the owners had not taken
into account the capacitance of the extra cable - dropped the manufacturers spec from 150kHz to
about 74, cut it in half when they put the card back in the machine
>Once we have the wow and flutter filtered out via Plangent, we can
>apply long-term speed drift compensation using a variety of tools
>(although some of the "autotune" tools might have too much of an
yeah ixnay on the auto-tune shiznit.
so far the drift of the transport of the machine making the recording is way worse than the drift of
the bias oscillator - famous case of a huge Who hit that goes flat when played on a servo PB
machine... the bias was reliable enough to correct the pitch over 8 minutes while the recorder had
run faster and faster. Not yet seen any cases where bias drift was an issue.
>I personally know Jamie and have found the process to be effective on a number of recordings,
>both for film and audio transfers. The technology appears to be solid. - John Spencer
Thanks, John - yup - it works fine on wow and flutter and even does a great distortion reduction
on stuff that doesn't have obvious problems - that's been the harder point to get across - given
sufficient bias it works on everything - and improves imaging and clarity on tapes where the
speed and wow issues seem immaterial by reducing heretofore undiagnosed flutter at very high
rates (over 300HZ often) at surprisingly high levels, and reducing scrape flutter in the 2K-3K
region which was the holy grail for the designers of the later-day machines like the ATR.
Apropos of ATR - we started with the ATR - but there's nothing other than noisy logic to defeat it
being used on other makes and models. Probably an A80RC with no logic would be really good
since it would be naturally very quiet. But the point is, it works, and hasn't been stumped yet.
>>In an earlier post you asked about staggered heads - David G referring to Richard H question
For most cases I bet picking one channel as the source for correction and some delay would be
helpful - though best would be to capture both separately... there would need to be some
finagling to make sure they tracked roughly and filter above that point - some correlation tricks
we could try... The problem with multiple heads is on playback - if something snags say at a guide
then the stall appears everywhere at once - and if theres a delay in the circuit to compensate for
head spacing then said delay is inappropriate for the snag case and and actually throws a wow
"hit" into the digitization where no wow exists and doesn't correct the one that happened. For
something like the staggered head I probably would just grab the earlier side, try some
intelligence to filter out what FM was common to both simultaneously (that would NOT want to be
delayed) then whatever was left delay it to process the other side - probably for safety lpf the
flutter repair based on the such that it couldn't be fooled. Could be done
>P. Maziewski: "Wow defect reduction based on interpolation techniques",
>Bulletin of the Polish Academy of Sciences, Technical Sciences, Vol. 54, No.
>4, 2006. per Rob Peretti
Very sore subject.
This is a direct lift of the work that Dr Patrick J Wolfe (ex Cambridge UK, now Harvard) did for us,
the Poles have been following on our heels ever since we demo'd the system in Berlin at AES in
2003. Sincere flattery I guess, but I raised hell with them and I'm disappointed that we've never
been contacted by PRESTO. They have lots of lovely EU tax money to play with to duplicate and re-
capitulate our work. We have the system complete and ready to go, but they seem intent upon
knocking it off.
Thanks for all the interest, and any other questions I'm sure Dave would be happy to answer or hit
me up at [log in to unmask]