Re: [ARSCLIST] 33 to 78

[Eric Jacobs <ericj@xxxxxxxxxxxxxxxxxxx>]

I have to re-pitch many of the transfers that I do, particularly
dictations and voice recordings made on mechanical media where
consistent speed was rare but often hum or other tell-tale
mechanical frequencies are present for reference.  Just a few
Hz can make a noticeable difference in terms of the natural
tone and timbre of the speaker's voice.

Re-pitching is a little extra work, but definitely worth it.

Having a precise re-pitch value is important.  Messing around
with look-up tables is not accurate - it's slow, and interpolation
of values is non-linear (so the look-up table gets you close, but
not dead on).

Getting the pitch shift right on the first try is really the
goal.  You want to minimize how much digital processing you
perform on a file if you can, so trial and error is not ideal
both in terms of efficiency and results.

Here's the magic speed/Hz to semitone calculator:

F1 = current speed/freq (in RPM or Hz)
F2 = new speed/freq (in RPM or Hz)

Shift (in semitones) = 12 * log (F2/F1) / log (2)
Shift (in cents) = Shift (in semitones) * 100

No more need for table look-up with this equation!

Along with re-pitch, it is important to have accurate equations
for digital re-EQ.  You don't want to be estimating EQ values by
interpolating from a NAB, RIAA, or other graph.  It's much
better to have equations where you can input turnover/roll-off
values and get direct gain values for the frequency of interest.
But that's another subject for another day...

Eric Jacobs
The Audio Archive


---- Original message ----
>Date: Thu, 7 Jul 2005 17:32:05 -0400
>From: Steven Smolian <smolians@xxxxxxxxx>
>Subject: [ARSCLIST] 33 to 78
>To: ARSCLIST@xxxxxxxxxxxx
>
>I'm having a problem dubing a bunch of 78s which skip a lot at that speed.
>
>They seem to play through at 33.
>
>Eq aside, if I record at 33, how many half-tones do I need to bump up the
sound file in the computer to get to 78.26 or thereabouts?
>
>Steve Smolian
>________________
>No virus found in this outgoing message.
>Checked by AVG Anti-Virus.
>Version: 7.0.323 / Virus Database: 267.8.10/43 - Release Date: 7/6/2005