[78-L] Wow/flutter, was Re: Cambridge Process

[Doug Pomeroy]

Hi,

Yes, Howarth's process looks at the bias frequency on tape, and often
he does this playing the tape at half speed.   With audio from disc
sources he is sometimes able to find other frequencies to track, such
as cutter turntable rumble, which can sometimes be multiplied and
used for pitch correction.  And, as you say, the sources must be the
original recordings, not copies.

Wow and flutter which is perfectly regular can often be corrected, once
the period of the fluctuation has be identified.  I have worked with two
mathematicians who can create algorithms for such jobs  -  Kevin Short,
who worked with Howarth on the recently discovered wire recording by
Woody Guthrie, and Dimitri Antsos, of Cal Tech, who has done work for
me on problem audio from metal part transfers.

Programs developed to correct off-pitch singers (AutoTune and Melodyne
from Celemony) require a monophonic input signal, where only one
instrument or voice is heard, sounding only a single note. However there
will be a new version of Melodyne which is said to overcome this  
limitation
in some cases.

I'm told one of the audio restoration programs (Sound Forge?) includes
a graph on which a line can be drawn which will effect a corresponding
correction of pitch.  I assume using this would require much trial  
and error,
and might work for gradual pitch changes over time, but not for very  
rapid
ones.

Doug Pomeroy
POMEROY AUDIO
Audio Restoration & Mastering Services
Transfers of metal masters, lacquers,
shellac and vinyl discs & tapes.
193 Baltic St
Brooklyn, NY 11201-6173
(718) 855-2650
pomeroyaudio at att.net
=======================================================

> Message: 15
> Date: Tue, 16 Dec 2008 20:28:14 +0100
> From: Chris Zwarg <doctordisc at truesoundtransfers.de>
> Subject: Re: [78-L] Wow/flutter, was Re:  Cambridge Process
> To: 78-L Mail List <78-l at klickitat.78online.com>
> Message-ID: <0MKwpI-1LCfaZ1MwR-0002k7 at mrelayeu.kundenserver.de>
> Content-Type: text/plain; charset="us-ascii"
>
> At 20:18 16.12.2008, you wrote:
>
>> Try Jamie Howarth at Plangent Processes.
>> www.plangentprocesses.com
>>
>> Doug
>>
>
> Yes that's the one I was thinking of. Of course this method will  
> only help if the wow/flutter is a defect of the actual tape you  
> have - dubbed wow will not be corrected as the process does not  
> analyze the musical content itself (has there been any attempt of  
> doing that for w/f correction? I imagine it could work similarly to  
> the "autotune" gadgets commonly used by pop singers with not-so- 
> perfect intonation).
>
> Chris Zwarg