On Jan 30, 2011, at 4:57 PM, Fred Thal wrote:

> Concerning the correction of possible equalization errors found in
> tapes recorded on a machine running off of desired speed, we first
> must know the conditions under which that recording machine was
> calibrated. Because one first performs a playback response
> calibration, before performing the record calibration, it is of
> interest to know if the machine was running off-speed when the
> playback response curve was set (yes or no). Two different sets of
> conditions can follow from this. Yet we seldom (if ever) have such
> information about the recorder’s calibration.

      Quite right.  It would be interesting and plausible both for a  
studio (in the seventies?) to have calibrated their machine long  
before some
age or wear-related issue caused a slower operation of the  
transport.   Believing their own machine to sound good, in-house,  
they might not have re-calibrated
using their alignment tape at the newer version of the standard speed  
(sluggish) or some elevated rate caused by various reasons which  
might have crept in after
the sole calibration was done (at proper speed).

      However, my question, however clumsily posed, was trying to ask  
about a machine that was believed to have been calibrated at the  
wrong speed and recorded on at that same wrong speed, to be played  
back, however, at a later time, on a different machine that is  
capable of a variety of playback speeds...

    [a snippet of how I tried to phrase that, below:]

"... Would it follow that, if a tape had apparently been recorded at  
an amazing 1.5x speed (i.e., 5.6 ips based on a fast-played 3.75 ips  
NAB Cal tape (presumably) at time of recording), those 7.5 ips- 
to-3.75 ips amplitude offsets would need be halved from where they  
are with the 2x (i.e., 7.5 ips) Cal tape...)"

    [note the "fast-played 3.75 ips NAB Cal tape (presumably) at time  
of recording..."]


Thank you for your other information, too, Mr. Thal...    It makes me  
feel less worried about the use of a vari-speed (under typical  
amounts of variance), without resorting to slide rulers.   At worst  
case scenario (2x or 1/2x), there's always the charts (as Richard  
pointed out).   It's nice to have a better sense of the curve, and  
how to work around it, so to speak.





Cheers,
     Andrew








>
> Andrew Hamilton:
>
>> by speeding up the playback machine to 15.2 ips - enough to
>> make the intended 1k tone actually play 1000 cps - does the
>> vari-sped playback (at the original record speed) magically
>> de-emphasize the haunted pre-emphasis?
>
> Yes, because the complementary record and playback equalizations are
> rooted in the frequency domain, they are (in a practical sense) blind
> to absolute tape speed and drift from nominal (and resulting changes
> to recorded wavelengths), provided those speed changes occur equally
> in BOTH recording and reproducing (playback).
>
>> What is the point at which an alteration should be made to
>> the pre- and/or de-emphasis EQ?
>
> Let's first simplify it by assuming there was no error introduced in
> the recorder's record equalization calibration by virtue of someone
> having first performed that machine's reproduce alignment with the
> machine running badly off speed.
>
> Then, there will be NO such point, IF the playback speed is matched to
> the record speed. Under this condition (which should always be the
> goal) the problems you are imagining disappear.
>
> Now, in the general case where the playback speed does NOT match the
> record speed, you are correct that a playback equalization
> (de-emphasis) tracking error will be introduced. As someone else
> responded, in most all practical examples experienced, this error is
> small enough to be negligible.
>
>> provided one had a way of dialing in the exact speed to make
>> the reproducer play back at the (actual) speed of the recorder
>
> You are correct to be emphasizing this, as it is one of the first
> things one needs to do in tape playback. Adjusting to known (or
> assumed) musical pitch, resolving to known (or assumed) recorded power
> line frequency hum, or possible other recorded pilot signal, including
> recording bias, are powerful tools for doing this.
>
> Fred Thal
> Audio Transfer Laboratory
> ATAE Studer