fisonic

Since I sometimes get asked, how exactly I obtained track durations for vinyl releases, here's a description:

(1.1) Direct timing from record
I am listening to the record via a quartz-locked, direct-driven, turntable (Technics SL-QD 33, effective rotational accuracy within 0,025% total WRMS, according to manufacturer data), and am using the stop-watch function of a digital quartz watch to measure the time from the first audible sound of a track to the very last audible bit.
Depending on the nature of the music - i.e. fade-in/out or not - I will dynamically turn the volume very high at the very beginning and the very end of a song. Thus the only potential systematical inaccuracy would be the inherent background noise level of my stereo system and the recording itself.
Fractions of a second provided by the stop watch will be rounded to the nearest whole second.

(1.2) Comparison with printed information
In case the release has durations printed on the sleeve or labels, I will use that information, as long as it is within a reasonable margin of error to the duration I obtained in (1.1).(*)

(2) Normalization (optional)
Sometimes, when a version of a song is available on other releases of the same title, and when those releases consistently have durations printed on them, which are within a reasonable margin of error to the duration I obtained in (1.1)(*), I might use those officially given times, for the sake of consistency. But only, if I verified that both tracks are exactly the same and both have the same actual duration.

(*) As an acceptable margin of error, I usually regard a tolerance of less than 2% of the absolute track duration, with an absolute maximum of 10 seconds for tracks longer than 8:20.

This all sounds much more complicated when I describe it, as it actually is. ;-)
And of course, I am not free of making mistakes. So, if you have reason to doubt a particular duration I submitted, please don't hesitate to tell me about it.