Rick explained further,
| Btw, dB is computed using log base 10, if you use the natural log, you
| will need to multiply the result by 0.4343. This would reduce the
| difference between what the two programs report to ~0.05 dB.
Rick, you know I'm old enough that my education came back in the days when
common logarithms were useful mathematical aides and not just the obsolete
curiosities they are today, so I know the difference, and I didn't mista-
kenly use natural logs.
The differences were generally about .02 of a dB, not enough to affect a
recording, just enough to leave me scratching my head.
GoldWave reported the peak amplitude to four decimal places instead of three
(like Exactaudiocopy) or two (like Audiograbber). I discovered that the
other two were always rounding *down* to reduce the number of decimal places.
Likewise, Cool Edit was rounding "down" (that is, toward a smaller negative
number, nearer zero) in trimming its peak in decibels to two decimal places,
so the roundings went in the opposite directions. The decibel figures were
always rounded louder and the peak amplitude figures always rounded softer,
and the disparity was *always* that the twenty times the common log of the
peak amplitude was softer than the direct figure in decibels. That probably
accounted for most of the differences, which usually were .01 or .02 dB,
sometimes .03.
On one track -- "Scotch and Soda" from the Capitol Collectors' Series best-of
of the Kingston Trio -- where the gap was .06 dB, and I haven't figured that
one out yet.
| How about just hedging the Exactaudiocopy's report upwards by 0.5 dB?
| You're not going to notice this.
Actually, I could hedge by just .1 of a dB and always be safe, and certainly
no one is going to notice that. I was mostly confused about why there was
always a discrepancy, and I think that rounding errors account for it, except
maybe on "Scotch and Soda."
-----------------------------------------------------------------
To stop getting this list send a message containing just the word
"unsubscribe" to [EMAIL PROTECTED]
