After a bit of coaching to read the MJ article more carefully, I've
updated the information I posted yesterday about SF Edit and placed it
in the FAQ:
Sony introduced Scale Factor Edit in 1999 (the first machine was
the http://www.minidisc.org/part_Sony_DHC-MD575.html DHC-MD575
bookshelf system). Scale factor edit is a function that changes
the volume level of previously recorded audio by changing all the
scale factors in each soundgroup. Users can adjust a whole track's
level, or taper the beginning and ending of tracks. It's a much
cheaper operation computationally than decoding the full waveform,
scaling it, and then recoding it.
The http://www.minidisc.org/mj_ja3es.html MJ Magazine MDS-JA3ES
article provides a good basis for understanding how this feature
works. ATRAC (and other modern audio coders) store audio in the
frequency domain. The samples are stored as floating point
numbers, with an exponent and a mantissa. The so-called Scale
Factor is the exponent, which is stored in 6 bits, giving 64 (2^6)
possible values and yielding final sample values in the range of
-120dB to +6dB. Each Scale Factor step is +/-2dB.
Scale Factor Edit does have a limitation: If you use Scale Factor
Edit to decrease volume past the point at which some Scale Factors
in the signal become zero, those zero values cannot be restored if
Scale Factor Edit is later used to increased volume. Likewise, if
Scale Factor Edit is used to increase volume past the point at
which some Scale Factors "max out" at the highest Scale Factor
value, previously distinct Scale Factor values will merge and the
signal's dynamic range (not to say fidelity) will decrease
unrecoverably.
Rick
-----------------------------------------------------------------
To stop getting this list send a message containing just the word
"unsubscribe" to [EMAIL PROTECTED]
