[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Date Index][Thread Index][Author Index]

Re: digital recording mystery



Now ya' got me thinkin'! 
 
Interesting problem...you've convinced me that the minidisc is the major culprit.
 
So I ran a couple of experiments.  (I used my Kyma system for all of this, operating it at 44.1 kHz sampling rate, and a Sharp MD MS-702.  Your milage may vary.) 
 
Experiment 1:
I generated a 20 kHz sine wave of 4 seconds duration.  I recorded it to minidisc.  I played it back and re-recorded it.  Items of note: 1) The playback signal was actually longer than 4 seconds.  2) The signal played back as a 20 kHz sine wave.
 
I'd expect the signal to be slightly longer due to filtering so (1) doesn't surprise me too much.  (2) is due to the adaptive nature of ATRAC.  I understand you can record a pure sine wave and play it back unchanged (well pretty much unchanged anyway considering the ADC and DAC).  And this experiment confirmed that.  But consider that there is not the storage capacity on the minidisc for a full bandwidth signal.  I.e., on a 74 minute minidisc, I can't record 74 minutes of a 20 kHz sine wave.  There's simple not enough room as Nyquist states.
 
Experiment 2:
Same procedure repeated with white noise instead of a sine wave.  1) I still see the playback signal as longer than 4 seconds, but only slightly.  2) The playback signal has been limited to 18 kHz.  There's good evidence of two filters, the first at approximately 17 kHz and the second at at 18 kHz.
 
Four seconds is not very long.  One of these days, I should repeat the experiment with real music of four minutes or more.
 
Dennis Leas
-------------------
dennis@mdbs.com