home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!caen!uwm.edu!csd4.csd.uwm.edu!info-high-audio-request
- From: jas@proteon.com (John A. Shriver)
- Newsgroups: rec.audio.high-end
- Subject: SPDIF jitter follow-ons
- Date: Mon, 16 Nov 92 10:42:09 EST
- Organization: University of Wisconsin - Milwaukee
- Lines: 22
- Approved: tjk@csd4.csd.uwm.edu
- Message-ID: <1eatflINN460@uwm.edu>
- NNTP-Posting-Host: 129.89.7.4
- Cc: winalski@adserv.enet.dec.com
- Originator: tjk@csd4.csd.uwm.edu
-
- Yeah, I almost left something in about using gigantic FIFO's in the
- DAC box.
-
- The problem with this is simple. Sort of like non-Euclidian geometry,
- no two ocsillators run at the same speed. The CD player will be
- encoding with one frequency down the SPDIF link, and the DAC will be
- draining the FIFO at some different frequency. The FIFO will either
- overflow or underflow.
-
- The tolerances on CD clocks, even by the specifications, aren't that
- tight. Moreover, some CD players violate these specifications
- flagrantly. One cheap Japanese CD player recently reviewd in HFN&RR
- used a ceramic resonator instead of a quartz oscillator, and was so
- far off-frequency that the audio output was over quarter-tone flat!
-
- If you used very precise clocks, and a gigantic FIFO, and were willing
- to tolerate a long startup delay until the FIFO was half-full, you
- might be able to make this work. However, it would cost more than the
- two-cable clock-feedback scheme, which eliminates the clock problem.
-
- However, a 32 bit FIFO would over- or under-run in seconds to minutes.
-
-