JackOfAll wrote: 
> Triode,
> 
> Was trying to figure out the logic behind the magic numbers, 23 and 37,
> found this....
> 
> > 
Code:
--------------------
  >   > 
  > /*
  > * spdif0_clk will be 454.7MHz divided by ccm dividers.
  > *
  > * 44.1KHz: 454.7MHz / 7 (ccm) / 23 (spdif) = 44,128 Hz ~ 0.06% error
  > * 48KHz:   454.7MHz / 4 (ccm) / 37 (spdif) = 48,004 Hz ~ 0.01% error
  > * 32KHz:   454.7MHz / 6 (ccm) / 37 (spdif) = 32,003 Hz ~ 0.01% error
  > */
  > 
--------------------
> > 

That's rather interesting. That means the SPDIF output is using the USB
PLL rather than the audio PLL. The audio PLL has a fractional N
synthesizer with 30 bits precision, which means it can be set to almost
any frequency desired to a high precision. With this you can get the
clockroot feeding SSI and SPDIF sections to be the standard audio clock
frequencies. The table on page 5022 of the manual shows how to do this
for different frequencies. 

The actual PLL registers take three numbers, a divider select, a
numerator and a denominator. The equation is:

fout = 24MHz * (div_sel + (num/denom)) The fraction should be between
0-1.0. 

There are then a series of dividers to get the audio frequency clock
(say 11.2896), then the SSI or SPDIF block divide this down to the
actual bit clock and word clock. 

I just don't understand why we are trying to come up with "magic numbers
" to get close to the real thing, when by using the audio PLL you CAN
get the real thing. The only reason I can think of is that somebody
wanted different sample rates on I2S and SPDIF. 

John S.

It seems to me this would be better than using a USB PLL frequency that
is not a nice multiple of the audio frequencies.


------------------------------------------------------------------------
JohnSwenson's Profile: http://forums.slimdevices.com/member.php?userid=5974
View this thread: http://forums.slimdevices.com/showthread.php?t=98544

_______________________________________________
unix mailing list
[email protected]
http://lists.slimdevices.com/mailman/listinfo/unix

Reply via email to