> I4ve some questions about the cpuload instrument statement and how a SA
> decoder reacts on limited CPU power:
I'll leave these to others who can answer them better, but just thought
I'd chime in to note that I just finished implementing cpuload a few
days ago in sfront, so expect it in sfront 0.62 at the end of the month.
So far, it's quite a cool variable to look at, I've been making plots of
cpuload as a function of time for some of the examples (PC.saol is a good
one for this experiment), and the cpuload(t) seems to reflect well what
I'm hearing out the speakers as far as artifacts and real-time slippage
goes -- I could easily imagine modifying PC.saol to do note dropping
using this number, although to what degree the scheme would be "normative"
in practice depends on how different cpuload(t) is going to be for
different decoders, especially the windowing and time-smoothing ...
--jl
-------------------------------------------------------------------------
John Lazzaro -- Research Specialist -- CS Division -- EECS -- UC Berkeley
lazzaro [at] cs [dot] berkeley [dot] edu www.cs.berkeley.edu/~lazzaro
-------------------------------------------------------------------------
This archive was generated by hypermail 2b29 : Mon Jan 28 2002 - 11:46:40 EST