|
Hi James,
Oops, I just synch'ed and see you have committed a change on this also. I think using SystemClock is better as using a timestamp with unsynched clocks has unknown results (which is what happens with non-zero lag and nil latency in your approach).
RJK On Jan 20, 2008, at 7:44 AM, ronald kuivila wrote: Hi James and Alberto,
I just committed the following:
schedBundle: #{ |lag, offset, server ...bundle | thisThread.clock.sched ( offset, { if (lag !=0 ) { SystemClock.sched( lag, { server.sendBundle(server.latency, *bundle) }) } { server.sendBundle(server.latency, *bundle) } }) },
schedBundleArray: #{ | lag, offset, server, bundleArray | thisThread.clock.sched ( offset, { if (lag !=0 ) { SystemClock.sched(lag, { server.sendBundle(server.latency, *bundleArray) }) } { server.sendBundle(server.latency, *bundleArray) } }) },
On Jan 20, 2008, at 7:37 AM, ronald kuivila wrote: Hi Alberto and James,
This is my fault. The problem is that lag is supposed to be realtime. I will change this to use SystemClock to schedule the lag and then Server.latency is used without math.
RJK On Jan 19, 2008, at 10:21 PM, James Harkins wrote: On Jan 19, 2008, at 7:16 PM, Alberto de Campo wrote: just tested again how latency nil behaves:
1. patterns break now with latency nil (they did play recently)
That's a side effect of some of the event refactoring. I'll commit a fix in a minute. It's an easy one.
2. events seem to get quantized to the next hardwareBufferSize!
request: A. I would much prefer having them start at the next block border, and B. ideally, have OffsetOuts start with their offset time into the block, so they start less time-quantized.
Now that's weird. Maybe that was always the case? I can't answer to that one. hjh
: H. James Harkins .::!:.:.......:.::........:..!.::.::...:..:...:.:.:.:..:
"Come said the Muse, Sing me a song no poet has yet chanted, Sing me the universal." -- Whitman
_______________________________________________ Sc-devel mailing list _______________________________________________ Sc-devel mailing list
_______________________________________________ Sc-devel mailing list
|