| Tue, 22 Sep 2015 04:09:13 +0100 |
Giulio Moro |
Now leveraging BBB's lock between xenomai clock and audio clock for ultra-accurate, low-latency clocking. CAVEAT: fractions of samples drifts will occurr every time the clock is changed
ClockSync
|
| Mon, 21 Sep 2015 03:12:21 +0100 |
Giulio Moro |
Added simple one-variable one-measurement Kalman filter, Pid controller(which output is not used). Virtual clock is now much more precise and reactive for period. Still it is lagging behind a bit on the overall offset.
ClockSync
|
| Mon, 21 Sep 2015 03:11:32 +0100 |
Giulio Moro |
Simple IirFilter class
ClockSync
|
| Tue, 15 Sep 2015 02:01:15 +0100 |
Giulio Moro |
Something weird happened when pllP=8, so I set pllP=7, which reduces (a bit) the clock adjustment precision
ClockSync
|
| Mon, 14 Sep 2015 21:39:54 +0100 |
Giulio Moro |
Merge
ClockSync
|
| Mon, 14 Sep 2015 21:38:09 +0100 |
Giulio Moro |
Further updates and refactoring of the I2c_Codec class
|
| Mon, 14 Sep 2015 21:31:09 +0100 |
Giulio Moro |
Added ifdefs for USE_JUCE compatibility
ClockSync
|
| Mon, 14 Sep 2015 17:35:18 +0100 |
Giulio Moro |
Merged
ClockSync
|
| Mon, 14 Sep 2015 17:31:24 +0100 |
Giulio Moro |
Added I2c methods to set R and P values for the PLL
|
| Mon, 14 Sep 2015 15:42:11 +0100 |
Giulio Moro |
Ultra-basic feedback for clock sync works^CIssues: response time of the IIR filter is too slow, requires PID and better filtering algorithm.
ClockSync
|
| Mon, 14 Sep 2015 14:57:54 +0100 |
Giulio Moro |
MovingAverage: added reset() method and edited getLength() to return current length
ClockSync
|
| Sun, 13 Sep 2015 21:34:47 +0100 |
Giulio Moro |
The reported offset is now meaningful. The whole thing is waaay too jittery.
ClockSync
|
| Sun, 13 Sep 2015 21:33:01 +0100 |
Giulio Moro |
Added xenomai version of Clock, now includes header
ClockSync
|
| Sun, 13 Sep 2015 21:31:37 +0100 |
Giulio Moro |
MovingAverage now returns a meaningful value also when the buffer is not full
ClockSync
|
