RCS file: /Users/implicit/Desktop/Source
`Code/cvs_strings/beads/audiosync/main/audiosync.c,v
`Working file: bdk/beads/audiosync/main/audiosync.c
`head: 1.26
`branch:
`locks: strict
`access list:
`symbolic names:
`BUILD_20060123: 1.26
`BUILD_20050908: 1.26
`BUILD_20050817: 1.26
`BUILD_20050722: 1.26
`BUILD_20050718: 1.26
`BUILD_20050627: 1.26
`BUILD_20050605: 1.26
`TRAVIS_20050527: 1.26.0.22
`dev_NewSchema-branch: 1.26.0.20
`rd_GuiToolkit-branch: 1.26.0.18
`ALL_20050422: 1.26
`TRIO_20050418: 1.26
`TRIO_20050404: 1.26
`blackfin-branch: 1.26.0.16
`TRIO-20041130: 1.26
`RADKIT_GOLD_0042: 1.26.0.14
`BEADS_SILVER_0056: 1.26
`RADKIT_GOLD_0041: 1.26.0.12
`BEADS_SILVER_0055: 1.26
`RADKIT_GOLD_0040: 1.26.0.10
`BEADS_SILVER_0054: 1.26
`BANDON_20040413: 1.26.0.8
`RADKIT_GOLD_0039: 1.26.0.6
`BEADS_SILVER_0053: 1.26
`BANDON_20040329: 1.26.0.4
`RADKIT_GOLD_0038: 1.26.0.2
`BEADS_SILVER_0052: 1.26
`RADKIT_GOLD_0037: 1.25.0.52
`BEADS_SILVER_0051: 1.25
`RADKIT_GOLD_0036: 1.25.0.50
`BEADS_SILVER_0050: 1.25
`RADKIT_GOLD_0035: 1.25.0.48
`BEADS_SILVER_0049: 1.25
`RADKIT_GOLD_0034: 1.25.0.46
`BANDON_20031224: 1.25.0.44
`BANDON_20031219: 1.25.0.42
`
`Page 1 of 7
`
`Implicit Exhibit 2016
`Sonos v. Implicit, IPR2018-0766, -0767
`
`

`

`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`BANDON_20031214: 1.25.0.40
`BANDON_20031204: 1.25.0.38
`BANDON_20031103: 1.25.0.36
`BANDON_20031024: 1.25.0.34
`BANDON_20031014: 1.25.0.32
`BEADS_SILVER_0048: 1.25
`BEADS_SILVER_0047: 1.25
`RADKIT_GOLD_0033: 1.25.0.30
`BANDON_20030903: 1.25.0.28
`BEADS_SILVER_0046: 1.25
`RADKIT_GOLD_0032: 1.25.0.26
`BEADS_SILVER_0045: 1.25
`RADKIT_GOLD_0031: 1.25.0.24
`BEADS_SILVER_0044: 1.25
`RADKIT_GOLD_0030: 1.25.0.22
`BANDON_20030815: 1.25.0.20
`BEADS_SILVER_0042: 1.25
`RADKIT_GOLD_0029: 1.25.0.18
`BEADS_SILVER_0041: 1.25
`RADKIT_GOLD_0028: 1.25.0.16
`BEADS_SILVER_0040: 1.25
`BEADS_SILVER_0039: 1.25
`RADKIT_GOLD_0026: 1.25.0.14
`BEADS_SILVER_0038: 1.25
`RADKIT_GOLD_0025: 1.25.0.12
`BEADS_SILVER_0037: 1.25
`RADKIT_GOLD_0024: 1.25.0.10
`BEADS_SILVER_0036: 1.25
`RADKIT_GOLD_0023: 1.25.0.8
`BEADS_SILVER_0035: 1.25
`RADKIT_GOLD_0022: 1.25.0.6
`BEADS_SILVER_0034: 1.25
`RADKIT_GOLD_0021: 1.25.0.4
`BEADS_SILVER_0033: 1.25
`RADKIT_GOLD_0020: 1.25.0.2
`BEADS_SILVER_0032: 1.25
`RADKIT_GOLD_0019: 1.23.0.2
`BEADS_SILVER_0031: 1.23
`RADKIT_GOLD_0018: 1.22.0.4
`BEADS_SILVER_0030: 1.22
`RADKIT_GOLD_0017: 1.22.0.2
`BEADS_SILVER_0029: 1.22
`RADKIT_GOLD_0016: 1.21.0.2
`BEADS_SILVER_0028: 1.21
`
`Page 2 of 7
`
`Implicit Exhibit 2016
`Sonos v. Implicit, IPR2018-0766, -0767
`
`

`

`RADKIT_GOLD_0015: 1.20.0.4
`
`BEADS_SILVER_0027: 1.20
`
`RADKIT_GOLD_0014: 1.20.0.2
`
`BEADS_SILVER_0026: 1.20
`
`RADKIT_GOLD_0013: 1.18.0.6
`
`BEADS_SILVER_0025: 1.18
`
`RADKIT_GOLD_0012: 1.18.0.4
`
`BEADS_SILVER_0024: 1.18
`
`RADKIT_GOLD_0011: 1.18.0.2
`
`BEADS_SILVER_0023: 1.18
`
`RADKIT_GOLD_0010_INTERNAL: 1.18
`
`BEADS_SILVER_0022: 1.18
`
`RADKIT_GOLD_0009: 1.16.0.10
`
`BEADS_SILVER_0021: 1.16
`
`BEADS_SILVER_0020: 1.16
`
`RADKIT_GOLD_0008_INTERNAL: 1.16
`
`BEADS_SILVER_0019: 1.16
`
`RADKIT_GOLD_0007: 1.16.0.8
`
`BEADS_SILVER_0018: 1.16
`
`BEADS_SILVER_0017: 1.16
`
`RADKIT_GOLD_0006: 1.16.0.6
`
`BEADS_SILVER_0016: 1.16
`
`RADKIT_GOLD_0005_INTERNAL: 1.16
`
`BEADS_SILVER_0015: 1.16
`
`RADKIT_GOLD_0004_INTERNAL: 1.16
`
`BEADS_SILVER_0014: 1.16
`
`RADKIT_GOLD_0003_INTERNAL: 1.16
`
`BEADS_SILVER_0013: 1.16
`
`RADKIT_GOLD_0002: 1.16.0.4
`
`BEADS_SILVER_0012: 1.16
`
`BEADS_SILVER_0011: 1.16
`
`RADKIT_GOLD_0001: 1.16.0.2
`
`BEADS_SILVER_0010: 1.16
`
`BEADS_SILVER_0009: 1.16
`
`BEADS_SILVER_0008: 1.12
`
`BEADS_SILVER_0007: 1.8
`
`SILVER: 1.26
`
`keyword substitution: kv
`total revisions: 26;
`selected revisions: 26
`description:
`----------------------------
`revision 1.26
`date: 2004-03-04 18:23:22 -0600; author: build; state: Exp; lines: +81 -83;
`Fix GCC type-punning warnings, which show up in retail mode on GCC 3.3.2.
`
`Page 3 of 7
`
`Implicit Exhibit 2016
`Sonos v. Implicit, IPR2018-0766, -0767
`
`

`

`Fix some coding convention violations (mostly whitespace).
`----------------------------
`revision 1.25
`date: 2002-06-10 16:45:14 -0500; author: davidc; state: Exp; lines: +5 -91;
`Removed the "preroll" edge, which is not used and whose purpose is unknown.
`----------------------------
`revision 1.24
`date: 2002-06-10 16:18:52 -0500; author: davidc; state: Exp; lines: +77 -53;
`Changed the default "smoothing" parameter to 1, so that the sliding average only includes the
`current value. Changed the code to silence all audio while the master clock is not set.
`Changed the code from silencing the first audio message to sending an empty message before
`sending the first message, which allows the render clock to be updated without any data loss.
`
`These changes help lessen the severity of audio chop that is heard on slave paths when the
`master is on a different machine than the source.
`
`The problem was that the delay for each audio chuck is the average of the last eight audio
`chunks played. This "smoothing" technique caused the master/render clock difference to
`oscillate while converging, which caused chopping. To explain: when audio is early, AudioSync
`delays the audio. When AudioSync has delayed the audio to make the audio on-time, the
`sliding average makes AudioSync act as if the audio is still early, causing it to further delay the
`audio. This makes the audio late. The audio did converge to being on-time, but not before
`sliding from early to on-time to late to on-time. Because AudioSync silences the audio when it
`is early or late and plays it when it is on-time, this oscillation causes audio to be silenced, then
`played, then silenced, then played, which is the chopping.
`
`The end result of these changes is that synchronization converge is sped up slightly and the out-
`of-sync slaves are silent instead of chopping. In the worst case, when the master is on a
`different machine than the source, this can cause between one and two seconds of silence.
`
`This is more of a work-around than a fix for the real problem, which is that is that it takes time
`for updates to the render clock of the master path to be propogated to the slaves. I don't know
`of a good way to solve this problem.
`----------------------------
`revision 1.23
`date: 2002-06-10 12:04:52 -0500; author: davidc; state: Exp; lines: +22 -7;
`Fixed a memory overwrite bug in SlidingAvg_Init. Memory would be overwritten if the sliding
`average length was less than five.
`----------------------------
`revision 1.22
`date: 2002-05-24 18:46:48 -0500; author: davidc; state: Exp; lines: +150 -148;
`Made it so that the "MasterClockOffset" attribute is treated as a signed value, so the audio can
`be sped-up in addition to being slowed down. Made it so that "MasterClockOffset" is ignored in
`master paths.
`
`Page 4 of 7
`
`Implicit Exhibit 2016
`Sonos v. Implicit, IPR2018-0766, -0767
`
`

`

`
`Refactored AudioSync_Adjust so that it can be called multiple times for a single message. This
`was useful when playing around with different strategies of how to incorporate the
`MasterClockOffset. I did not end up calling AudioSunc_Adjust from more than one place, but I
`kept the changes to make it easier to play around with the MasterClockOffset strategy in the
`future.
`----------------------------
`revision 1.21
`date: 2002-05-16 22:26:03 -0500; author: davidc; state: Exp; lines: +4 -4;
`Untabified
`----------------------------
`revision 1.20
`date: 2002-04-30 21:55:17 -0500; author: davidc; state: Exp; lines: +60 -42;
`Added some code that reads in the "MasterClockOffset" from the path and adds that the to
`master clock's epoch. This hack helps simpletv synchronize audio and video.
`----------------------------
`revision 1.19
`date: 2002-04-30 20:11:26 -0500; author: davidc; state: Exp; lines: +19 -16;
`Fixed defect #257, which is that audiosync would go into an inifite loop when it received an
`empty audio package that was within the EARLY_COPY range. I fixed this by not attempting to
`duplicate empty audio messages.
`----------------------------
`revision 1.18
`date: 2002-04-03 15:15:46 -0600; author: davidc; state: Exp; lines: +2 -2;
`Made the "Audio sync reports" DEBUGOUT a data trace (level 4) rather than a detail (level 5) so
`that it can be turned on without turning the others on.
`----------------------------
`revision 1.17
`date: 2002-04-02 23:23:52 -0600; author: davidc; state: Exp; lines: +3 -1;
`Added static keyword to eliminate unnecessary namespace pollution.
`----------------------------
`revision 1.16
`date: 2001-11-09 18:53:11 -0600; author: guyc; state: Exp; lines: +1 -26;
`Removed old and configuration constants.
`----------------------------
`revision 1.15
`date: 2001-11-09 16:52:01 -0600; author: guyc; state: Exp; lines: +11 -11;
`Changed capitalization of path parameters.
`----------------------------
`revision 1.14
`date: 2001-11-09 16:49:04 -0600; author: guyc; state: Exp; lines: +120 -20;
`Now optionally takes parameters (lots of them) from the
`path context. Default values have changed back to
`favour faster convergence, which may not work will with
`
`Page 5 of 7
`
`Implicit Exhibit 2016
`Sonos v. Implicit, IPR2018-0766, -0767
`
`

`

`the ipaq demo - need to use rule seeds in the ipaq demo
`to get it back to good.
`----------------------------
`revision 1.13
`date: 2001-11-09 14:36:49 -0600; author: guyc; state: Exp; lines: +23 -6;
`Changed audiosync to try to detect the situation where
`it is the master and not silence the initial packet.
`----------------------------
`revision 1.12
`date: 2001-10-23 11:53:51 -0500; author: guyc; state: Exp; lines: +2 -1;
`Fixed memory leak. Again.
`----------------------------
`revision 1.11
`date: 2001-10-22 18:34:28 -0500; author: guyc; state: Exp; lines: +12 -10;
`Changed a bunch of debug_stringprints to debugout_detail's, and
`removed explicit debug level setting.
`----------------------------
`revision 1.10
`date: 2001-10-18 15:59:04 -0500; author: guyc; state: Exp; lines: +165 -37;
`Added "preroll" edge to allow sync when a remote edge is
`possibly very late.
`----------------------------
`revision 1.9
`date: 2001-10-15 16:33:01 -0500; author: guyc; state: Exp; lines: +25 -1;
`Added code to silence initial unsync-ed packet.
`----------------------------
`revision 1.8
`date: 2001-10-11 10:18:27 -0500; author: guyc; state: Exp; lines: +114 -8;
`Various changes to improve audio quality and
`avoid major screwups when sample clock is unavailable.
`----------------------------
`revision 1.7
`date: 2001-10-03 19:01:25 -0500; author: guyc; state: Exp; lines: +54 -18;
`Minor cleanups.
`----------------------------
`revision 1.6
`date: 2001-10-03 18:11:55 -0500; author: guyc; state: Exp; lines: +47 -41;
`Now uses a hybrid of resampling, dropping or duplicating, depending
`on the severity of the error.
`----------------------------
`revision 1.5
`date: 2001-10-03 17:32:14 -0500; author: guyc; state: Exp; lines: +83 -2;
`Added resampling algorithm.
`----------------------------
`
`Page 6 of 7
`
`Implicit Exhibit 2016
`Sonos v. Implicit, IPR2018-0766, -0767
`
`

`

`revision 1.4
`date: 2001-10-03 16:58:43 -0500; author: guyc; state: Exp; lines: +27 -7;
`Changed padding method to use audio duplication instead of silence.
`----------------------------
`revision 1.3
`date: 2001-10-03 10:56:20 -0500; author: guyc; state: Exp; lines: +14 -5;
`Fixed sign error in sliding average error calculation.
`----------------------------
`revision 1.2
`date: 2001-10-02 19:34:28 -0500; author: guyc; state: Exp; lines: +85 -15;
`Added sliding average to latency calculation
`----------------------------
`revision 1.1
`date: 2001-09-28 18:04:40 -0500; author: guyc; state: Exp;
`Initial checkin of audiosync. Works using very simple silence/dropping
`logic. Requires finess to make it work with gradual time drift.
`=============================================================================
`
`
`Page 7 of 7
`
`Implicit Exhibit 2016
`Sonos v. Implicit, IPR2018-0766, -0767
`
`