--- /dev/null
+Random assorted programs to sync multiple audio streams. Very useful when you
+have two sound cards each capable of recording eight channels each, but they're
+not in sync, since real-life oscillators are non-ideal. (I'm sure this is a
+really common use case.) Using a sync channel (yes, that means you're "wasting"
+one channel on each card) it can "stretch" the channels using a phase-locked
+loop (PLL) so they are reasonably in sync all the time. (It won't neccessarily
+help you getting the _start_ right, but once you've lined the recordings up
+they will _stay_ lined up and not drift apart.)
+
+VERY UNPOLISHED. Needed them once, probably will never again.
+
+Basically, the tools are, in semi-logical order:
+
+ - gen-random-skew: Generate semi-random skew; used for testing only.
+
+ - gen-sine SAMPLES FREQUENCY: Generate a FREQUENCY Hz sine running
+ for SAMPLES samples. Not generally used anymore.
+
+ - gen-square SAMPLES FREQUENCY: The same, only with a square wave.
+ This will be your sync channel; output it from one of the sound
+ cards (if you can) or from an external machine. You'll need the
+ sync signal on one channel of each card. Don't worry about clipping,
+ it's all just max/min anyhow.
+
+ - estimate-skew REFSYNC INPSYNC SKEW: Compare the sync channel from
+ REFSYNC with that from INPSYNC and output the estimated skew (or
+ rather, recommended normalized resampling frequency) to SKEW.
+
+ - do-skew INPUT SKEW OUTPUT: Take sound data from INPUT, resample it
+ using the information from SKEW (usually from estimate-skew) and
+ write it to OUTPUT.
+
+Copyright 2006 Steinar H. Gunderson <sgunderson@bigfoot.com>. Licensed under
+the GNU General Public License, version 2.
+
projects / audiosync / blobdiff