From aadff3e1efd9ec8d12c41c69a136fa4ff479d573 Mon Sep 17 00:00:00 2001
From: Jules Laplace <julescarbon@gmail.com>
Date: Fri, 25 May 2018 00:17:46 +0200
Subject: modify dataset to jitter differently

---
 datasets/dataset.pl | 42 ++++++++++++++++++++++++++++++++++++------
 1 file changed, 36 insertions(+), 6 deletions(-)

(limited to 'datasets')

diff --git a/datasets/dataset.pl b/datasets/dataset.pl
index cf1c1e2..a89e3f8 100755
--- a/datasets/dataset.pl
+++ b/datasets/dataset.pl
@@ -4,7 +4,9 @@ use strict;
 use Getopt::Std;
 
 our $opt_c;
-getopts('c');
+our $opt_j;
+getopts('c'); # normalize the file
+getopts('j'); # jitter the pitch of each subsample by N rather than repitching big chunks
 print "<<<< DATASET >>>>\n";
 
 my $fmt = <<FMT;
@@ -45,11 +47,8 @@ sub process($) {
     }
   }
 
-  my $scale = sprintf("%.09f", 5e-7 * $seconds);
 
   print "Seconds: $seconds\n";
-  print "Scale factor: $scale\n";
-  print "\n";
 
   my $a_tmp = "a_" . $filename;
   my $b_tmp = "b_" . $filename;
@@ -59,8 +58,39 @@ sub process($) {
   } else {
     $a_tmp = $filename;
   }
-  system("./spread.sh", $a_tmp, $b_tmp, 0.999, $scale, 1.001);
-  system("./split44k.sh", $b_tmp, 8, $name);
+
+  # jitter mide:  make copies of the original file and join together,
+  #   split into chunks, and then pitchshift the chunks.
+  if ($opt_j) {
+    my $copies = 4080 / $seconds;
+    print "Copies to make: $copies\n";
+    print "\n";
+    for (my $i = 0; $i < $copies; $i++) {
+     system('cp', $a_tmp, sprintf("tmp_%d.wav", $i));
+    }
+    system 'sox', glob 'tmp_*', $b_tmp;
+    system 'rm', glob 'tmp_*';
+    system './split44k.sh', $b_tmp, 8, $name;
+    opendir my $dh, $name or die "cant opendir $name: $!";
+    while (readdir $dh) {
+      next if /^\./;
+      my $fn = $_;
+      my $n = rand( 2 * $j ) - $j;
+      system 'sox', $name . '/' . $fn, $name . '/z_' . $fn, 'pitch', $n;
+      system 'rm', $name . '/' . $fn;
+    }
+    closedir $dh;
+  }
+
+  # standard mode: make a bunch of copies and then join them together,
+  #   then subdivide into 8s chunks
+  else {
+    my $scale = sprintf("%.09f", 5e-7 * $seconds);
+    print "Scale factor: $scale\n";
+    print "\n";
+    system("./spread.sh", $a_tmp, $b_tmp, 0.999, $scale, 1.001);
+    system("./split44k.sh", $b_tmp, 8, $name);
+  }
   if (!$opt_c) {
     system("/bin/rm", $a_tmp);
   }
-- 
cgit v1.2.3-70-g09d2