# HG changeset patch
# User Chris Cannam <c.cannam@qmul.ac.uk>
# Date 1385637595 0
# Node ID fe450824850e91dd167af5c8f2ddbd075627321c
# Parent  e64ea86fe781cfe0440581eae16c8894c9ad65cd
Further latency adjustment

diff -r e64ea86fe781 -r fe450824850e cpp-qm-dsp/ConstantQ.cpp
--- a/cpp-qm-dsp/ConstantQ.cpp	Thu Nov 28 11:00:33 2013 +0000
+++ b/cpp-qm-dsp/ConstantQ.cpp	Thu Nov 28 11:19:55 2013 +0000
@@ -142,12 +142,21 @@
 
     cerr << "maxLatPlusDrop = " << maxLatPlusDrop << endl;
 
-    int maxDrop = *std::max_element(drops.begin(), drops.end());
-    cerr << "maxDrop " << maxDrop << endl;
+///    int maxDrop = *std::max_element(drops.begin(), drops.end());
+//    cerr << "maxDrop " << maxDrop << endl;
     
 //    m_totalLatency = ceil(double(maxLatPlusDrop) / m_bigBlockSize) * m_bigBlockSize;
     int bigHop = m_p.fftHop * pow(2, m_octaves) / 2;
-    m_totalLatency = ceil(double(maxLatPlusDrop) / bigHop) * bigHop;
+
+    // we want to design m_totalLatency such that m_totalLatency -
+    // latencies[0] - drops[0] is a multiple of m_p.fftHop, so that we
+    // can get identical results in octave 0 to our reference
+    // implementation, making for easier testing (though other octaves
+    // will differ because of different resampler implementations)
+
+    m_totalLatency = maxLatPlusDrop;
+    int lat0 = m_totalLatency - latencies[0] - drops[0];
+    m_totalLatency = ceil(double(lat0 / m_p.fftHop) * m_p.fftHop) + latencies[0] + drops[0];
 
 //    m_totalLatency = MathUtilities::nextPowerOfTwo(maxLatency);
     cerr << "total latency = " << m_totalLatency << endl;