changeset 146:d27f13651e67

* don't store fft values scaled by fftsize/2; that's a special requirement for the spectrogram, and other applications will not expect it -- make the spectrogram do that scaling itself * add a higher-resolution memory cache (still polar, though) as an alternative to the 16-bit compact cache * don't use the memory cache if we want rectangular coords (unless the disc cache is totally infeasible) as conversion slows it down anyway * avoid redundant rectangular -> polar -> rectangular conversion when storing values in a rectangular-mode disc cache
author Chris Cannam
date Fri, 01 Jun 2007 13:56:35 +0000
parents dfaf4f9330b5
children 140ea2569bcc
files transform/FeatureExtractionPluginTransform.cpp
diffstat 1 files changed, 2 insertions(+), 2 deletions(-) [+]
line wrap: on
line diff
--- a/transform/FeatureExtractionPluginTransform.cpp	Fri May 18 16:47:38 2007 +0000
+++ b/transform/FeatureExtractionPluginTransform.cpp	Fri Jun 01 13:56:35 2007 +0000
@@ -303,8 +303,8 @@
             if (blockFrame >= endFrame) break;
         }
 
-//	std::cerr << "FeatureExtractionPluginTransform::run: blockFrame "
-//		  << blockFrame << std::endl;
+	std::cerr << "FeatureExtractionPluginTransform::run: blockFrame "
+		  << blockFrame << std::endl;
 
 	long completion =
 	    (((blockFrame - startFrame) / m_context.stepSize) * 99) /