[FFmpeg-soc] [soc]: r1285 - in qcelp: qcelpdata.h qcelpdec.c
reynaldo
subversion at mplayerhq.hu
Mon Sep 3 00:25:21 CEST 2007
Author: reynaldo
Date: Mon Sep 3 00:25:21 2007
New Revision: 1285
Log:
Removes pointless macro stating spec has errors -- Another from Michael's review
Modified:
qcelp/qcelpdata.h
qcelp/qcelpdec.c
Modified: qcelp/qcelpdata.h
==============================================================================
--- qcelp/qcelpdata.h (original)
+++ qcelp/qcelpdata.h Mon Sep 3 00:25:21 2007
@@ -26,16 +26,6 @@
*/
/**
- * TIA/EIA/IS-733 Spec has errors on the predictor determination formula
- * at equation 2.4.6.1-4 -- The predictor there needs 6 to be subtracted
- * from it to give RI compliants results. The problem is it ignores the
- * fact that codebook subframes 4,8,12 and 16 on a FULL_RATE frame use a
- * different quantizer table.
- */
-
-#define QCELP_FIX_SPEC_PREDICTOR(p) (p-6)
-
-/**
* TIA/EIA/IS-733 Spec has an omission on the codebook index determination
* formula for RATE_FULL and RATE_HALF frames at section 2.4.8.1.1. It says
* you have to subtract the decoded index parameter to the given scaled
Modified: qcelp/qcelpdec.c
==============================================================================
--- qcelp/qcelpdec.c (original)
+++ qcelp/qcelpdec.c Mon Sep 3 00:25:21 2007
@@ -160,14 +160,17 @@ void qcelp_decode_params(AVCodecContext
g0[i]=4*cbgain[i];
/*
- * Spec has errors on the predictor determination formula
- * it needs 6 to be subtracted from it to give RI results.
+ * TIA/EIA/IS-733 Spec has errors on the predictor determination
+ * formula at equation 2.4.6.1-4 -- The predictor there needs 6
+ * to be subtracted from it to give RI compliants results. The
+ * problem is it ignores the fact that codebook subframes 4, 8,
+ * 12 and 16 on a FULL_RATE frame use a different quantizer
+ * table.
*/
if(frame->rate == RATE_FULL && i > 0 && !((i+1) & 3))
- predictor=QCELP_FIX_SPEC_PREDICTOR
- (av_clip(floor((g1[i-1]+g1[i-2]+g1[i-3])/3.0), 6,
- 38));
+ predictor=av_clip(floor((g1[i-1]+g1[i-2]+g1[i-3])/3.0), 6,
+ 38)-6;
else
predictor=0;
More information about the FFmpeg-soc
mailing list