Index: main/dsp.c
===================================================================
--- main/dsp.c	(revision 360932)
+++ main/dsp.c	(working copy)
@@ -209,9 +209,9 @@
 #define DTMF_GSIZE		102
 
 /* How many successive hits needed to consider begin of a digit */
-#define DTMF_HITS_TO_BEGIN	4
+#define DTMF_HITS_TO_BEGIN	2
 /* How many successive misses needed to consider end of a digit */
-#define DTMF_MISSES_TO_END	4
+#define DTMF_MISSES_TO_END	3
 
 /*!
  * \brief The default silence threshold we will use if an alternate
@@ -723,41 +723,42 @@
 			}
 		} 
 
-		if (hit == s->td.dtmf.lasthit) {
-			if (s->td.dtmf.current_hit) {
-				/* We are in the middle of a digit already */
-				if (hit) {
-					if (hit != s->td.dtmf.current_hit) {
-						/* Look for a start of a new digit.
-						   This is because hits_to_begin may be smaller than misses_to_end
-						   and we may find the beginning of new digit before we consider last one ended. */
-						s->td.dtmf.current_hit = 0;
-					} else {
-						/* Current hit was same as last, so increment digit duration (of last digit) */
-						s->digitlen[s->current_digits - 1] += DTMF_GSIZE;
-					}
-				} else {
-					/* No Digit */
-					s->td.dtmf.misses++;
-					if (s->td.dtmf.misses == s->td.dtmf.misses_to_end) {
-						/* There were enough misses to consider digit ended */
-						s->td.dtmf.current_hit = 0;
-					}
+		if (s->td.dtmf.current_hit) {
+			/* We are in the middle of a digit already */
+			if (hit != s->td.dtmf.current_hit) {
+				s->td.dtmf.misses++;
+				if (s->td.dtmf.misses == s->td.dtmf.misses_to_end) {
+					/* There were enough misses to consider digit ended */
+					s->td.dtmf.current_hit = 0;
 				}
-			} else if (hit) {
-				/* Detecting new digit */
+			} else {
+				s->td.dtmf.misses = 0;
+				/* Current hit was same as last, so increment digit duration (of last digit) */
+				s->digitlen[s->current_digits - 1] += DTMF_GSIZE;
+			}
+		}
+
+		/* Look for a start of a new digit no matter if we are already in the middle of some
+		   digit or not. This is because hits_to_begin may be smaller than misses_to_end
+		   and we may find begin of new digit before we consider last one ended. */
+		if (hit) {
+			if (hit == s->td.dtmf.lasthit) {
 				s->td.dtmf.hits++;
-				if (s->td.dtmf.hits == s->td.dtmf.hits_to_begin) {
-					store_digit(s, hit);
-					s->td.dtmf.current_hit = hit;
-				}
+			} else {
+				s->td.dtmf.hits = 1;
 			}
+
+			if (s->td.dtmf.hits == s->td.dtmf.hits_to_begin && hit != s->td.dtmf.current_hit) {
+				store_digit(s, hit);
+				s->td.dtmf.current_hit = hit;
+				s->td.dtmf.misses = 0;
+			}
 		} else {
-			s->td.dtmf.hits = 1;
-			s->td.dtmf.misses = 1;
-			s->td.dtmf.lasthit = hit;
+			s->td.dtmf.hits = 0;
 		}
 
+		s->td.dtmf.lasthit = hit;
+
 		/* If we had a hit in this block, include it into mute fragment */
 		if (squelch && hit) {
 			if (mute.end < sample - DTMF_GSIZE) {