diff --git a/speech/gen_dataVals_from_wave_viewer.m b/speech/gen_dataVals_from_wave_viewer.m
index ccee774..724d8d0 100644
--- a/speech/gen_dataVals_from_wave_viewer.m
+++ b/speech/gen_dataVals_from_wave_viewer.m
@@ -302,10 +302,15 @@
     onset_type = 'default amplitude onset';
 end
 
+% find boundaries for onset detection. Only consider samples that could 
+%  have formants, ie, that exist in ftrack_taxis
+startIx = find(sigmat.ampl_taxis > sigmat.ftrack_taxis(1), 1);
+endIx = find(sigmat.ampl_taxis > sigmat.ftrack_taxis(end), 1) - 1; % subtracting 1 so endIx is inside the ftrack_taxis range
+
 % find onset
-onsetIndAmp = find(sigmat.ampl > ampl_thresh4voicing);
+onsetIndAmp = find(sigmat.ampl(startIx:endIx) > ampl_thresh4voicing);
 if onsetIndAmp
-    onsetIndAmp = onsetIndAmp(1) + 1;
+    onsetIndAmp = onsetIndAmp(1) + startIx;
 else
     onsetIndAmp = 1;
     onset_type = 'no amplitude onset found';
@@ -341,7 +346,7 @@
 
 % find the index of ftrack_taxis that's closest to and greater than ampl_taxis's onset index
 [~, onsetIndFtrack] = find(sigmat.ftrack_taxis - sigmat.ampl_taxis(onsetIndAmp)>0, 1); 
-if ~isempty(onsetIndFrack)
+if ~isempty(onsetIndFtrack)
     offsetIndFtrack = find(ampl_Ftrack(onsetIndFtrack:end) < ampl_thresh4voicing);
 else
     offsetIndFtrack = [];