diff --git a/pycaption/scc/__init__.py b/pycaption/scc/__init__.py
index 5e4f5d02..dbfc51d4 100644
--- a/pycaption/scc/__init__.py
+++ b/pycaption/scc/__init__.py
@@ -164,7 +164,7 @@ def __init__(self, *args, **kw):
         )
 
         self.last_command = ''
-        self.cue = []
+        self.double_starter = None
 
         self.buffer_dict = NotifyingDict()
 
@@ -309,29 +309,18 @@ def _translate_line(self, line):
         parts = r.findall(line.lower())
 
         self.time_translator.start_at(parts[0][0])
-        cue_starters = ['9425', '9426', '94a7', '9429', '9420']
         word_list = parts[0][2].split(' ')
+
         for idx, word in enumerate(word_list):
-            # ignore empty results or invalid commands
             word = word.strip()
-            # if line_starts_a_cue:
-            if word in cue_starters:
-                if len(word_list) > 0 and word_list[idx] == word_list[idx-1]:
-                    self.cue.append(word)
-                else:
-                    self.cue = [word]
-            else:
-                self.cue.append(word)
-
-            previous_is_pac_or_tab = len(self.cue) > 1 and (
+            previous_is_pac_or_tab = len(word_list) > 1 and (
                     _is_pac_command(word_list[idx - 1]) or word_list[idx - 1] in PAC_TAB_OFFSET_COMMANDS
             )
-            pacs_are_doubled = len(self.cue) > 1 and self.cue[0] == self.cue[1]
             if len(word) == 4:
                 self._translate_word(
                     word=word,
                     previous_is_pac_or_tab=previous_is_pac_or_tab,
-                    pacs_are_doubled=pacs_are_doubled
+                    pacs_are_doubled=self.double_starter
                 )
 
     def _translate_word(self, word, previous_is_pac_or_tab, pacs_are_doubled):
@@ -367,12 +356,17 @@ def _handle_double_command(self, word, pacs_are_doubled):
         # If we have doubled commands we're skipping also
         # doubled special characters and doubled extended characters
         # with only one member of each pair being displayed.
-
+        cue_starter_commands = ['9425', '9426', '94a7', '9429', '9420']
         doubled_types = word != "94a1" and word in COMMANDS or _is_pac_command(word)
         if pacs_are_doubled:
-            doubled_types = doubled_types or word in SPECIAL_CHARS or word in EXTENDED_CHARS or word == "94a1"
+            doubled_types = doubled_types or word in EXTENDED_CHARS or word == "94a1"
+
+        if word in cue_starter_commands and word != self.last_command:
+            self.double_starter = False
 
         if doubled_types and word == self.last_command:
+            if word in cue_starter_commands:
+                self.double_starter = True
             self.last_command = ''
             return True
             # Fix for the <position> <tab offset> <position> <tab offset>
diff --git a/pycaption/scc/specialized_collections.py b/pycaption/scc/specialized_collections.py
index 920d0d7c..deaf56b9 100644
--- a/pycaption/scc/specialized_collections.py
+++ b/pycaption/scc/specialized_collections.py
@@ -1,5 +1,4 @@
 import collections
-import unicodedata
 
 from ..base import CaptionList, Caption, CaptionNode
 from ..geometry import (
@@ -9,7 +8,7 @@
 from .constants import (
     PAC_BYTES_TO_POSITIONING_MAP, COMMANDS, PAC_TAB_OFFSET_COMMANDS,
     MICROSECONDS_PER_CODEWORD, BACKGROUND_COLOR_CODES,
-    MID_ROW_CODES, EXTENDED_CHARS, SPECIAL_CHARS
+    MID_ROW_CODES, EXTENDED_CHARS
 )
 
 PopOnCue = collections.namedtuple("PopOnCue", "buffer, start, end")
diff --git a/tests/test_scc.py b/tests/test_scc.py
index ece0e1ce..13a1da8c 100644
--- a/tests/test_scc.py
+++ b/tests/test_scc.py
@@ -223,11 +223,10 @@ def test_skip_duplicate_tab_offset(self, sample_scc_duplicate_tab_offset):
     def test_skip_duplicate_special_characters(
             self, sample_scc_duplicate_special_characters):
         expected_lines = [
-            '®°½¿™¢£♪à èâêîôû',
+            '®®°°½½¿¿™™¢¢££♪♪àà  èèââêêîîôôûû',
             '®°½¿™¢£♪à èâêîôû',
             '®°AA½¿™¢£♪à èâêAAîôû'
         ]
-
         caption_set = SCCReader().read(sample_scc_duplicate_special_characters)
         actual_lines = [
             node.content
@@ -278,7 +277,7 @@ def test_freeze_rollup_captions_contents(self, sample_scc_roll_up_ru2):
             'HELPING THE LOCAL NEIGHBORHOODS',
             'AND IMPROVING THE LIVES OF ALL',
             'WE SERVE.',
-            '®°½',
+            '®°½½',
             'ABû',
             'ÁÉÓ¡',
             "WHERE YOU'RE STANDING NOW,",