diff --git a/intern/pro_tools.py b/intern/pro_tools.py
index e7ddb35..d35f934 100644
--- a/intern/pro_tools.py
+++ b/intern/pro_tools.py
@@ -1,8 +1,13 @@
+import pprint
+
 import bpy
 import ptsl
 from ptsl import PTSL_pb2 as pt
 
 
+from .geom_utils import compute_relative_vector, room_norm_vector, speaker_active_time_range
+
+
 def mixdown_speaker(speaker: bpy.types.Object, scene: bpy.types.Scene) -> str:
     return ""
 
@@ -50,9 +55,54 @@ def insert_audio(speaker: bpy.types.Object, scene: bpy.types.Scene,
     
 
 def apply_pan_automation(speaker: bpy.types.Object, scene: bpy.types.Scene,
-                         track_id: str, client: ptsl.Client):
+                         track_id: str, room_size: float, client: ptsl.Client):
     
-    pass
+    CId_SetTrackControlBreakpoints = 150
+    # fps = scene.render.fps 
+    speaker_interval = speaker_active_time_range(speaker)
+    assert scene.camera, "Scene does not have a camera"
+    
+    values: list[tuple[dict, dict, dict]] = []
+
+    for frame in range(speaker_interval.start_frame, 
+                       speaker_interval.end_frame + 1):
+        scene.frame_set(frame)
+        relative_vector = compute_relative_vector(camera=scene.camera, 
+                                                  target=speaker)
+        norm_vec = room_norm_vector(relative_vector, 
+                                    room_size=room_size)
+        
+        time_struct = {'location': str(frame), 'time_type': 'TLType_Frames'} 
+
+        x_pos = {'time': time_struct, 'value': round(norm_vec.x, 4)}
+        y_pos = {'time': time_struct, 'value': round(norm_vec.y, 4)}
+        z_pos = {'time': time_struct, 'value': round(norm_vec.z, 4)}
+
+        values.append((x_pos, y_pos, z_pos))
+
+    for i, control in enumerate(['PCParameter_X', 'PCParameter_Y',
+                                 'PCParameter_Z']):
+
+        params ={
+            'track_id': track_id,
+            'control_id': {
+                'section': 'TSId_MainOut',
+                'control_type': 'TCType_Pan',
+                'pan': {
+                    'pan_space': 'PSpace_3dCart3layer', 
+                    'parameter': control,
+                    'channel': 'SChannel_Mono'
+                },
+            },
+            'breakpoints': [v[i] for v in values]
+
+        } 
+
+        print("Sending parameter changes:")
+        pprint.pprint(params)
+
+        client.run_command(CId_SetTrackControlBreakpoints, params)
+
 
 
 def send_to_pro_tools(speakers: list[bpy.types.Object], 
@@ -67,7 +117,8 @@ def send_to_pro_tools(speakers: list[bpy.types.Object],
         for speaker in speakers:
             new_track_id = create_track(speaker, client)
             insert_audio(speaker, scene, new_track_id, client)
-            apply_pan_automation(speaker, scene, new_track_id, client)
+            apply_pan_automation(speaker, scene, new_track_id, room_size, 
+                                 client)
 
         client.close()
         return True
diff --git a/operator_protools_export.py b/operator_protools_export.py
index 3f74688..bc47438 100644
--- a/operator_protools_export.py
+++ b/operator_protools_export.py
@@ -23,7 +23,8 @@ class SendToProTools(bpy.types.Operator):
     @classmethod
     def poll(cls, context):
         return len(filter_speakers(context.selected_objects)) > 0 and \
-                bpy.app.online_access
+                bpy.app.online_access and \
+                context.scene.camera is not None
 
     def invoke(self, context, event):
         wm = context.window_manager