diff --git a/youtube_dlc/YoutubeDL.py b/youtube_dlc/YoutubeDL.py
index 8906607457..92ca687a87 100644
--- a/youtube_dlc/YoutubeDL.py
+++ b/youtube_dlc/YoutubeDL.py
@@ -2349,7 +2349,7 @@ class YoutubeDL(object):
                             downloaded.append(fname)
                             partial_success, real_download = dl(fname, new_info)
                             success = success and partial_success
-                        info_dict['__postprocessors'] = postprocessors
+                        info_dict['__postprocessors'].append(postprocessors)
                         info_dict['__files_to_merge'] = downloaded
                         # Even if there were no downloads, it is being merged only now
                         info_dict['__real_download'] = True
diff --git a/youtube_dlc/postprocessor/sponskrub.py b/youtube_dlc/postprocessor/sponskrub.py
index 4e9bec2571..0ba22138e8 100644
--- a/youtube_dlc/postprocessor/sponskrub.py
+++ b/youtube_dlc/postprocessor/sponskrub.py
@@ -43,6 +43,10 @@ class SponSkrubPP(PostProcessor):
         if self.path is None:
             return [], information
 
+        filename = information['filepath']
+        if not os.path.exists(encodeFilename(filename)):  # no download
+            return [], information
+
         if information['extractor_key'].lower() != 'youtube':
             self.to_screen('Skipping sponskrub since it is not a YouTube video')
             return [], information
@@ -58,7 +62,6 @@ class SponSkrubPP(PostProcessor):
             if not information.get('__real_download', False):
                 self.report_warning('If sponskrub is run multiple times, unintended parts of the video could be cut out.')
 
-        filename = information['filepath']
         temp_filename = prepend_extension(filename, self._temp_ext)
         if os.path.exists(encodeFilename(temp_filename)):
             os.remove(encodeFilename(temp_filename))