diff --git a/cps/book_formats.py b/cps/book_formats.py
index f7622ced..5acbc4c4 100644
--- a/cps/book_formats.py
+++ b/cps/book_formats.py
@@ -41,16 +41,21 @@ except ImportError as e:
 
 
 def process(tmp_file_path, original_file_name, original_file_extension):
+    meta = None
     try:
         if ".PDF" == original_file_extension.upper():
-            return pdf_meta(tmp_file_path, original_file_name, original_file_extension)
+            meta = pdf_meta(tmp_file_path, original_file_name, original_file_extension)
         if ".EPUB" == original_file_extension.upper() and use_epub_meta is True:
-            return epub.get_epub_info(tmp_file_path, original_file_name, original_file_extension)
+            meta = epub.get_epub_info(tmp_file_path, original_file_name, original_file_extension)
         if ".FB2" == original_file_extension.upper() and use_fb2_meta is True:
-            return fb2.get_fb2_info(tmp_file_path, original_file_extension)
+            meta = fb2.get_fb2_info(tmp_file_path, original_file_extension)
     except Exception as e:
         logger.warning('cannot parse metadata, using default: %s', e)
-    return default_meta(tmp_file_path, original_file_name, original_file_extension)
+
+    if meta and meta.title.strip() and meta.author.strip():
+        return meta
+    else:
+        return default_meta(tmp_file_path, original_file_name, original_file_extension)
 
 
 def default_meta(tmp_file_path, original_file_name, original_file_extension):
@@ -76,8 +81,8 @@ def pdf_meta(tmp_file_path, original_file_name, original_file_extension):
         doc_info = None
 
     if doc_info is not None:
-        author = doc_info.author if doc_info.author is not None else u"Unknown"
-        title = doc_info.title if doc_info.title is not None else original_file_name
+        author = doc_info.author if doc_info.author else u"Unknown"
+        title = doc_info.title if doc_info.title else original_file_name
         subject = doc_info.subject
     else:
         author = u"Unknown"
@@ -115,4 +120,4 @@ def get_versions():
         PVersion=PyPdfVersion
     else:
         PVersion=_(u'not installed')
-    return {'ImageVersion':IVersion,'PyPdfVersion':PVersion}
\ No newline at end of file
+    return {'ImageVersion':IVersion,'PyPdfVersion':PVersion}
diff --git a/cps/epub.py b/cps/epub.py
index 6473d401..50714b6f 100644
--- a/cps/epub.py
+++ b/cps/epub.py
@@ -89,7 +89,7 @@ def get_epub_info(tmp_file_path, original_file_name, original_file_extension):
                 else:
                     coverfile = extractCover(epubZip, coversection[0], coverpath, tmp_file_path)
 
-    if epub_metadata['title'] is None:
+    if not epub_metadata['title']:
         title = original_file_name
     else:
         title = epub_metadata['title']
diff --git a/cps/helper.py b/cps/helper.py
index 4c52def8..562d09af 100755
--- a/cps/helper.py
+++ b/cps/helper.py
@@ -261,6 +261,9 @@ def get_valid_filename(value, replace_whitespace=True):
         value = re.sub('[\*\+:\\\"/<>\?]+', u'_', value, flags=re.U)
 
     value = value[:128]
+    if not value:
+        raise ValueError("Filename cannot be empty")
+
     return value
 
 def get_sorted_author(value):