diff --git a/beetsplug/echoplus.py b/beetsplug/echoplus.py
index 83505f06c..2386a68f3 100644
--- a/beetsplug/echoplus.py
+++ b/beetsplug/echoplus.py
@@ -52,7 +52,7 @@ def _picker(value, rang, mapping):
     return m # in case of floating point precision problems
 
 def _mapping(mapstr):
-    """Split mapstr at comma and return the stripped values as array."""
+    """Split mapstr at comma and returned the stripped values as array."""
     return [ m.strip() for m in mapstr.split(u',') ]
 
 def _guess_mood(valence, energy):
@@ -105,7 +105,7 @@ def fetch_item_attributes(lib, item, write, force, reapply):
 
     log.debug(u'echoplus: {} - {} [{}] force:{} reapply:{}'.format(
         item.artist, item.title, item.length, force, reapply))
-    # permanently store the raw values?  not supported yet
+    # permanently store the raw values?
     store_raw = config['echoplus']['store_raw'].get(bool)
 
     # if we want to set mood, we need to make sure, that valence and energy
@@ -156,8 +156,7 @@ def fetch_item_attributes(lib, item, write, force, reapply):
         reapply = True
 
         # (re-)fetch audio_summary and store it to the raw values.  if we do
-        # not want to keep the raw values, we clean them up later (not
-        # implemented yet)
+        # not want to keep the raw values, we clean them up later
 
         audio_summary = get_audio_summary(item.artist, item.title,
                 item.length, allow_upload, item.path)
@@ -287,20 +286,22 @@ def get_audio_summary(artist, title, duration, upload, path):
 
     if (not pick or min_distance > 1.0) and upload:
         log.debug(u'echoplus: uploading file "{}" to EchoNest'.format(path))
-        t = _echonest_fun(pyechonest.track.track_from_filename, filename=path)
-        if t:
-            log.debug(u'echoplus: track {} - {} [{:2.2f}]'.format(t.artist, t.title,
-                t.duration))
-            # FIXME:  maybe make pyechonest "nicer"?
-            result = {}
-            result['energy'] = t.energy
-            result['liveness'] = t.liveness
-            result['speechiness'] = t.speechiness
-            result['acousticness'] = t.acousticness
-            result['danceability'] = t.danceability
-            result['valence'] = t.valence
-            result['tempo'] = t.tempo
-            return result
+        # FIXME: same loop as above...  make this better
+        for i in range(RETRIES):
+            t = _echonest_fun(pyechonest.track.track_from_filename, filename=path)
+            if t:
+                log.debug(u'echoplus: track {} - {} [{:2.2f}]'.format(t.artist, t.title,
+                    t.duration))
+                # FIXME:  maybe make pyechonest "nicer"?
+                result = {}
+                result['energy'] = t.energy
+                result['liveness'] = t.liveness
+                result['speechiness'] = t.speechiness
+                result['acousticness'] = t.acousticness
+                result['danceability'] = t.danceability
+                result['valence'] = t.valence
+                result['tempo'] = t.tempo
+                return result
     elif not pick:
         return None
     return pick.audio_summary
diff --git a/docs/faq.rst b/docs/faq.rst
index f7d516bec..4408e0c50 100644
--- a/docs/faq.rst
+++ b/docs/faq.rst
@@ -202,9 +202,9 @@ Why does beets…
 There are a number of possibilities:
 
 -  First, make sure the album is in `the MusicBrainz
-   database <http://musicbrainz.org/>`__. You can search on their site to make
-   sure it's cataloged there. (If not, anyone can edit MusicBrainz---so
-   consider adding the data yourself.)
+   database <http://musicbrainz.org/>`__ the MusicBrainz database. You
+   can search on their site to make sure it's cataloged there. (If not,
+   anyone can edit MusicBrainz---so consider adding the data yourself.)
 -  If the album in question is a multi-disc release, see the relevant
    FAQ answer above.
 -  The music files' metadata might be insufficient. Try using the "enter