diff --git a/youtube_dl/extractor/libsyn.py b/youtube_dl/extractor/libsyn.py
index 4750b03a3..07bd18d64 100644
--- a/youtube_dl/extractor/libsyn.py
+++ b/youtube_dl/extractor/libsyn.py
@@ -11,14 +11,14 @@ class LibsynIE(InfoExtractor):
     _VALID_URL = r'(?P<mainurl>https?://html5-player\.libsyn\.com/embed/episode/id/(?P<id>[0-9]+))'
 
     _TESTS = [{
-        'url': 'http://html5-player.libsyn.com/embed/episode/id/3377616/',
-        'md5': '443360ee1b58007bc3dcf09b41d093bb',
+        'url': 'http://html5-player.libsyn.com/embed/episode/id/6324076/',
+        'md5': '5b497505660690028d482f9a28431cca',
         'info_dict': {
-            'id': '3377616',
+            'id': '6324076',
             'ext': 'mp3',
-            'title': "The Daily Show Podcast without Jon Stewart - Episode 12: Bassem Youssef: Egypt's Jon Stewart",
-            'description': 'md5:601cb790edd05908957dae8aaa866465',
-            'upload_date': '20150220',
+            'title': "Verily, Octopi Sing - Pandas are not smallpox",
+            'description': 'md5:0d20ad472ae296f22a0c9df23a6d78f1',
+            'upload_date': '20180303',
             'thumbnail': 're:^https?://.*',
         },
     }, {
@@ -28,6 +28,7 @@ class LibsynIE(InfoExtractor):
             'id': '3727166',
             'ext': 'mp3',
             'title': 'Clients From Hell Podcast - How a Sex Toy Company Kickstarted my Freelance Career',
+            'description': 'md5:996a28b4f829ed4ffb8302a53d825704',
             'upload_date': '20150818',
             'thumbnail': 're:^https?://.*',
         }
@@ -39,25 +40,30 @@ class LibsynIE(InfoExtractor):
         url = m.group('mainurl')
         webpage = self._download_webpage(url, video_id)
 
-        formats = [{
-            'url': media_url,
-        } for media_url in set(re.findall(r'var\s+mediaURL(?:Libsyn)?\s*=\s*"([^"]+)"', webpage))]
-
-        podcast_title = self._search_regex(
-            r'<h2>([^<]+)</h2>', webpage, 'podcast title', default=None)
-        episode_title = self._search_regex(
-            r'(?:<div class="episode-title">|<h3>)([^<]+)</', webpage, 'episode title')
+        media_url_json = self._search_regex(r'"media_url(?:_libsyn)?"\s*:\s*("(\\"|[^"])+")', webpage, 'media_url')
+        media_url = self._parse_json(media_url_json, video_id)
 
+        podcast_title = self._html_search_regex( r'<h3\b[^>]*>([^<]+)</h3>', webpage, 'podcast title', default=None)
+        episode_title_json = self._search_regex(r'"item_title"\s*:\s*("(\\"|[^"])+")', webpage, 'episode title', default=None)
+        episode_title = self._parse_json(episode_title_json, video_id, fatal=False)
+        if episode_title is None:
+            # Fallback: scrape from page
+            episode_title = self._html_search_regex(r'<h4\b[^>]*>([^<]+)', webpage, 'episode title')
         title = '%s - %s' % (podcast_title, episode_title) if podcast_title else episode_title
 
         description = self._html_search_regex(
-            r'<div id="info_text_body">(.+?)</div>', webpage,
-            'description', default=None)
-        thumbnail = self._search_regex(
-            r'<img[^>]+class="info-show-icon"[^>]+src="([^"]+)"',
-            webpage, 'thumbnail', fatal=False)
-        release_date = unified_strdate(self._search_regex(
-            r'<div class="release_date">Released: ([^<]+)<', webpage, 'release date', fatal=False))
+            r'<(\w+)\b[^>]*id="info_text_body"[^>]*>(?P<description>.+?)<\/\1>',
+            webpage, 'description', default=None, group='description')
+
+        thumbnail_json = self._search_regex(r'"thumbnail_url"\s*:\s*("(\\"|[^"])+")', webpage, 'thumbnail')
+        thumbnail = self._parse_json(thumbnail_json, video_id, fatal=False)
+
+        release_date_json = self._search_regex(r'"release_date"\s*:\s*("(\\"|[^"])+")', webpage, 'release date')
+        release_date = unified_strdate(self._parse_json(release_date_json, video_id, fatal=False))
+        if release_date is None:
+            # Fallback: scrape from page
+            release_date = unified_strdate(self._search_regex(
+                r'<div class="release_date">Released: ([^<]+)<', webpage, 'release date', fatal=False))
 
         return {
             'id': video_id,
@@ -65,5 +71,5 @@ class LibsynIE(InfoExtractor):
             'description': description,
             'thumbnail': thumbnail,
             'upload_date': release_date,
-            'formats': formats,
+            'url': media_url,
         }