diff --git a/docs/api-guide/renderers.md b/docs/api-guide/renderers.md
index 7fc1fc1fd..d46d05686 100644
--- a/docs/api-guide/renderers.md
+++ b/docs/api-guide/renderers.md
@@ -88,7 +88,7 @@ The client may additionally include an `'indent'` media type parameter, in which
 
 **.format**: `'.json'`
 
-**.charset**: `utf-8`
+**.charset**: `None`
 
 ## UnicodeJSONRenderer
 
@@ -110,7 +110,7 @@ Both the `JSONRenderer` and `UnicodeJSONRenderer` styles conform to [RFC 4627][r
 
 **.format**: `'.json'`
 
-**.charset**: `utf-8`
+**.charset**: `None`
 
 ## JSONPRenderer
 
@@ -295,12 +295,15 @@ By default renderer classes are assumed to be using the `UTF-8` encoding.  To us
 
 Note that if a renderer class returns a unicode string, then the response content will be coerced into a bytestring by the `Response` class, with the `charset` attribute set on the renderer used to determine the encoding.
 
-If the renderer returns a bytestring representing raw binary content, you should set a charset value of `None`, which will ensure the `Content-Type` header of the response will not have a `charset` value set.  Doing so will also ensure that the browsable API will not attempt to display the binary content as a string.
+If the renderer returns a bytestring representing raw binary content, you should set a charset value of `None`, which will ensure the `Content-Type` header of the response will not have a `charset` value set.
+
+In some cases you may also want to set the `render_style` attribute to `'binary'`.  Doing so will also ensure that the browsable API will not attempt to display the binary content as a string.
 
     class JPEGRenderer(renderers.BaseRenderer):
         media_type = 'image/jpeg'
         format = 'jpg'
         charset = None
+        render_style = 'binary'
 
         def render(self, data, media_type=None, renderer_context=None):
             return data
diff --git a/rest_framework/renderers.py b/rest_framework/renderers.py
index 9885c8ddc..c07b1652c 100644
--- a/rest_framework/renderers.py
+++ b/rest_framework/renderers.py
@@ -36,6 +36,7 @@ class BaseRenderer(object):
     media_type = None
     format = None
     charset = 'utf-8'
+    render_style = 'text'
 
     def render(self, data, accepted_media_type=None, renderer_context=None):
         raise NotImplemented('Renderer class requires .render() to be implemented')
@@ -51,16 +52,17 @@ class JSONRenderer(BaseRenderer):
     format = 'json'
     encoder_class = encoders.JSONEncoder
     ensure_ascii = True
-    charset = 'utf-8'
-    # Note that JSON encodings must be utf-8, utf-16 or utf-32.
+    charset = None
+    # JSON is a binary encoding, that can be encoded as utf-8, utf-16 or utf-32.
     # See: http://www.ietf.org/rfc/rfc4627.txt
+    # Also: http://lucumr.pocoo.org/2013/7/19/application-mimetypes-and-encodings/
 
     def render(self, data, accepted_media_type=None, renderer_context=None):
         """
         Render `data` into JSON.
         """
         if data is None:
-            return ''
+            return bytes()
 
         # If 'indent' is provided in the context, then pretty print the result.
         # E.g. If we're being called by the BrowsableAPIRenderer.
@@ -85,13 +87,12 @@ class JSONRenderer(BaseRenderer):
         # and may (or may not) be unicode.
         # On python 3.x json.dumps() returns unicode strings.
         if isinstance(ret, six.text_type):
-            return bytes(ret.encode(self.charset))
+            return bytes(ret.encode('utf-8'))
         return ret
 
 
 class UnicodeJSONRenderer(JSONRenderer):
     ensure_ascii = False
-    charset = 'utf-8'
     """
     Renderer which serializes to JSON.
     Does *not* apply JSON's character escaping for non-ascii characters.
@@ -108,6 +109,7 @@ class JSONPRenderer(JSONRenderer):
     format = 'jsonp'
     callback_parameter = 'callback'
     default_callback = 'callback'
+    charset = 'utf-8'
 
     def get_callback(self, renderer_context):
         """
@@ -414,7 +416,10 @@ class BrowsableAPIRenderer(BaseRenderer):
         renderer_context['indent'] = 4
         content = renderer.render(data, accepted_media_type, renderer_context)
 
-        if renderer.charset is None:
+        render_style = getattr(renderer, 'render_style', 'text')
+        assert render_style in ['text', 'binary'], 'Expected .render_style ' \
+            '"text" or "binary", but got "%s"' % render_style
+        if render_style == 'binary':
             return '[%d bytes of binary content]' % len(content)
 
         return content
diff --git a/rest_framework/routers.py b/rest_framework/routers.py
index 930011d39..3fee1e494 100644
--- a/rest_framework/routers.py
+++ b/rest_framework/routers.py
@@ -189,7 +189,11 @@ class SimpleRouter(BaseRouter):
         Given a viewset, return the portion of URL regex that is used
         to match against a single instance.
         """
-        base_regex = '(?P<{lookup_field}>[^/]+)'
+        if self.trailing_slash:
+            base_regex = '(?P<{lookup_field}>[^/]+)'
+        else:
+            # Don't consume `.json` style suffixes
+            base_regex = '(?P<{lookup_field}>[^/.]+)'
         lookup_field = getattr(viewset, 'lookup_field', 'pk')
         return base_regex.format(lookup_field=lookup_field)
 
diff --git a/rest_framework/tests/test_routers.py b/rest_framework/tests/test_routers.py
index 5fcccb741..e723f7d45 100644
--- a/rest_framework/tests/test_routers.py
+++ b/rest_framework/tests/test_routers.py
@@ -146,7 +146,7 @@ class TestTrailingSlashRemoved(TestCase):
         self.urls = self.router.urls
 
     def test_urls_can_have_trailing_slash_removed(self):
-        expected = ['^notes$', '^notes/(?P<pk>[^/]+)$']
+        expected = ['^notes$', '^notes/(?P<pk>[^/.]+)$']
         for idx in range(len(expected)):
             self.assertEqual(expected[idx], self.urls[idx].regex.pattern)