diff --git a/gst-libs/gst/video/gstvideofilter.c b/gst-libs/gst/video/gstvideofilter.c
index 16b0cb25f0..c883dc999c 100644
--- a/gst-libs/gst/video/gstvideofilter.c
+++ b/gst-libs/gst/video/gstvideofilter.c
@@ -40,37 +40,98 @@
 #include "gstvideofilter.h"
 
 #include <gst/video/video.h>
+#include <gst/video/gstvideometa.h>
+#include <gst/video/gstvideopool.h>
 
 GST_DEBUG_CATEGORY_STATIC (gst_video_filter_debug);
 #define GST_CAT_DEFAULT gst_video_filter_debug
 
-static void gst_video_filter_class_init (gpointer g_class, gpointer class_data);
-static void gst_video_filter_init (GTypeInstance * instance, gpointer g_class);
+#define gst_video_filter_parent_class parent_class
+G_DEFINE_ABSTRACT_TYPE (GstVideoFilter, gst_video_filter,
+    GST_TYPE_BASE_TRANSFORM);
 
-static GstBaseTransformClass *parent_class = NULL;
-
-GType
-gst_video_filter_get_type (void)
+/* Answer the allocation query downstream. This is only called for
+ * non-passthrough cases */
+static gboolean
+gst_video_filter_propose_allocation (GstBaseTransform * trans, GstQuery * query)
 {
-  static GType video_filter_type = 0;
+  GstVideoFilter *filter = GST_VIDEO_FILTER_CAST (trans);
+  GstBufferPool *pool;
+  GstCaps *caps;
+  gboolean need_pool;
+  guint size;
 
-  if (!video_filter_type) {
-    static const GTypeInfo video_filter_info = {
-      sizeof (GstVideoFilterClass),
-      NULL,
-      NULL,
-      gst_video_filter_class_init,
-      NULL,
-      NULL,
-      sizeof (GstVideoFilter),
-      0,
-      gst_video_filter_init,
-    };
+  gst_query_parse_allocation (query, &caps, &need_pool);
 
-    video_filter_type = g_type_register_static (GST_TYPE_BASE_TRANSFORM,
-        "GstVideoFilter", &video_filter_info, G_TYPE_FLAG_ABSTRACT);
+  size = GST_VIDEO_INFO_SIZE (&filter->in_info);
+
+  if (need_pool) {
+    GstStructure *structure;
+
+    pool = gst_video_buffer_pool_new ();
+
+    structure = gst_buffer_pool_get_config (pool);
+    gst_buffer_pool_config_set (structure, caps, size, 0, 0, 0, 15);
+    if (!gst_buffer_pool_set_config (pool, structure))
+      goto config_failed;
+  } else
+    pool = NULL;
+
+  gst_query_set_allocation_params (query, size, 0, 0, 0, 15, pool);
+  gst_object_unref (pool);
+
+  gst_query_add_allocation_meta (query, GST_VIDEO_META_API);
+
+  return TRUE;
+
+  /* ERRORS */
+config_failed:
+  {
+    GST_ERROR_OBJECT (filter, "failed to set config");
+    gst_object_unref (pool);
+    return FALSE;
   }
-  return video_filter_type;
+}
+
+/* configure the allocation query that was answered downstream, we can configure
+ * some properties on it. Only called in passthrough mode. */
+static gboolean
+gst_video_filter_decide_allocation (GstBaseTransform * trans, GstQuery * query)
+{
+  GstBufferPool *pool = NULL;
+  guint size, min, max, prefix, alignment;
+
+  gst_query_parse_allocation_params (query, &size, &min, &max, &prefix,
+      &alignment, &pool);
+
+  if (pool) {
+    GstStructure *config;
+
+    config = gst_buffer_pool_get_config (pool);
+    gst_buffer_pool_config_add_option (config,
+        GST_BUFFER_POOL_OPTION_VIDEO_META);
+    gst_buffer_pool_set_config (pool, config);
+  }
+  return TRUE;
+}
+
+
+/* our output size only depends on the caps, not on the input caps */
+static gboolean
+gst_video_filter_transform_size (GstBaseTransform * btrans,
+    GstPadDirection direction, GstCaps * caps, gsize size,
+    GstCaps * othercaps, gsize * othersize)
+{
+  gboolean ret = TRUE;
+  GstVideoInfo info;
+
+  g_assert (size);
+
+  ret = gst_video_info_from_caps (&info, othercaps);
+  if (ret)
+    *othersize = info.size;
+
+  return ret;
 }
 
 static gboolean
@@ -92,8 +153,134 @@ gst_video_filter_get_unit_size (GstBaseTransform * btrans, GstCaps * caps,
   return TRUE;
 }
 
+static gboolean
+gst_video_filter_set_caps (GstBaseTransform * trans, GstCaps * incaps,
+    GstCaps * outcaps)
+{
+  GstVideoFilter *filter = GST_VIDEO_FILTER_CAST (trans);
+  GstVideoFilterClass *fclass;
+  GstVideoInfo in_info, out_info;
+  gboolean res;
+
+  /* input caps */
+  if (!gst_video_info_from_caps (&in_info, incaps))
+    goto invalid_caps;
+
+  /* output caps */
+  if (!gst_video_info_from_caps (&out_info, outcaps))
+    goto invalid_caps;
+
+  fclass = GST_VIDEO_FILTER_GET_CLASS (filter);
+  if (fclass->set_info)
+    res = fclass->set_info (filter, incaps, &in_info, outcaps, &out_info);
+  else
+    res = TRUE;
+
+  if (res) {
+    filter->in_info = in_info;
+    filter->out_info = out_info;
+  }
+  filter->negotiated = res;
+
+  return res;
+
+  /* ERRORS */
+invalid_caps:
+  {
+    GST_ERROR_OBJECT (filter, "invalid caps");
+    filter->negotiated = FALSE;
+    return FALSE;
+  }
+}
+
+static GstFlowReturn
+gst_video_filter_transform (GstBaseTransform * trans, GstBuffer * inbuf,
+    GstBuffer * outbuf)
+{
+  GstFlowReturn res;
+  GstVideoFilter *filter = GST_VIDEO_FILTER_CAST (trans);
+  GstVideoFilterClass *fclass;
+
+  if (G_UNLIKELY (!filter->negotiated))
+    goto unknown_format;
+
+  fclass = GST_VIDEO_FILTER_GET_CLASS (filter);
+  if (fclass->transform_frame) {
+    GstVideoFrame in_frame, out_frame;
+
+    if (!gst_video_frame_map (&in_frame, &filter->in_info, inbuf, GST_MAP_READ))
+      goto invalid_buffer;
+
+    if (!gst_video_frame_map (&out_frame, &filter->out_info, outbuf,
+            GST_MAP_WRITE))
+      goto invalid_buffer;
+
+    res = fclass->transform_frame (filter, &in_frame, &out_frame);
+
+    gst_video_frame_unmap (&out_frame);
+    gst_video_frame_unmap (&in_frame);
+  } else
+    res = GST_FLOW_OK;
+
+  return res;
+
+  /* ERRORS */
+unknown_format:
+  {
+    GST_ELEMENT_ERROR (filter, CORE, NOT_IMPLEMENTED, (NULL),
+        ("unknown format"));
+    return GST_FLOW_NOT_NEGOTIATED;
+  }
+invalid_buffer:
+  {
+    GST_ELEMENT_WARNING (filter, CORE, NOT_IMPLEMENTED, (NULL),
+        ("invalid video buffer received"));
+    return GST_FLOW_OK;
+  }
+}
+
+static GstFlowReturn
+gst_video_filter_transform_ip (GstBaseTransform * trans, GstBuffer * buf)
+{
+  GstFlowReturn res;
+  GstVideoFilter *filter = GST_VIDEO_FILTER_CAST (trans);
+  GstVideoFilterClass *fclass;
+
+  if (G_UNLIKELY (!filter->negotiated))
+    goto unknown_format;
+
+  fclass = GST_VIDEO_FILTER_GET_CLASS (filter);
+  if (fclass->transform_frame_ip) {
+    GstVideoFrame frame;
+
+    if (!gst_video_frame_map (&frame, &filter->in_info, buf, GST_MAP_READWRITE))
+      goto invalid_buffer;
+
+    res = fclass->transform_frame_ip (filter, &frame);
+
+    gst_video_frame_unmap (&frame);
+  } else
+    res = GST_FLOW_OK;
+
+  return res;
+
+  /* ERRORS */
+unknown_format:
+  {
+    GST_ELEMENT_ERROR (filter, CORE, NOT_IMPLEMENTED, (NULL),
+        ("unknown format"));
+    return GST_FLOW_NOT_NEGOTIATED;
+  }
+invalid_buffer:
+  {
+    GST_ELEMENT_WARNING (filter, CORE, NOT_IMPLEMENTED, (NULL),
+        ("invalid video buffer received"));
+    return GST_FLOW_OK;
+  }
+}
+
 static void
-gst_video_filter_class_init (gpointer g_class, gpointer class_data)
+gst_video_filter_class_init (GstVideoFilterClass * g_class)
 {
   GstBaseTransformClass *trans_class;
   GstVideoFilterClass *klass;
@@ -101,23 +288,30 @@ gst_video_filter_class_init (gpointer g_class, gpointer class_data)
   klass = (GstVideoFilterClass *) g_class;
   trans_class = (GstBaseTransformClass *) klass;
 
+  trans_class->set_caps = GST_DEBUG_FUNCPTR (gst_video_filter_set_caps);
+  trans_class->propose_allocation =
+      GST_DEBUG_FUNCPTR (gst_video_filter_propose_allocation);
+  trans_class->decide_allocation =
+      GST_DEBUG_FUNCPTR (gst_video_filter_decide_allocation);
+  trans_class->transform_size =
+      GST_DEBUG_FUNCPTR (gst_video_filter_transform_size);
   trans_class->get_unit_size =
       GST_DEBUG_FUNCPTR (gst_video_filter_get_unit_size);
-
-  parent_class = g_type_class_peek_parent (klass);
+  trans_class->transform = GST_DEBUG_FUNCPTR (gst_video_filter_transform);
+  trans_class->transform_ip = GST_DEBUG_FUNCPTR (gst_video_filter_transform_ip);
 
   GST_DEBUG_CATEGORY_INIT (gst_video_filter_debug, "videofilter", 0,
       "videofilter");
 }
 
 static void
-gst_video_filter_init (GTypeInstance * instance, gpointer g_class)
+gst_video_filter_init (GstVideoFilter * instance)
 {
   GstVideoFilter *videofilter = GST_VIDEO_FILTER (instance);
 
   GST_DEBUG_OBJECT (videofilter, "gst_video_filter_init");
 
-  videofilter->inited = FALSE;
+  videofilter->negotiated = FALSE;
   /* enable QoS */
   gst_base_transform_set_qos_enabled (GST_BASE_TRANSFORM (videofilter), TRUE);
 }
diff --git a/gst-libs/gst/video/gstvideofilter.h b/gst-libs/gst/video/gstvideofilter.h
index bb0183e47c..4259657cd6 100644
--- a/gst-libs/gst/video/gstvideofilter.h
+++ b/gst-libs/gst/video/gstvideofilter.h
@@ -22,6 +22,7 @@
 #define __GST_VIDEO_FILTER_H__
 
 #include <gst/base/gstbasetransform.h>
+#include <gst/video/video.h>
 
 G_BEGIN_DECLS
 
@@ -40,11 +41,14 @@ typedef struct _GstVideoFilterClass GstVideoFilterClass;
   (G_TYPE_CHECK_INSTANCE_TYPE((obj),GST_TYPE_VIDEO_FILTER))
 #define GST_IS_VIDEO_FILTER_CLASS(klass) \
   (G_TYPE_CHECK_CLASS_TYPE((klass),GST_TYPE_VIDEO_FILTER))
+#define GST_VIDEO_FILTER_CAST(obj)  ((GstVideoFilter *)(obj))
 
 struct _GstVideoFilter {
   GstBaseTransform element;
 
-  gboolean inited;
+  gboolean negotiated;
+  GstVideoInfo in_info;
+  GstVideoInfo out_info;
 };
 
 /**
@@ -55,6 +59,15 @@ struct _GstVideoFilter {
  */
 struct _GstVideoFilterClass {
   GstBaseTransformClass parent_class;
+
+  gboolean      (*set_info)           (GstVideoFilter *filter,
+                                       GstCaps *incaps, GstVideoInfo *in_info,
+                                       GstCaps *outcaps, GstVideoInfo *out_info);
+
+  /* transform */
+  GstFlowReturn (*transform_frame)    (GstVideoFilter *filter,
+                                       GstVideoFrame *inframe, GstVideoFrame *outframe);
+  GstFlowReturn (*transform_frame_ip) (GstVideoFilter *trans, GstVideoFrame *frame);
 };
 
 GType gst_video_filter_get_type (void);
diff --git a/gst/videoconvert/gstvideoconvert.c b/gst/videoconvert/gstvideoconvert.c
index c41fd19411..0167cf75da 100644
--- a/gst/videoconvert/gstvideoconvert.c
+++ b/gst/videoconvert/gstvideoconvert.c
@@ -38,6 +38,7 @@
 #endif
 
 #include "gstvideoconvert.h"
+
 #include <gst/video/video.h>
 #include <gst/video/gstvideometa.h>
 #include <gst/video/gstvideopool.h>
@@ -77,13 +78,11 @@ static void gst_video_convert_set_property (GObject * object,
 static void gst_video_convert_get_property (GObject * object,
     guint property_id, GValue * value, GParamSpec * pspec);
 
-static gboolean gst_video_convert_set_caps (GstBaseTransform * btrans,
-    GstCaps * incaps, GstCaps * outcaps);
-static gboolean gst_video_convert_transform_size (GstBaseTransform * btrans,
-    GstPadDirection direction, GstCaps * caps, gsize size,
-    GstCaps * othercaps, gsize * othersize);
-static GstFlowReturn gst_video_convert_transform (GstBaseTransform * btrans,
-    GstBuffer * inbuf, GstBuffer * outbuf);
+static gboolean gst_video_convert_set_info (GstVideoFilter * filter,
+    GstCaps * incaps, GstVideoInfo * in_info, GstCaps * outcaps,
+    GstVideoInfo * out_info);
+static GstFlowReturn gst_video_convert_transform_frame (GstVideoFilter * filter,
+    GstVideoFrame * in_frame, GstVideoFrame * out_frame);
 
 static GType
 dither_method_get_type (void)
@@ -159,96 +158,26 @@ gst_video_convert_transform_caps (GstBaseTransform * btrans,
   return result;
 }
 
-/* Answer the allocation query downstream. This is only called for
- * non-passthrough cases */
 static gboolean
-gst_video_convert_propose_allocation (GstBaseTransform * trans,
-    GstQuery * query)
-{
-  GstVideoConvert *space = GST_VIDEO_CONVERT_CAST (trans);
-  GstBufferPool *pool;
-  GstCaps *caps;
-  gboolean need_pool;
-  guint size;
-
-  gst_query_parse_allocation (query, &caps, &need_pool);
-
-  size = GST_VIDEO_INFO_SIZE (&space->from_info);
-
-  if (need_pool) {
-    GstStructure *structure;
-
-    pool = gst_video_buffer_pool_new ();
-
-    structure = gst_buffer_pool_get_config (pool);
-    gst_buffer_pool_config_set (structure, caps, size, 0, 0, 0, 15);
-    if (!gst_buffer_pool_set_config (pool, structure))
-      goto config_failed;
-  } else
-    pool = NULL;
-
-  gst_query_set_allocation_params (query, size, 0, 0, 0, 15, pool);
-  gst_object_unref (pool);
-
-  gst_query_add_allocation_meta (query, GST_VIDEO_META_API);
-
-  return TRUE;
-
-  /* ERRORS */
-config_failed:
-  {
-    GST_ERROR_OBJECT (space, "failed to set config.");
-    gst_object_unref (pool);
-    return FALSE;
-  }
-}
-
-/* configure the allocation query that was answered downstream, we can configure
- * some properties on it. Only called in passthrough mode. */
-static gboolean
-gst_video_convert_decide_allocation (GstBaseTransform * trans, GstQuery * query)
-{
-  GstBufferPool *pool = NULL;
-  guint size, min, max, prefix, alignment;
-
-  gst_query_parse_allocation_params (query, &size, &min, &max, &prefix,
-      &alignment, &pool);
-
-  if (pool) {
-    GstStructure *config;
-
-    config = gst_buffer_pool_get_config (pool);
-    gst_buffer_pool_config_add_option (config,
-        GST_BUFFER_POOL_OPTION_VIDEO_META);
-    gst_buffer_pool_set_config (pool, config);
-  }
-  return TRUE;
-}
-
-static gboolean
-gst_video_convert_set_caps (GstBaseTransform * btrans, GstCaps * incaps,
-    GstCaps * outcaps)
+gst_video_convert_set_info (GstVideoFilter * filter,
+    GstCaps * incaps, GstVideoInfo * in_info, GstCaps * outcaps,
+    GstVideoInfo * out_info)
 {
   GstVideoConvert *space;
-  GstVideoInfo in_info;
-  GstVideoInfo out_info;
   ColorSpaceColorSpec in_spec, out_spec;
   gboolean interlaced;
 
-  space = GST_VIDEO_CONVERT_CAST (btrans);
+  space = GST_VIDEO_CONVERT_CAST (filter);
 
   if (space->convert) {
     videoconvert_convert_free (space->convert);
   }
 
   /* input caps */
-  if (!gst_video_info_from_caps (&in_info, incaps))
-    goto invalid_caps;
-
-  if (in_info.finfo->flags & GST_VIDEO_FORMAT_FLAG_RGB) {
+  if (GST_VIDEO_INFO_IS_RGB (in_info)) {
     in_spec = COLOR_SPEC_RGB;
-  } else if (in_info.finfo->flags & GST_VIDEO_FORMAT_FLAG_YUV) {
-    if (in_info.colorimetry.matrix == GST_VIDEO_COLOR_MATRIX_BT709)
+  } else if (GST_VIDEO_INFO_IS_YUV (in_info)) {
+    if (in_info->colorimetry.matrix == GST_VIDEO_COLOR_MATRIX_BT709)
       in_spec = COLOR_SPEC_YUV_BT709;
     else
       in_spec = COLOR_SPEC_YUV_BT470_6;
@@ -257,13 +186,10 @@ gst_video_convert_set_caps (GstBaseTransform * btrans, GstCaps * incaps,
   }
 
   /* output caps */
-  if (!gst_video_info_from_caps (&out_info, outcaps))
-    goto invalid_caps;
-
-  if (out_info.finfo->flags & GST_VIDEO_FORMAT_FLAG_RGB) {
+  if (GST_VIDEO_INFO_IS_RGB (out_info)) {
     out_spec = COLOR_SPEC_RGB;
-  } else if (out_info.finfo->flags & GST_VIDEO_FORMAT_FLAG_YUV) {
-    if (out_info.colorimetry.matrix == GST_VIDEO_COLOR_MATRIX_BT709)
+  } else if (GST_VIDEO_INFO_IS_YUV (out_info)) {
+    if (out_info->colorimetry.matrix == GST_VIDEO_COLOR_MATRIX_BT709)
       out_spec = COLOR_SPEC_YUV_BT709;
     else
       out_spec = COLOR_SPEC_YUV_BT470_6;
@@ -272,42 +198,39 @@ gst_video_convert_set_caps (GstBaseTransform * btrans, GstCaps * incaps,
   }
 
   /* these must match */
-  if (in_info.width != out_info.width || in_info.height != out_info.height ||
-      in_info.fps_n != out_info.fps_n || in_info.fps_d != out_info.fps_d)
+  if (in_info->width != out_info->width || in_info->height != out_info->height
+      || in_info->fps_n != out_info->fps_n || in_info->fps_d != out_info->fps_d)
     goto format_mismatch;
 
   /* if present, these must match too */
-  if (in_info.par_n != out_info.par_n || in_info.par_d != out_info.par_d)
+  if (in_info->par_n != out_info->par_n || in_info->par_d != out_info->par_d)
     goto format_mismatch;
 
   /* if present, these must match too */
-  if ((in_info.flags & GST_VIDEO_FLAG_INTERLACED) !=
-      (out_info.flags & GST_VIDEO_FLAG_INTERLACED))
+  if ((in_info->flags & GST_VIDEO_FLAG_INTERLACED) !=
+      (out_info->flags & GST_VIDEO_FLAG_INTERLACED))
     goto format_mismatch;
 
-  space->from_info = in_info;
   space->from_spec = in_spec;
-  space->to_info = out_info;
   space->to_spec = out_spec;
 
-  interlaced = (in_info.flags & GST_VIDEO_FLAG_INTERLACED) != 0;
+  interlaced = (in_info->flags & GST_VIDEO_FLAG_INTERLACED) != 0;
 
   space->convert =
-      videoconvert_convert_new (GST_VIDEO_INFO_FORMAT (&out_info), out_spec,
-      GST_VIDEO_INFO_FORMAT (&in_info), in_spec, in_info.width, in_info.height);
+      videoconvert_convert_new (GST_VIDEO_INFO_FORMAT (out_info), out_spec,
+      GST_VIDEO_INFO_FORMAT (in_info), in_spec, in_info->width,
+      in_info->height);
   if (space->convert == NULL)
     goto no_convert;
 
   videoconvert_convert_set_interlaced (space->convert, interlaced);
 
   /* palette, only for from data */
-  if (GST_VIDEO_INFO_FORMAT (&space->from_info) ==
+  if (GST_VIDEO_INFO_FORMAT (in_info) ==
       GST_VIDEO_FORMAT_RGB8_PALETTED
-      && GST_VIDEO_INFO_FORMAT (&space->to_info) ==
-      GST_VIDEO_FORMAT_RGB8_PALETTED) {
+      && GST_VIDEO_INFO_FORMAT (out_info) == GST_VIDEO_FORMAT_RGB8_PALETTED) {
     goto format_mismatch;
-  } else if (GST_VIDEO_INFO_FORMAT (&space->from_info) ==
-      GST_VIDEO_FORMAT_RGB8_PALETTED) {
+  } else if (GST_VIDEO_INFO_FORMAT (in_info) == GST_VIDEO_FORMAT_RGB8_PALETTED) {
     GstBuffer *palette;
     guint32 *data;
 
@@ -324,8 +247,7 @@ gst_video_convert_set_caps (GstBaseTransform * btrans, GstCaps * incaps,
     gst_buffer_unmap (palette, data, -1);
 
     gst_buffer_unref (palette);
-  } else if (GST_VIDEO_INFO_FORMAT (&space->to_info) ==
-      GST_VIDEO_FORMAT_RGB8_PALETTED) {
+  } else if (GST_VIDEO_INFO_FORMAT (out_info) == GST_VIDEO_FORMAT_RGB8_PALETTED) {
     const guint32 *palette;
     GstBuffer *p_buf;
 
@@ -337,37 +259,25 @@ gst_video_convert_set_caps (GstBaseTransform * btrans, GstCaps * incaps,
     gst_buffer_unref (p_buf);
   }
 
-  GST_DEBUG ("reconfigured %d %d", GST_VIDEO_INFO_FORMAT (&space->from_info),
-      GST_VIDEO_INFO_FORMAT (&space->to_info));
-
-  space->negotiated = TRUE;
+  GST_DEBUG ("reconfigured %d %d", GST_VIDEO_INFO_FORMAT (in_info),
+      GST_VIDEO_INFO_FORMAT (out_info));
 
   return TRUE;
 
   /* ERRORS */
-invalid_caps:
-  {
-    GST_ERROR_OBJECT (space, "invalid caps");
-    goto error_done;
-  }
 format_mismatch:
   {
     GST_ERROR_OBJECT (space, "input and output formats do not match");
-    goto error_done;
+    return FALSE;
   }
 no_convert:
   {
     GST_ERROR_OBJECT (space, "could not create converter");
-    goto error_done;
+    return FALSE;
   }
 invalid_palette:
   {
     GST_ERROR_OBJECT (space, "invalid palette");
-    goto error_done;
-  }
-error_done:
-  {
-    space->negotiated = FALSE;
     return FALSE;
   }
 }
@@ -394,6 +304,7 @@ gst_video_convert_class_init (GstVideoConvertClass * klass)
   GstElementClass *gstelement_class = (GstElementClass *) klass;
   GstBaseTransformClass *gstbasetransform_class =
       (GstBaseTransformClass *) klass;
+  GstVideoFilterClass *gstvideofilter_class = (GstVideoFilterClass *) klass;
 
   gobject_class->set_property = gst_video_convert_set_property;
   gobject_class->get_property = gst_video_convert_get_property;
@@ -411,19 +322,14 @@ gst_video_convert_class_init (GstVideoConvertClass * klass)
 
   gstbasetransform_class->transform_caps =
       GST_DEBUG_FUNCPTR (gst_video_convert_transform_caps);
-  gstbasetransform_class->set_caps =
-      GST_DEBUG_FUNCPTR (gst_video_convert_set_caps);
-  gstbasetransform_class->transform_size =
-      GST_DEBUG_FUNCPTR (gst_video_convert_transform_size);
-  gstbasetransform_class->propose_allocation =
-      GST_DEBUG_FUNCPTR (gst_video_convert_propose_allocation);
-  gstbasetransform_class->decide_allocation =
-      GST_DEBUG_FUNCPTR (gst_video_convert_decide_allocation);
-  gstbasetransform_class->transform =
-      GST_DEBUG_FUNCPTR (gst_video_convert_transform);
 
   gstbasetransform_class->passthrough_on_same_caps = TRUE;
 
+  gstvideofilter_class->set_info =
+      GST_DEBUG_FUNCPTR (gst_video_convert_set_info);
+  gstvideofilter_class->transform_frame =
+      GST_DEBUG_FUNCPTR (gst_video_convert_transform_frame);
+
   g_object_class_install_property (gobject_class, PROP_DITHER,
       g_param_spec_enum ("dither", "Dither", "Apply dithering while converting",
           dither_method_get_type (), DITHER_NONE,
@@ -433,7 +339,6 @@ gst_video_convert_class_init (GstVideoConvertClass * klass)
 static void
 gst_video_convert_init (GstVideoConvert * space)
 {
-  space->negotiated = FALSE;
 }
 
 void
@@ -472,78 +377,26 @@ gst_video_convert_get_property (GObject * object, guint property_id,
   }
 }
 
-static gboolean
-gst_video_convert_transform_size (GstBaseTransform * btrans,
-    GstPadDirection direction, GstCaps * caps, gsize size,
-    GstCaps * othercaps, gsize * othersize)
-{
-  gboolean ret = TRUE;
-  GstVideoInfo info;
-
-  g_assert (size);
-
-  ret = gst_video_info_from_caps (&info, othercaps);
-  if (ret)
-    *othersize = info.size;
-
-  return ret;
-}
-
 static GstFlowReturn
-gst_video_convert_transform (GstBaseTransform * btrans, GstBuffer * inbuf,
-    GstBuffer * outbuf)
+gst_video_convert_transform_frame (GstVideoFilter * filter,
+    GstVideoFrame * in_frame, GstVideoFrame * out_frame)
 {
   GstVideoConvert *space;
-  GstVideoFrame in_frame, out_frame;
 
-  space = GST_VIDEO_CONVERT_CAST (btrans);
+  space = GST_VIDEO_CONVERT_CAST (filter);
 
-  GST_DEBUG ("from %s -> to %s", GST_VIDEO_INFO_NAME (&space->from_info),
-      GST_VIDEO_INFO_NAME (&space->to_info));
-
-  if (G_UNLIKELY (!space->negotiated))
-    goto unknown_format;
+  GST_DEBUG ("from %s -> to %s", GST_VIDEO_INFO_NAME (&filter->in_info),
+      GST_VIDEO_INFO_NAME (&filter->out_info));
 
   videoconvert_convert_set_dither (space->convert, space->dither);
 
-  if (!gst_video_frame_map (&in_frame, &space->from_info, inbuf, GST_MAP_READ))
-    goto invalid_buffer;
-
-  if (!gst_video_frame_map (&out_frame, &space->to_info, outbuf, GST_MAP_WRITE))
-    goto invalid_buffer;
-
-  videoconvert_convert_convert (space->convert, &out_frame, &in_frame);
-
-  gst_video_frame_unmap (&out_frame);
-  gst_video_frame_unmap (&in_frame);
+  videoconvert_convert_convert (space->convert, out_frame, in_frame);
 
   /* baseclass copies timestamps */
-  GST_DEBUG ("from %s -> to %s done", GST_VIDEO_INFO_NAME (&space->from_info),
-      GST_VIDEO_INFO_NAME (&space->to_info));
+  GST_DEBUG ("from %s -> to %s done", GST_VIDEO_INFO_NAME (&filter->in_info),
+      GST_VIDEO_INFO_NAME (&filter->out_info));
 
   return GST_FLOW_OK;
-
-  /* ERRORS */
-unknown_format:
-  {
-    GST_ELEMENT_ERROR (space, CORE, NOT_IMPLEMENTED, (NULL),
-        ("attempting to convert colorspaces between unknown formats"));
-    return GST_FLOW_NOT_NEGOTIATED;
-  }
-invalid_buffer:
-  {
-    GST_ELEMENT_WARNING (space, CORE, NOT_IMPLEMENTED, (NULL),
-        ("invalid video buffer received"));
-    return GST_FLOW_OK;
-  }
-#if 0
-not_supported:
-  {
-    GST_ELEMENT_ERROR (space, CORE, NOT_IMPLEMENTED, (NULL),
-        ("cannot convert between formats"));
-    return GST_FLOW_NOT_SUPPORTED;
-  }
-#endif
 }
 
 static gboolean
diff --git a/gst/videoconvert/gstvideoconvert.h b/gst/videoconvert/gstvideoconvert.h
index abd827052f..fc9a85f686 100644
--- a/gst/videoconvert/gstvideoconvert.h
+++ b/gst/videoconvert/gstvideoconvert.h
@@ -47,10 +47,6 @@ typedef struct _GstVideoConvertClass GstVideoConvertClass;
 struct _GstVideoConvert {
   GstVideoFilter element;
 
-  GstVideoInfo from_info;
-  GstVideoInfo to_info;
-  gboolean negotiated;
-
   ColorSpaceColorSpec from_spec;
   ColorSpaceColorSpec to_spec;
 
diff --git a/gst/videoscale/gstvideoscale.c b/gst/videoscale/gstvideoscale.c
index c20c5f8303..65d24cbff5 100644
--- a/gst/videoscale/gstvideoscale.c
+++ b/gst/videoscale/gstvideoscale.c
@@ -186,19 +186,15 @@ static gboolean gst_video_scale_src_event (GstBaseTransform * trans,
 /* base transform vmethods */
 static GstCaps *gst_video_scale_transform_caps (GstBaseTransform * trans,
     GstPadDirection direction, GstCaps * caps, GstCaps * filter);
-static gboolean gst_video_scale_set_caps (GstBaseTransform * trans,
-    GstCaps * in, GstCaps * out);
-static gboolean gst_video_scale_get_unit_size (GstBaseTransform * trans,
-    GstCaps * caps, gsize * size);
-static gboolean gst_video_scale_propose_allocation (GstBaseTransform * trans,
-    GstQuery * query);
-static gboolean gst_video_scale_decide_allocation (GstBaseTransform * trans,
-    GstQuery * query);
-static GstFlowReturn gst_video_scale_transform (GstBaseTransform * trans,
-    GstBuffer * in, GstBuffer * out);
 static void gst_video_scale_fixate_caps (GstBaseTransform * base,
     GstPadDirection direction, GstCaps * caps, GstCaps * othercaps);
 
+static gboolean gst_video_scale_set_info (GstVideoFilter * filter,
+    GstCaps * in, GstVideoInfo * in_info, GstCaps * out,
+    GstVideoInfo * out_info);
+static GstFlowReturn gst_video_scale_transform_frame (GstVideoFilter * filter,
+    GstVideoFrame * in, GstVideoFrame * out);
+
 static void gst_video_scale_set_property (GObject * object, guint prop_id,
     const GValue * value, GParamSpec * pspec);
 static void gst_video_scale_get_property (GObject * object, guint prop_id,
@@ -213,6 +209,7 @@ gst_video_scale_class_init (GstVideoScaleClass * klass)
   GObjectClass *gobject_class = (GObjectClass *) klass;
   GstElementClass *element_class = (GstElementClass *) klass;
   GstBaseTransformClass *trans_class = (GstBaseTransformClass *) klass;
+  GstVideoFilterClass *filter_class = (GstVideoFilterClass *) klass;
 
   gobject_class->finalize = (GObjectFinalizeFunc) gst_video_scale_finalize;
   gobject_class->set_property = gst_video_scale_set_property;
@@ -270,16 +267,12 @@ gst_video_scale_class_init (GstVideoScaleClass * klass)
 
   trans_class->transform_caps =
       GST_DEBUG_FUNCPTR (gst_video_scale_transform_caps);
-  trans_class->set_caps = GST_DEBUG_FUNCPTR (gst_video_scale_set_caps);
-  trans_class->get_unit_size =
-      GST_DEBUG_FUNCPTR (gst_video_scale_get_unit_size);
-  trans_class->propose_allocation =
-      GST_DEBUG_FUNCPTR (gst_video_scale_propose_allocation);
-  trans_class->decide_allocation =
-      GST_DEBUG_FUNCPTR (gst_video_scale_decide_allocation);
-  trans_class->transform = GST_DEBUG_FUNCPTR (gst_video_scale_transform);
   trans_class->fixate_caps = GST_DEBUG_FUNCPTR (gst_video_scale_fixate_caps);
   trans_class->src_event = GST_DEBUG_FUNCPTR (gst_video_scale_src_event);
+
+  filter_class->set_info = GST_DEBUG_FUNCPTR (gst_video_scale_set_info);
+  filter_class->transform_frame =
+      GST_DEBUG_FUNCPTR (gst_video_scale_transform_frame);
 }
 
 static void
@@ -450,92 +443,21 @@ gst_video_scale_transform_caps (GstBaseTransform * trans,
   return ret;
 }
 
-/* Answer the allocation query downstream. This is only called for
- * non-passthrough cases */
 static gboolean
-gst_video_scale_propose_allocation (GstBaseTransform * trans, GstQuery * query)
+gst_video_scale_set_info (GstVideoFilter * filter, GstCaps * in,
+    GstVideoInfo * in_info, GstCaps * out, GstVideoInfo * out_info)
 {
-  GstVideoScale *scale = GST_VIDEO_SCALE_CAST (trans);
-  GstBufferPool *pool;
-  GstCaps *caps;
-  gboolean need_pool;
-  guint size;
-
-  gst_query_parse_allocation (query, &caps, &need_pool);
-
-  size = GST_VIDEO_INFO_SIZE (&scale->from_info);
-
-  if (need_pool) {
-    GstStructure *structure;
-
-    pool = gst_video_buffer_pool_new ();
-
-    structure = gst_buffer_pool_get_config (pool);
-    gst_buffer_pool_config_set (structure, caps, size, 0, 0, 0, 15);
-    if (!gst_buffer_pool_set_config (pool, structure))
-      goto config_failed;
-  } else
-    pool = NULL;
-
-  gst_query_set_allocation_params (query, size, 0, 0, 0, 15, pool);
-  gst_object_unref (pool);
-
-  gst_query_add_allocation_meta (query, GST_VIDEO_META_API);
-
-  return TRUE;
-
-  /* ERRORS */
-config_failed:
-  {
-    GST_ERROR_OBJECT (scale, "failed to set config.");
-    gst_object_unref (pool);
-    return FALSE;
-  }
-}
-
-/* configure the allocation query that was answered downstream, we can configure
- * some properties on it. Only called in passthrough mode. */
-static gboolean
-gst_video_scale_decide_allocation (GstBaseTransform * trans, GstQuery * query)
-{
-  GstBufferPool *pool = NULL;
-  guint size, min, max, prefix, alignment;
-
-  gst_query_parse_allocation_params (query, &size, &min, &max, &prefix,
-      &alignment, &pool);
-
-  if (pool) {
-    GstStructure *config;
-
-    config = gst_buffer_pool_get_config (pool);
-    gst_buffer_pool_config_add_option (config,
-        GST_BUFFER_POOL_OPTION_VIDEO_META);
-    gst_buffer_pool_set_config (pool, config);
-  }
-  return TRUE;
-}
-
-static gboolean
-gst_video_scale_set_caps (GstBaseTransform * trans, GstCaps * in, GstCaps * out)
-{
-  GstVideoScale *videoscale = GST_VIDEO_SCALE (trans);
-  gboolean ret;
-  GstVideoInfo in_info, out_info;
+  GstVideoScale *videoscale = GST_VIDEO_SCALE (filter);
   gint from_dar_n, from_dar_d, to_dar_n, to_dar_d;
 
-  ret = gst_video_info_from_caps (&in_info, in);
-  ret &= gst_video_info_from_caps (&out_info, out);
-  if (!ret)
-    goto invalid_formats;
-
-  if (!gst_util_fraction_multiply (in_info.width,
-          in_info.height, out_info.par_n, out_info.par_d, &from_dar_n,
+  if (!gst_util_fraction_multiply (in_info->width,
+          in_info->height, out_info->par_n, out_info->par_d, &from_dar_n,
           &from_dar_d)) {
     from_dar_n = from_dar_d = -1;
   }
 
-  if (!gst_util_fraction_multiply (out_info.width,
-          out_info.height, out_info.par_n, out_info.par_d, &to_dar_n,
+  if (!gst_util_fraction_multiply (out_info->width,
+          out_info->height, out_info->par_n, out_info->par_d, &to_dar_n,
           &to_dar_d)) {
     to_dar_n = to_dar_d = -1;
   }
@@ -546,17 +468,17 @@ gst_video_scale_set_caps (GstBaseTransform * trans, GstCaps * in, GstCaps * out)
       gint n, d, to_h, to_w;
 
       if (from_dar_n != -1 && from_dar_d != -1
-          && gst_util_fraction_multiply (from_dar_n, from_dar_d, out_info.par_n,
-              out_info.par_d, &n, &d)) {
-        to_h = gst_util_uint64_scale_int (out_info.width, d, n);
-        if (to_h <= out_info.height) {
-          videoscale->borders_h = out_info.height - to_h;
+          && gst_util_fraction_multiply (from_dar_n, from_dar_d,
+              out_info->par_n, out_info->par_d, &n, &d)) {
+        to_h = gst_util_uint64_scale_int (out_info->width, d, n);
+        if (to_h <= out_info->height) {
+          videoscale->borders_h = out_info->height - to_h;
           videoscale->borders_w = 0;
         } else {
-          to_w = gst_util_uint64_scale_int (out_info.height, n, d);
-          g_assert (to_w <= out_info.width);
+          to_w = gst_util_uint64_scale_int (out_info->height, n, d);
+          g_assert (to_w <= out_info->width);
           videoscale->borders_h = 0;
-          videoscale->borders_w = out_info.width - to_w;
+          videoscale->borders_w = out_info->width - to_w;
         }
       } else {
         GST_WARNING_OBJECT (videoscale, "Can't calculate borders");
@@ -568,42 +490,19 @@ gst_video_scale_set_caps (GstBaseTransform * trans, GstCaps * in, GstCaps * out)
 
   if (videoscale->tmp_buf)
     g_free (videoscale->tmp_buf);
-  videoscale->tmp_buf = g_malloc (out_info.width * 8 * 4);
+  videoscale->tmp_buf = g_malloc (out_info->width * 8 * 4);
 
-  gst_base_transform_set_passthrough (trans,
-      (in_info.width == out_info.width && in_info.height == out_info.height));
+  gst_base_transform_set_passthrough (GST_BASE_TRANSFORM (filter),
+      (in_info->width == out_info->width
+          && in_info->height == out_info->height));
 
   GST_DEBUG_OBJECT (videoscale, "from=%dx%d (par=%d/%d dar=%d/%d), size %"
       G_GSIZE_FORMAT " -> to=%dx%d (par=%d/%d dar=%d/%d borders=%d:%d), "
       "size %" G_GSIZE_FORMAT,
-      in_info.width, in_info.height, out_info.par_n, out_info.par_d,
-      from_dar_n, from_dar_d, in_info.size, out_info.width,
-      out_info.height, out_info.par_n, out_info.par_d, to_dar_n, to_dar_d,
-      videoscale->borders_w, videoscale->borders_h, out_info.size);
-
-  videoscale->from_info = in_info;
-  videoscale->to_info = out_info;
-
-  return TRUE;
-
-  /* ERRORS */
-invalid_formats:
-  {
-    GST_DEBUG_OBJECT (videoscale, "could not parse formats");
-    return FALSE;
-  }
-}
-
-static gboolean
-gst_video_scale_get_unit_size (GstBaseTransform * trans, GstCaps * caps,
-    gsize * size)
-{
-  GstVideoInfo info;
-
-  if (!gst_video_info_from_caps (&info, caps))
-    return FALSE;
-
-  *size = info.size;
+      in_info->width, in_info->height, out_info->par_n, out_info->par_d,
+      from_dar_n, from_dar_d, in_info->size, out_info->width,
+      out_info->height, out_info->par_n, out_info->par_d, to_dar_n, to_dar_d,
+      videoscale->borders_w, videoscale->borders_h, out_info->size);
 
   return TRUE;
 }
@@ -1149,12 +1048,11 @@ _get_black_for_format (GstVideoFormat format)
 }
 
 static GstFlowReturn
-gst_video_scale_transform (GstBaseTransform * trans, GstBuffer * in,
-    GstBuffer * out)
+gst_video_scale_transform_frame (GstVideoFilter * filter,
+    GstVideoFrame * in_frame, GstVideoFrame * out_frame)
 {
-  GstVideoScale *videoscale = GST_VIDEO_SCALE (trans);
+  GstVideoScale *videoscale = GST_VIDEO_SCALE (filter);
   GstFlowReturn ret = GST_FLOW_OK;
-  GstVideoFrame in_frame, out_frame;
   VSImage dest[4] = { {NULL,}, };
   VSImage src[4] = { {NULL,}, };
   gint method;
@@ -1168,23 +1066,20 @@ gst_video_scale_transform (GstBaseTransform * trans, GstBuffer * in,
   add_borders = videoscale->add_borders;
   GST_OBJECT_UNLOCK (videoscale);
 
-  format = GST_VIDEO_INFO_FORMAT (&videoscale->from_info);
+  format = GST_VIDEO_INFO_FORMAT (&filter->in_info);
   black = _get_black_for_format (format);
 
-  if (videoscale->from_info.width == 1) {
+  if (filter->in_info.width == 1) {
     method = GST_VIDEO_SCALE_NEAREST;
   }
   if (method == GST_VIDEO_SCALE_4TAP &&
-      (videoscale->from_info.width < 4 || videoscale->from_info.height < 4)) {
+      (filter->in_info.width < 4 || filter->in_info.height < 4)) {
     method = GST_VIDEO_SCALE_BILINEAR;
   }
 
-  gst_video_frame_map (&in_frame, &videoscale->from_info, in, GST_MAP_READ);
-  gst_video_frame_map (&out_frame, &videoscale->to_info, out, GST_MAP_WRITE);
-
-  for (i = 0; i < GST_VIDEO_FRAME_N_PLANES (&in_frame); i++) {
-    gst_video_scale_setup_vs_image (&src[i], &in_frame, i, 0, 0);
-    gst_video_scale_setup_vs_image (&dest[i], &out_frame, i,
+  for (i = 0; i < GST_VIDEO_FRAME_N_PLANES (in_frame); i++) {
+    gst_video_scale_setup_vs_image (&src[i], in_frame, i, 0, 0);
+    gst_video_scale_setup_vs_image (&dest[i], out_frame, i,
         videoscale->borders_w, videoscale->borders_h);
   }
 
@@ -1410,13 +1305,6 @@ gst_video_scale_transform (GstBaseTransform * trans, GstBuffer * in,
       goto unsupported;
   }
 
-  GST_LOG_OBJECT (videoscale, "pushing buffer of %" G_GSIZE_FORMAT " bytes",
-      gst_buffer_get_size (out));
-
-done:
-  gst_video_frame_unmap (&out_frame);
-  gst_video_frame_unmap (&in_frame);
-
   return ret;
 
   /* ERRORS */
@@ -1424,22 +1312,21 @@ unsupported:
   {
     GST_ELEMENT_ERROR (videoscale, STREAM, NOT_IMPLEMENTED, (NULL),
         ("Unsupported format %d for scaling method %d", format, method));
-    ret = GST_FLOW_ERROR;
-    goto done;
+    return GST_FLOW_ERROR;
   }
 unknown_mode:
   {
     GST_ELEMENT_ERROR (videoscale, STREAM, NOT_IMPLEMENTED, (NULL),
         ("Unknown scaling method %d", videoscale->method));
-    ret = GST_FLOW_ERROR;
-    goto done;
+    return GST_FLOW_ERROR;
   }
 }
 
 static gboolean
 gst_video_scale_src_event (GstBaseTransform * trans, GstEvent * event)
 {
-  GstVideoScale *videoscale = GST_VIDEO_SCALE (trans);
+  GstVideoScale *videoscale = GST_VIDEO_SCALE_CAST (trans);
+  GstVideoFilter *filter = GST_VIDEO_FILTER_CAST (trans);
   gboolean ret;
   gdouble a;
   GstStructure *structure;
@@ -1455,12 +1342,11 @@ gst_video_scale_src_event (GstBaseTransform * trans, GstEvent * event)
       structure = (GstStructure *) gst_event_get_structure (event);
       if (gst_structure_get_double (structure, "pointer_x", &a)) {
         gst_structure_set (structure, "pointer_x", G_TYPE_DOUBLE,
-            a * videoscale->from_info.width / videoscale->to_info.width, NULL);
+            a * filter->in_info.width / filter->out_info.width, NULL);
       }
       if (gst_structure_get_double (structure, "pointer_y", &a)) {
         gst_structure_set (structure, "pointer_y", G_TYPE_DOUBLE,
-            a * videoscale->from_info.height / videoscale->to_info.height,
-            NULL);
+            a * filter->in_info.height / filter->out_info.height, NULL);
       }
       break;
     default:
diff --git a/gst/videoscale/gstvideoscale.h b/gst/videoscale/gstvideoscale.h
index 26ce7bcfd2..85c2782305 100644
--- a/gst/videoscale/gstvideoscale.h
+++ b/gst/videoscale/gstvideoscale.h
@@ -80,10 +80,6 @@ struct _GstVideoScale {
   int submethod;
   double envelope;
 
-  /* negotiated stuff */
-  GstVideoInfo from_info;
-  GstVideoInfo to_info;
-
   gint borders_h;
   gint borders_w;