Extend EncodedVideoChunkMetadata for Spatial SVC

index.src.html

@@ -54,6 +54,15 @@
 
 <pre class='biblio'>
 {
+  "AV1-RTP": {
+    "title": "RTP Payload Format for AV1",
+    "href": "https://aomediacodec.github.io/av1-rtp-spec/",
+    "authors": [
+      "AV1 RTC SG"
+    ],
+    "status": "Standard",
+    "publisher": "Alliance for Open Media"
+  },
   "H.273": {
     "href": "https://www.itu.int/rec/T-REC-H.273/en",
     "title": "Coding-independent code points for video signal type identification",
@@ -128,6 +137,24 @@
 :: A grouping of {{EncodedVideoChunk}}s whose timestamp cadence produces a
     particular framerate. See {{VideoEncoderConfig/scalabilityMode}}.
 
+: <dfn>Spatial Layer</dfn>
+:: A grouping of {{EncodedVideoChunk}}s which produces a particular
+    resolution. See {{VideoEncoderConfig/scalabilityMode}}.
+
+: <dfn>Chains</dfn>
+:: A sequence of frames for which it can be determined instantly if a frame
+     from that sequence has been lost. Defined in Section A.5 of [[AV1-RTP]].
+
+: <dfn>Frame Number</dfn>
+:: A 16-bit number that increases monotonically in decode order. Defined in
+     Section A.8.3 of [[AV1-RTP]].
+
+: <dfn>Decode Target</dfn>
+:: A numerical index determined by the encoder that indicates the set of
+     frames needed to decode a sequence of {{EncodedVideoChunk}}s at a
+     given spatial and temporal fidelity. Values do not necessarily
+     correspond to a given [=Temporal Layer=] or [=Spatial Layer=].
+
 : <dfn>Progressive Image</dfn>
 :: An image that supports decoding to multiple levels of detail, with lower
     levels becoming available while the encoded data is not yet fully buffered.
@@ -1639,11 +1666,20 @@
                 |svc|.{{SvcOutputMetadata/temporalLayerId}}.
             4. Assign |svc| to
                 |chunkMetadata|.{{EncodedVideoChunkMetadata/svc}}.
-        8. If |encoderConfig|.{{VideoEncoderConfig/alpha}} is set to `"keep"`:
+        8. If |encoderConfig|.{{VideoEncoderConfig/scalabilityMode}}
+            describes more than one [=Spatial Layer=]:
+            1. Let |svc| be a new {{SvcOutputMetadata}} instance.
+            2. Let |spatial_layer_id| be the zero-based index describing the
+                spatial layer for |output|.
+            3. Assign |spatial_layer_id| to
+                |svc|.{{SvcOutputMetadata/spatialLayerId}}.
+            4. Assign |svc| to
+                |chunkMetadata|.{{EncodedVideoChunkMetadata/svc}}.
+        9. If |encoderConfig|.{{VideoEncoderConfig/alpha}} is set to `"keep"`:
             1. Let |alphaSideData| be the encoded alpha data in |output|.
             2. Assign |alphaSideData| to
                 |chunkMetadata|.{{EncodedVideoChunkMetadata/alphaSideData}}.
-        9. Invoke {{VideoEncoder/[[output callback]]}} with |chunk| and
+        10. Invoke {{VideoEncoder/[[output callback]]}} with |chunk| and
             |chunkMetadata|.
   </dd>
   <dt><dfn>Reset VideoEncoder</dfn> (with |exception|)</dt>
@@ -1692,6 +1728,11 @@
 
 dictionary SvcOutputMetadata {
   unsigned long temporalLayerId;
+  unsigned long spatialLayerId;
+  unsigned long frameNumber;
+  sequence <unsigned long> dependencies;
+  sequence <unsigned long> decodeTargets;
+  record <USVString, unsigned long> chainLinks;
 };
 </xmp>
 
@@ -1708,9 +1749,27 @@
     channel data.
 
 : <dfn dict-member for=SvcOutputMetadata>temporalLayerId</dfn>
-:: A number that identifies the [=temporal layer=] for the associated
+:: A number that identifies the [=Temporal Layer=] for the associated
     {{EncodedVideoChunk}}.
-
+
+: <dfn dict-member for=SvcOutputMetadata>spatialLayerId</dfn>
+:: A number that identifies the [=Spatial Layer=] for the associated
+    {{EncodedVideoChunk}}.
+
+: <dfn dict-member for=SvcOutputMetadata>frameNumber</dfn>
+:: A number that identifies the frame.
+
+: <dfn dict-member for=SvcOutputMetadata>dependencies</dfn>
+:: A sequence of frame number values that this {{EncodedVideoChunk}}
+     depends on.
+
+: <dfn dict-member for=SvcOutputMetadata>decodeTargets</dfn>
+:: A sequence of [=Decode Target=] values that this {{EncodedVideoChunk}}
+     participates in.
+
+: <dfn dict-member for=SvcOutputMetadata>chainLinks</dfn>
+:: A mapping of [=Decode Target=] values to the frame number of the
+     last important frame to decode prior to this {{EncodedVideoChunk}}.
 
 Configurations{#configurations}
 ===============================