On 02/13/2015 11:18 PM, Diego Biurrun wrote:
On Fri, Feb 13, 2015 at 10:43:16PM +0100, Thomas Volkert wrote:
--- a/libavformat/rtpdec_hevc.c
+++ b/libavformat/rtpdec_hevc.c
@@ -355,6 +345,78 @@ static int hevc_handle_packet(AVFormatContext *ctx,
PayloadContext *rtp_hevc_ctx
memcpy(pkt->data + sizeof(start_sequence), buf, len);
break;
+ /* aggregated packet (AP) - with two or more NAL units */
+ case 48:
+ /* pass the HEVC payload header */
+ buf += RTP_HEVC_PAYLOAD_HEADER_SIZE;
+ len -= RTP_HEVC_PAYLOAD_HEADER_SIZE;
+
+ /* pass the HEVC DONL field */
+ if (rtp_hevc_ctx->using_donl_field) {
+ buf += RTP_HEVC_DONL_FIELD_SIZE;
+ len -= RTP_HEVC_DONL_FIELD_SIZE;
+ }
+
+ /*
+ * pass 0: determine overall size of the A/V packet
+ * pass 1: create resulting A/V packet
+ */
+ {
+ int pass = 0;
+ int pkt_size = 0;
+ uint8_t *pkt_data = NULL;
Is there a reason to introduce a new block here instead of moving the
variable declarations up a bit?
I tried to avoid to reinvent everything.
My code is intentionally kept similar to the one of the H.264 parser.
Maybe we will find a way to generalize the code and use it both in the
H.264 and HEVC parser in the near future.
Best regards,
Thomas.
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
