avcodec/libaomenc: Avoid copying data, allow user-supplied buffers

Here the packet size is known before allocating the packet because
the encoder provides said information (and works with internal buffers
itself), so one can use this information to avoid the implicit use of
another intermediate buffer for the packet data; and by switching to
ff_get_encode_buffer() one can also allow user-supplied buffers.

Reviewed-by: James Almer <jamrial@gmail.com>
Reviewed-by: James Zern <jzern@google.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This commit is contained in:
Andreas Rheinhardt 2021-04-25 01:43:26 +02:00
parent c685249ff8
commit 044daa7c45

View File

@ -36,6 +36,7 @@
#include "av1.h"
#include "avcodec.h"
#include "encode.h"
#include "internal.h"
#include "packet_internal.h"
#include "profiles.h"
@ -970,7 +971,7 @@ static int storeframe(AVCodecContext *avctx, struct FrameListData *cx_frame,
{
AOMContext *ctx = avctx->priv_data;
int av_unused pict_type;
int ret = ff_alloc_packet2(avctx, pkt, cx_frame->sz, 0);
int ret = ff_get_encode_buffer(avctx, pkt, cx_frame->sz, 0);
if (ret < 0) {
av_log(avctx, AV_LOG_ERROR,
"Error getting output packet of size %"SIZE_SPECIFIER".\n", cx_frame->sz);
@ -1344,11 +1345,12 @@ AVCodec ff_libaom_av1_encoder = {
.long_name = NULL_IF_CONFIG_SMALL("libaom AV1"),
.type = AVMEDIA_TYPE_VIDEO,
.id = AV_CODEC_ID_AV1,
.capabilities = AV_CODEC_CAP_DR1 | AV_CODEC_CAP_DELAY |
AV_CODEC_CAP_OTHER_THREADS,
.priv_data_size = sizeof(AOMContext),
.init = av1_init,
.encode2 = aom_encode,
.close = aom_free,
.capabilities = AV_CODEC_CAP_DELAY | AV_CODEC_CAP_OTHER_THREADS,
.caps_internal = FF_CODEC_CAP_AUTO_THREADS,
.profiles = NULL_IF_CONFIG_SMALL(ff_av1_profiles),
.priv_class = &class_aom,