gstreamer/subprojects/gst-plugins-bad/sys/nvcodec/gstnvencoder.h

313 lines
9.3 KiB
C
Raw Normal View History

nvcodec: Add new Direct3D11/CUDA mode encoder implementation Adding new encoder elements nvd3d11{h264,h265}enc for Direct3D11 input support and re-written nvcuda{h264,h265}enc elements. Newly writeen elements have some differences compared with old nv{h264,h265}enc including non-backward compatible changes. * RGBA is not a supported input format any more: New elements will support only YUV formats to avoid implicit conversion done by hardware. Ideally it should be done by upstream element in order to have more control on it. Moreover, RGBA support can cause redundant RGBA -> YUV conversion if multiple encoders are used for the same RGBA input * Subsampled planar format support is dropped: I420 and YV12 format are not supported formats for Direct3D11. Although it's supported in CUDA mode, it's not a hardware friendly memory layout and it will waste GPU memory since UV planes will have large padding due to the memory layout requirement of NVENC. * GL support is dropped: Similar to the RGBA case, GL support in encoder would be suboptimal if GL input is used by multiple encoders, because each encoder will copy GL memory into CUDA memory. Upstream cudaupload element can be used for GL <-> CUDA interop instead. * No more pre-allocation of encoder input surfaces. New implementation will use input CUDA memory without copy (zero-copy) or will copy into a NVENC's input buffer struct in case of system memory input. Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1997>
2022-03-15 19:58:16 +00:00
/* GStreamer
* Copyright (C) 2022 Seungha Yang <seungha@centricular.com>
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Library General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Library General Public License for more details.
*
* You should have received a copy of the GNU Library General Public
* License along with this library; if not, write to the
* Free Software Foundation, Inc., 51 Franklin St, Fifth Floor,
* Boston, MA 02110-1301, USA.
*/
#pragma once
#include <gst/gst.h>
#include <gst/video/video.h>
#ifdef G_OS_WIN32
nvcodec: Add new Direct3D11/CUDA mode encoder implementation Adding new encoder elements nvd3d11{h264,h265}enc for Direct3D11 input support and re-written nvcuda{h264,h265}enc elements. Newly writeen elements have some differences compared with old nv{h264,h265}enc including non-backward compatible changes. * RGBA is not a supported input format any more: New elements will support only YUV formats to avoid implicit conversion done by hardware. Ideally it should be done by upstream element in order to have more control on it. Moreover, RGBA support can cause redundant RGBA -> YUV conversion if multiple encoders are used for the same RGBA input * Subsampled planar format support is dropped: I420 and YV12 format are not supported formats for Direct3D11. Although it's supported in CUDA mode, it's not a hardware friendly memory layout and it will waste GPU memory since UV planes will have large padding due to the memory layout requirement of NVENC. * GL support is dropped: Similar to the RGBA case, GL support in encoder would be suboptimal if GL input is used by multiple encoders, because each encoder will copy GL memory into CUDA memory. Upstream cudaupload element can be used for GL <-> CUDA interop instead. * No more pre-allocation of encoder input surfaces. New implementation will use input CUDA memory without copy (zero-copy) or will copy into a NVENC's input buffer struct in case of system memory input. Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1997>
2022-03-15 19:58:16 +00:00
#include <gst/d3d11/gstd3d11.h>
#endif
#ifdef HAVE_CUDA_GST_GL
#include <gst/gl/gl.h>
#endif
nvcodec: Add new Direct3D11/CUDA mode encoder implementation Adding new encoder elements nvd3d11{h264,h265}enc for Direct3D11 input support and re-written nvcuda{h264,h265}enc elements. Newly writeen elements have some differences compared with old nv{h264,h265}enc including non-backward compatible changes. * RGBA is not a supported input format any more: New elements will support only YUV formats to avoid implicit conversion done by hardware. Ideally it should be done by upstream element in order to have more control on it. Moreover, RGBA support can cause redundant RGBA -> YUV conversion if multiple encoders are used for the same RGBA input * Subsampled planar format support is dropped: I420 and YV12 format are not supported formats for Direct3D11. Although it's supported in CUDA mode, it's not a hardware friendly memory layout and it will waste GPU memory since UV planes will have large padding due to the memory layout requirement of NVENC. * GL support is dropped: Similar to the RGBA case, GL support in encoder would be suboptimal if GL input is used by multiple encoders, because each encoder will copy GL memory into CUDA memory. Upstream cudaupload element can be used for GL <-> CUDA interop instead. * No more pre-allocation of encoder input surfaces. New implementation will use input CUDA memory without copy (zero-copy) or will copy into a NVENC's input buffer struct in case of system memory input. Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1997>
2022-03-15 19:58:16 +00:00
#include <string.h>
#include <gst/cuda/gstcuda.h>
nvcodec: Add new Direct3D11/CUDA mode encoder implementation Adding new encoder elements nvd3d11{h264,h265}enc for Direct3D11 input support and re-written nvcuda{h264,h265}enc elements. Newly writeen elements have some differences compared with old nv{h264,h265}enc including non-backward compatible changes. * RGBA is not a supported input format any more: New elements will support only YUV formats to avoid implicit conversion done by hardware. Ideally it should be done by upstream element in order to have more control on it. Moreover, RGBA support can cause redundant RGBA -> YUV conversion if multiple encoders are used for the same RGBA input * Subsampled planar format support is dropped: I420 and YV12 format are not supported formats for Direct3D11. Although it's supported in CUDA mode, it's not a hardware friendly memory layout and it will waste GPU memory since UV planes will have large padding due to the memory layout requirement of NVENC. * GL support is dropped: Similar to the RGBA case, GL support in encoder would be suboptimal if GL input is used by multiple encoders, because each encoder will copy GL memory into CUDA memory. Upstream cudaupload element can be used for GL <-> CUDA interop instead. * No more pre-allocation of encoder input surfaces. New implementation will use input CUDA memory without copy (zero-copy) or will copy into a NVENC's input buffer struct in case of system memory input. Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1997>
2022-03-15 19:58:16 +00:00
#include "nvEncodeAPI.h"
#include "gstnvenc.h"
G_BEGIN_DECLS
#define GST_TYPE_NV_ENCODER (gst_nv_encoder_get_type())
#define GST_NV_ENCODER(obj) (G_TYPE_CHECK_INSTANCE_CAST((obj), GST_TYPE_NV_ENCODER, GstNvEncoder))
#define GST_NV_ENCODER_CLASS(klass) (G_TYPE_CHECK_CLASS_CAST((klass), GST_TYPE_NV_ENCODER, GstNvEncoderClass))
#define GST_IS_NV_ENCODER(obj) (G_TYPE_CHECK_INSTANCE_TYPE((obj), GST_TYPE_NV_ENCODER))
#define GST_IS_NV_ENCODER_CLASS(klass) (G_TYPE_CHECK_CLASS_TYPE((klass), GST_TYPE_NV_ENCODER))
#define GST_NV_ENCODER_GET_CLASS(obj) (G_TYPE_INSTANCE_GET_CLASS((obj), GST_TYPE_NV_ENCODER, GstNvEncoderClass))
#define GST_NV_ENCODER_CAST(obj) ((GstNvEncoder *)obj)
typedef struct _GstNvEncoder GstNvEncoder;
typedef struct _GstNvEncoderClass GstNvEncoderClass;
typedef struct _GstNvEncoderPrivate GstNvEncoderPrivate;
typedef enum
{
GST_NV_ENCODER_RECONFIGURE_NONE,
GST_NV_ENCODER_RECONFIGURE_BITRATE,
GST_NV_ENCODER_RECONFIGURE_FULL,
} GstNvEncoderReconfigure;
#define GST_TYPE_NV_ENCODER_PRESET (gst_nv_encoder_preset_get_type())
GType gst_nv_encoder_preset_get_type (void);
typedef enum
{
GST_NV_ENCODER_PRESET_DEFAULT,
GST_NV_ENCODER_PRESET_HP,
GST_NV_ENCODER_PRESET_HQ,
GST_NV_ENCODER_PRESET_LOW_LATENCY_DEFAULT,
GST_NV_ENCODER_PRESET_LOW_LATENCY_HQ,
GST_NV_ENCODER_PRESET_LOW_LATENCY_HP,
GST_NV_ENCODER_PRESET_LOSSLESS_DEFAULT,
GST_NV_ENCODER_PRESET_LOSSLESS_HP,
GST_NV_ENCODER_PRESET_P1,
GST_NV_ENCODER_PRESET_P2,
GST_NV_ENCODER_PRESET_P3,
GST_NV_ENCODER_PRESET_P4,
GST_NV_ENCODER_PRESET_P5,
GST_NV_ENCODER_PRESET_P6,
GST_NV_ENCODER_PRESET_P7,
nvcodec: Add new Direct3D11/CUDA mode encoder implementation Adding new encoder elements nvd3d11{h264,h265}enc for Direct3D11 input support and re-written nvcuda{h264,h265}enc elements. Newly writeen elements have some differences compared with old nv{h264,h265}enc including non-backward compatible changes. * RGBA is not a supported input format any more: New elements will support only YUV formats to avoid implicit conversion done by hardware. Ideally it should be done by upstream element in order to have more control on it. Moreover, RGBA support can cause redundant RGBA -> YUV conversion if multiple encoders are used for the same RGBA input * Subsampled planar format support is dropped: I420 and YV12 format are not supported formats for Direct3D11. Although it's supported in CUDA mode, it's not a hardware friendly memory layout and it will waste GPU memory since UV planes will have large padding due to the memory layout requirement of NVENC. * GL support is dropped: Similar to the RGBA case, GL support in encoder would be suboptimal if GL input is used by multiple encoders, because each encoder will copy GL memory into CUDA memory. Upstream cudaupload element can be used for GL <-> CUDA interop instead. * No more pre-allocation of encoder input surfaces. New implementation will use input CUDA memory without copy (zero-copy) or will copy into a NVENC's input buffer struct in case of system memory input. Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1997>
2022-03-15 19:58:16 +00:00
} GstNvEncoderPreset;
#define GST_TYPE_NV_ENCODER_RC_MODE (gst_nv_encoder_rc_mode_get_type())
GType gst_nv_encoder_rc_mode_get_type (void);
typedef enum
{
GST_NV_ENCODER_RC_MODE_CONSTQP,
GST_NV_ENCODER_RC_MODE_VBR,
GST_NV_ENCODER_RC_MODE_CBR,
GST_NV_ENCODER_RC_MODE_CBR_LOWDELAY_HQ,
GST_NV_ENCODER_RC_MODE_CBR_HQ,
GST_NV_ENCODER_RC_MODE_VBR_HQ,
} GstNvEncoderRCMode;
#define GST_TYPE_NV_ENCODER_SEI_INSERT_MODE (gst_nv_encoder_sei_insert_mode_get_type ())
GType gst_nv_encoder_sei_insert_mode_get_type (void);
typedef enum
{
GST_NV_ENCODER_SEI_INSERT,
GST_NV_ENCODER_SEI_INSERT_AND_DROP,
GST_NV_ENCODER_SEI_DISABLED,
} GstNvEncoderSeiInsertMode;
#define GST_TYPE_NV_ENCODER_MULTI_PASS (gst_nv_encoder_multi_pass_get_type ())
GType gst_nv_encoder_multi_pass_get_type (void);
typedef enum
{
GST_NV_ENCODER_MULTI_PASS_DEFAULT = 0,
GST_NV_ENCODER_MULTI_PASS_DISABLED = 1,
GST_NV_ENCODER_TWO_PASS_QUARTER_RESOLUTION = 2,
GST_NV_ENCODER_TWO_PASS_FULL_RESOLUTION = 3,
} GstNvEncoderMultiPass;
#define GST_TYPE_NV_ENCODER_TUNE (gst_nv_encoder_tune_get_type ())
GType gst_nv_encoder_tune_get_type (void);
typedef enum
{
GST_NV_ENCODER_TUNE_DEFAULT = 0,
GST_NV_ENCODER_TUNE_HIGH_QUALITY = 1,
GST_NV_ENCODER_TUNE_LOW_LATENCY = 2,
GST_NV_ENCODER_TUNE_ULTRA_LOW_LATENCY = 3,
GST_NV_ENCODER_TUNE_LOSSLESS = 4,
} GstNvEncoderTune;
typedef struct
{
gint max_bframes;
gint ratecontrol_modes;
gint field_encoding;
gint monochrome;
gint fmo;
gint qpelmv;
gint bdirect_mode;
gint cabac;
gint adaptive_transform;
gint stereo_mvc;
gint temoral_layers;
gint hierarchical_pframes;
gint hierarchical_bframes;
gint level_max;
gint level_min;
gint separate_colour_plane;
gint width_max;
gint height_max;
gint temporal_svc;
gint dyn_res_change;
gint dyn_bitrate_change;
gint dyn_force_constqp;
gint dyn_rcmode_change;
gint subframe_readback;
gint constrained_encoding;
gint intra_refresh;
gint custom_vbv_buf_size;
gint dynamic_slice_mode;
gint ref_pic_invalidation;
gint preproc_support;
gint async_encoding_support;
gint mb_num_max;
gint mb_per_sec_max;
gint yuv444_encode;
gint lossless_encode;
gint sao;
gint meonly_mode;
gint lookahead;
gint temporal_aq;
gint supports_10bit_encode;
gint num_max_ltr_frames;
gint weighted_prediction;
gint bframe_ref_mode;
gint emphasis_level_map;
gint width_min;
gint height_min;
gint multiple_ref_frames;
} GstNvEncoderDeviceCaps;
typedef enum
{
GST_NV_ENCODER_DEVICE_D3D11,
GST_NV_ENCODER_DEVICE_CUDA,
GST_NV_ENCODER_DEVICE_AUTO_SELECT,
} GstNvEncoderDeviceMode;
typedef struct
{
GstCaps *sink_caps;
GstCaps *src_caps;
guint cuda_device_id;
gint64 adapter_luid;
GstNvEncoderDeviceMode device_mode;
GstNvEncoderDeviceCaps device_caps;
GList *formats;
GList *profiles;
/* auto gpu select mode */
guint adapter_luid_size;
gint64 adapter_luid_list[8];
guint cuda_device_id_size;
guint cuda_device_id_list[8];
gint ref_count;
} GstNvEncoderClassData;
typedef struct
{
GstNvEncoderDeviceMode device_mode;
guint cuda_device_id;
gint64 adapter_luid;
GstObject *device;
} GstNvEncoderDeviceData;
nvcodec: Add new Direct3D11/CUDA mode encoder implementation Adding new encoder elements nvd3d11{h264,h265}enc for Direct3D11 input support and re-written nvcuda{h264,h265}enc elements. Newly writeen elements have some differences compared with old nv{h264,h265}enc including non-backward compatible changes. * RGBA is not a supported input format any more: New elements will support only YUV formats to avoid implicit conversion done by hardware. Ideally it should be done by upstream element in order to have more control on it. Moreover, RGBA support can cause redundant RGBA -> YUV conversion if multiple encoders are used for the same RGBA input * Subsampled planar format support is dropped: I420 and YV12 format are not supported formats for Direct3D11. Although it's supported in CUDA mode, it's not a hardware friendly memory layout and it will waste GPU memory since UV planes will have large padding due to the memory layout requirement of NVENC. * GL support is dropped: Similar to the RGBA case, GL support in encoder would be suboptimal if GL input is used by multiple encoders, because each encoder will copy GL memory into CUDA memory. Upstream cudaupload element can be used for GL <-> CUDA interop instead. * No more pre-allocation of encoder input surfaces. New implementation will use input CUDA memory without copy (zero-copy) or will copy into a NVENC's input buffer struct in case of system memory input. Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1997>
2022-03-15 19:58:16 +00:00
struct _GstNvEncoder
{
GstVideoEncoder parent;
GstNvEncoderPrivate *priv;
};
struct _GstNvEncoderClass
{
GstVideoEncoderClass parent_class;
gboolean (*set_format) (GstNvEncoder * encoder,
GstVideoCodecState * state,
gpointer session,
NV_ENC_INITIALIZE_PARAMS * init_params,
NV_ENC_CONFIG * config);
gboolean (*set_output_state) (GstNvEncoder * encoder,
GstVideoCodecState * state,
gpointer session);
GstBuffer * (*create_output_buffer) (GstNvEncoder * encoder,
NV_ENC_LOCK_BITSTREAM * bitstream);
GstNvEncoderReconfigure (*check_reconfigure) (GstNvEncoder * encoder,
NV_ENC_CONFIG * config);
gboolean (*select_device) (GstNvEncoder * encoder,
const GstVideoInfo * info,
GstBuffer * buffer,
GstNvEncoderDeviceData * data);
guint (*calculate_min_buffers) (GstNvEncoder * encoder);
nvcodec: Add new Direct3D11/CUDA mode encoder implementation Adding new encoder elements nvd3d11{h264,h265}enc for Direct3D11 input support and re-written nvcuda{h264,h265}enc elements. Newly writeen elements have some differences compared with old nv{h264,h265}enc including non-backward compatible changes. * RGBA is not a supported input format any more: New elements will support only YUV formats to avoid implicit conversion done by hardware. Ideally it should be done by upstream element in order to have more control on it. Moreover, RGBA support can cause redundant RGBA -> YUV conversion if multiple encoders are used for the same RGBA input * Subsampled planar format support is dropped: I420 and YV12 format are not supported formats for Direct3D11. Although it's supported in CUDA mode, it's not a hardware friendly memory layout and it will waste GPU memory since UV planes will have large padding due to the memory layout requirement of NVENC. * GL support is dropped: Similar to the RGBA case, GL support in encoder would be suboptimal if GL input is used by multiple encoders, because each encoder will copy GL memory into CUDA memory. Upstream cudaupload element can be used for GL <-> CUDA interop instead. * No more pre-allocation of encoder input surfaces. New implementation will use input CUDA memory without copy (zero-copy) or will copy into a NVENC's input buffer struct in case of system memory input. Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1997>
2022-03-15 19:58:16 +00:00
};
GType gst_nv_encoder_get_type (void);
void gst_nv_encoder_preset_to_native (GstNvEncoderPreset preset,
GstNvEncoderTune tune,
GUID * preset_guid,
NV_ENC_TUNING_INFO * tune_info);
void gst_nv_encoder_rc_mode_to_native (GstNvEncoderRCMode rc_mode,
GstNvEncoderMultiPass multipass,
NV_ENC_PARAMS_RC_MODE * rc_mode_native,
NV_ENC_MULTI_PASS * multipass_native);
nvcodec: Add new Direct3D11/CUDA mode encoder implementation Adding new encoder elements nvd3d11{h264,h265}enc for Direct3D11 input support and re-written nvcuda{h264,h265}enc elements. Newly writeen elements have some differences compared with old nv{h264,h265}enc including non-backward compatible changes. * RGBA is not a supported input format any more: New elements will support only YUV formats to avoid implicit conversion done by hardware. Ideally it should be done by upstream element in order to have more control on it. Moreover, RGBA support can cause redundant RGBA -> YUV conversion if multiple encoders are used for the same RGBA input * Subsampled planar format support is dropped: I420 and YV12 format are not supported formats for Direct3D11. Although it's supported in CUDA mode, it's not a hardware friendly memory layout and it will waste GPU memory since UV planes will have large padding due to the memory layout requirement of NVENC. * GL support is dropped: Similar to the RGBA case, GL support in encoder would be suboptimal if GL input is used by multiple encoders, because each encoder will copy GL memory into CUDA memory. Upstream cudaupload element can be used for GL <-> CUDA interop instead. * No more pre-allocation of encoder input surfaces. New implementation will use input CUDA memory without copy (zero-copy) or will copy into a NVENC's input buffer struct in case of system memory input. Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1997>
2022-03-15 19:58:16 +00:00
void gst_nv_encoder_set_device_mode (GstNvEncoder * encoder,
GstNvEncoderDeviceMode mode,
guint cuda_device_id,
gint64 adapter_luid);
GstNvEncoderClassData * gst_nv_encoder_class_data_new (void);
GstNvEncoderClassData * gst_nv_encoder_class_data_ref (GstNvEncoderClassData * cdata);
void gst_nv_encoder_class_data_unref (GstNvEncoderClassData * cdata);
void gst_nv_encoder_get_encoder_caps (gpointer session,
const GUID * encode_guid,
GstNvEncoderDeviceCaps * device_caps);
void gst_nv_encoder_merge_device_caps (const GstNvEncoderDeviceCaps * a,
const GstNvEncoderDeviceCaps * b,
GstNvEncoderDeviceCaps * merged);
nvcodec: Add new Direct3D11/CUDA mode encoder implementation Adding new encoder elements nvd3d11{h264,h265}enc for Direct3D11 input support and re-written nvcuda{h264,h265}enc elements. Newly writeen elements have some differences compared with old nv{h264,h265}enc including non-backward compatible changes. * RGBA is not a supported input format any more: New elements will support only YUV formats to avoid implicit conversion done by hardware. Ideally it should be done by upstream element in order to have more control on it. Moreover, RGBA support can cause redundant RGBA -> YUV conversion if multiple encoders are used for the same RGBA input * Subsampled planar format support is dropped: I420 and YV12 format are not supported formats for Direct3D11. Although it's supported in CUDA mode, it's not a hardware friendly memory layout and it will waste GPU memory since UV planes will have large padding due to the memory layout requirement of NVENC. * GL support is dropped: Similar to the RGBA case, GL support in encoder would be suboptimal if GL input is used by multiple encoders, because each encoder will copy GL memory into CUDA memory. Upstream cudaupload element can be used for GL <-> CUDA interop instead. * No more pre-allocation of encoder input surfaces. New implementation will use input CUDA memory without copy (zero-copy) or will copy into a NVENC's input buffer struct in case of system memory input. Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1997>
2022-03-15 19:58:16 +00:00
gboolean _gst_nv_enc_result (NVENCSTATUS status,
GObject * self,
const gchar * file,
const gchar * function,
gint line);
#define gst_nv_enc_result(status,self) \
_gst_nv_enc_result (status, (GObject *) self, __FILE__, GST_FUNCTION, __LINE__)
nvcodec: Add new Direct3D11/CUDA mode encoder implementation Adding new encoder elements nvd3d11{h264,h265}enc for Direct3D11 input support and re-written nvcuda{h264,h265}enc elements. Newly writeen elements have some differences compared with old nv{h264,h265}enc including non-backward compatible changes. * RGBA is not a supported input format any more: New elements will support only YUV formats to avoid implicit conversion done by hardware. Ideally it should be done by upstream element in order to have more control on it. Moreover, RGBA support can cause redundant RGBA -> YUV conversion if multiple encoders are used for the same RGBA input * Subsampled planar format support is dropped: I420 and YV12 format are not supported formats for Direct3D11. Although it's supported in CUDA mode, it's not a hardware friendly memory layout and it will waste GPU memory since UV planes will have large padding due to the memory layout requirement of NVENC. * GL support is dropped: Similar to the RGBA case, GL support in encoder would be suboptimal if GL input is used by multiple encoders, because each encoder will copy GL memory into CUDA memory. Upstream cudaupload element can be used for GL <-> CUDA interop instead. * No more pre-allocation of encoder input surfaces. New implementation will use input CUDA memory without copy (zero-copy) or will copy into a NVENC's input buffer struct in case of system memory input. Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1997>
2022-03-15 19:58:16 +00:00
G_DEFINE_AUTOPTR_CLEANUP_FUNC(GstNvEncoder, gst_object_unref)
G_END_DECLS
#ifdef __cplusplus
#ifndef G_OS_WIN32
inline bool is_equal_guid(const GUID & lhs, const GUID & rhs)
{
return memcmp(&lhs, &rhs, sizeof (GUID)) == 0;
nvcodec: Add new Direct3D11/CUDA mode encoder implementation Adding new encoder elements nvd3d11{h264,h265}enc for Direct3D11 input support and re-written nvcuda{h264,h265}enc elements. Newly writeen elements have some differences compared with old nv{h264,h265}enc including non-backward compatible changes. * RGBA is not a supported input format any more: New elements will support only YUV formats to avoid implicit conversion done by hardware. Ideally it should be done by upstream element in order to have more control on it. Moreover, RGBA support can cause redundant RGBA -> YUV conversion if multiple encoders are used for the same RGBA input * Subsampled planar format support is dropped: I420 and YV12 format are not supported formats for Direct3D11. Although it's supported in CUDA mode, it's not a hardware friendly memory layout and it will waste GPU memory since UV planes will have large padding due to the memory layout requirement of NVENC. * GL support is dropped: Similar to the RGBA case, GL support in encoder would be suboptimal if GL input is used by multiple encoders, because each encoder will copy GL memory into CUDA memory. Upstream cudaupload element can be used for GL <-> CUDA interop instead. * No more pre-allocation of encoder input surfaces. New implementation will use input CUDA memory without copy (zero-copy) or will copy into a NVENC's input buffer struct in case of system memory input. Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1997>
2022-03-15 19:58:16 +00:00
}
inline bool operator==(const GUID & lhs, const GUID & rhs)
{
return is_equal_guid(lhs, rhs);
}
inline bool operator!=(const GUID & lhs, const GUID & rhs)
{
return !is_equal_guid(lhs, rhs);
nvcodec: Add new Direct3D11/CUDA mode encoder implementation Adding new encoder elements nvd3d11{h264,h265}enc for Direct3D11 input support and re-written nvcuda{h264,h265}enc elements. Newly writeen elements have some differences compared with old nv{h264,h265}enc including non-backward compatible changes. * RGBA is not a supported input format any more: New elements will support only YUV formats to avoid implicit conversion done by hardware. Ideally it should be done by upstream element in order to have more control on it. Moreover, RGBA support can cause redundant RGBA -> YUV conversion if multiple encoders are used for the same RGBA input * Subsampled planar format support is dropped: I420 and YV12 format are not supported formats for Direct3D11. Although it's supported in CUDA mode, it's not a hardware friendly memory layout and it will waste GPU memory since UV planes will have large padding due to the memory layout requirement of NVENC. * GL support is dropped: Similar to the RGBA case, GL support in encoder would be suboptimal if GL input is used by multiple encoders, because each encoder will copy GL memory into CUDA memory. Upstream cudaupload element can be used for GL <-> CUDA interop instead. * No more pre-allocation of encoder input surfaces. New implementation will use input CUDA memory without copy (zero-copy) or will copy into a NVENC's input buffer struct in case of system memory input. Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1997>
2022-03-15 19:58:16 +00:00
}
#endif /* G_OS_WIN32 */
#endif /* __cplusplus */