diff mbox series

[FFmpeg-devel,4/4,v2] avcodec/h264dec: add support for LCEVC enhancement

Message ID 20240721225350.298-4-jamrial@gmail.com
State New
Headers show
Series [FFmpeg-devel,1/4,v2] avutil/frame: add an LCEVC enhancement data payload side data type | expand

Checks

Context Check Description
andriy/make_x86 success Make finished
andriy/make_fate_x86 success Make fate finished

Commit Message

James Almer July 21, 2024, 10:53 p.m. UTC
Signed-off-by: James Almer <jamrial@gmail.com>
---
 configure                 |  2 +-
 libavcodec/avcodec.h      |  5 +++++
 libavcodec/h264_picture.c |  1 +
 libavcodec/h264_slice.c   | 15 ++++++++++++++
 libavcodec/h264dec.c      | 41 ++++++++++++++++++++++++++++++++++++---
 libavcodec/h264dec.h      |  3 +++
 6 files changed, 63 insertions(+), 4 deletions(-)

Comments

Anton Khirnov July 26, 2024, 7:55 a.m. UTC | #1
Quoting James Almer (2024-07-22 00:53:50)
> Signed-off-by: James Almer <jamrial@gmail.com>
> ---
>  configure                 |  2 +-
>  libavcodec/avcodec.h      |  5 +++++
>  libavcodec/h264_picture.c |  1 +
>  libavcodec/h264_slice.c   | 15 ++++++++++++++
>  libavcodec/h264dec.c      | 41 ++++++++++++++++++++++++++++++++++++---
>  libavcodec/h264dec.h      |  3 +++
>  6 files changed, 63 insertions(+), 4 deletions(-)

Does this need to be integrated into individual decoders? Can't the
generic path apply it based on the presence of AV_FRAME_DATA_LCEVC sdide
data and AV_CODEC_EXPORT_DATA_ENHANCEMENTS being set?
James Almer July 26, 2024, 12:10 p.m. UTC | #2
On 7/26/2024 4:55 AM, Anton Khirnov wrote:
> Quoting James Almer (2024-07-22 00:53:50)
>> Signed-off-by: James Almer <jamrial@gmail.com>
>> ---
>>   configure                 |  2 +-
>>   libavcodec/avcodec.h      |  5 +++++
>>   libavcodec/h264_picture.c |  1 +
>>   libavcodec/h264_slice.c   | 15 ++++++++++++++
>>   libavcodec/h264dec.c      | 41 ++++++++++++++++++++++++++++++++++++---
>>   libavcodec/h264dec.h      |  3 +++
>>   6 files changed, 63 insertions(+), 4 deletions(-)
> 
> Does this need to be integrated into individual decoders? Can't the
> generic path apply it based on the presence of AV_FRAME_DATA_LCEVC sdide
> data and AV_CODEC_EXPORT_DATA_ENHANCEMENTS being set?

Individual decoders still need to be aware of lcevc metadata being 
present to do things like allocating the output frame, because afaik i 
can't call get_buffer2() from the main/user thread. So is it worth 
moving things to the generic code when DR1 decoders will still need 
extra considerations?
Anton Khirnov July 26, 2024, 12:19 p.m. UTC | #3
Quoting James Almer (2024-07-26 14:10:46)
> On 7/26/2024 4:55 AM, Anton Khirnov wrote:
> > Quoting James Almer (2024-07-22 00:53:50)
> >> Signed-off-by: James Almer <jamrial@gmail.com>
> >> ---
> >>   configure                 |  2 +-
> >>   libavcodec/avcodec.h      |  5 +++++
> >>   libavcodec/h264_picture.c |  1 +
> >>   libavcodec/h264_slice.c   | 15 ++++++++++++++
> >>   libavcodec/h264dec.c      | 41 ++++++++++++++++++++++++++++++++++++---
> >>   libavcodec/h264dec.h      |  3 +++
> >>   6 files changed, 63 insertions(+), 4 deletions(-)
> > 
> > Does this need to be integrated into individual decoders? Can't the
> > generic path apply it based on the presence of AV_FRAME_DATA_LCEVC sdide
> > data and AV_CODEC_EXPORT_DATA_ENHANCEMENTS being set?
> 
> Individual decoders still need to be aware of lcevc metadata being 
> present to do things like allocating the output frame, because afaik i 
> can't call get_buffer2() from the main/user thread. So is it worth 
> moving things to the generic code when DR1 decoders will still need 
> extra considerations?

Can it not be handled in ff_get_buffer()? And even if not - there'd
still be a lot less code that needs to be duplicated across every
decoder supporting this.
James Almer July 26, 2024, 12:21 p.m. UTC | #4
On 7/26/2024 9:19 AM, Anton Khirnov wrote:
> Quoting James Almer (2024-07-26 14:10:46)
>> On 7/26/2024 4:55 AM, Anton Khirnov wrote:
>>> Quoting James Almer (2024-07-22 00:53:50)
>>>> Signed-off-by: James Almer <jamrial@gmail.com>
>>>> ---
>>>>    configure                 |  2 +-
>>>>    libavcodec/avcodec.h      |  5 +++++
>>>>    libavcodec/h264_picture.c |  1 +
>>>>    libavcodec/h264_slice.c   | 15 ++++++++++++++
>>>>    libavcodec/h264dec.c      | 41 ++++++++++++++++++++++++++++++++++++---
>>>>    libavcodec/h264dec.h      |  3 +++
>>>>    6 files changed, 63 insertions(+), 4 deletions(-)
>>>
>>> Does this need to be integrated into individual decoders? Can't the
>>> generic path apply it based on the presence of AV_FRAME_DATA_LCEVC sdide
>>> data and AV_CODEC_EXPORT_DATA_ENHANCEMENTS being set?
>>
>> Individual decoders still need to be aware of lcevc metadata being
>> present to do things like allocating the output frame, because afaik i
>> can't call get_buffer2() from the main/user thread. So is it worth
>> moving things to the generic code when DR1 decoders will still need
>> extra considerations?
> 
> Can it not be handled in ff_get_buffer()? And even if not - there'd
> still be a lot less code that needs to be duplicated across every
> decoder supporting this.

I'll try that.
diff mbox series

Patch

diff --git a/configure b/configure
index d1f32684a6..7bc8c3b0af 100755
--- a/configure
+++ b/configure
@@ -2968,7 +2968,7 @@  h263i_decoder_select="h263_decoder"
 h263p_decoder_select="h263_decoder"
 h263p_encoder_select="h263_encoder"
 h264_decoder_select="cabac golomb h264chroma h264dsp h264parse h264pred h264qpel h264_sei videodsp"
-h264_decoder_suggest="error_resilience"
+h264_decoder_suggest="error_resilience liblcevc_dec"
 hap_decoder_select="snappy texturedsp"
 hap_encoder_deps="libsnappy"
 hap_encoder_select="texturedspenc"
diff --git a/libavcodec/avcodec.h b/libavcodec/avcodec.h
index 2da63c87ea..6fc46207db 100644
--- a/libavcodec/avcodec.h
+++ b/libavcodec/avcodec.h
@@ -418,6 +418,11 @@  typedef struct RcOverride{
  * Do not apply film grain, export it instead.
  */
 #define AV_CODEC_EXPORT_DATA_FILM_GRAIN (1 << 3)
+/**
+ * Decoding only.
+ * Do not apply picture enhancement layers, export them instead.
+ */
+#define AV_CODEC_EXPORT_DATA_ENHANCEMENTS (1 << 4)
 
 /**
  * The decoder will keep a reference to the frame and may reuse it later.
diff --git a/libavcodec/h264_picture.c b/libavcodec/h264_picture.c
index 3234141dbd..1cdb926d15 100644
--- a/libavcodec/h264_picture.c
+++ b/libavcodec/h264_picture.c
@@ -103,6 +103,7 @@  static void h264_copy_picture_params(H264Picture *dst, const H264Picture *src)
     dst->mb_height     = src->mb_height;
     dst->mb_stride     = src->mb_stride;
     dst->needs_fg      = src->needs_fg;
+    dst->needs_lcevc   = src->needs_lcevc;
 }
 
 int ff_h264_ref_picture(H264Picture *dst, const H264Picture *src)
diff --git a/libavcodec/h264_slice.c b/libavcodec/h264_slice.c
index a66b75ca80..60b6fa4624 100644
--- a/libavcodec/h264_slice.c
+++ b/libavcodec/h264_slice.c
@@ -191,6 +191,11 @@  static int alloc_picture(H264Context *h, H264Picture *pic)
 
     av_assert0(!pic->f->data[0]);
 
+    if (pic->needs_lcevc) {
+        pic->f->width  = FFMAX(h->avctx->width,  h->avctx->coded_width)  * 2 / FFMAX(h->avctx->sample_aspect_ratio.den, 1);
+        pic->f->height = FFMAX(h->avctx->height, h->avctx->coded_height) * 2 / FFMAX(h->avctx->sample_aspect_ratio.num, 1);
+    }
+
     pic->tf.f = pic->f;
     ret = ff_thread_get_ext_buffer(h->avctx, &pic->tf,
                                    pic->reference ? AV_GET_BUFFER_FLAG_REF : 0);
@@ -206,6 +211,11 @@  static int alloc_picture(H264Context *h, H264Picture *pic)
             goto fail;
     }
 
+    if (pic->needs_lcevc) {
+        pic->f->width  = FFMAX(h->avctx->width,  h->avctx->coded_width);
+        pic->f->height = FFMAX(h->avctx->height, h->avctx->coded_height);
+    }
+
     ret = ff_hwaccel_frame_priv_alloc(h->avctx, &pic->hwaccel_picture_private);
     if (ret < 0)
         goto fail;
@@ -459,6 +469,8 @@  int ff_h264_update_thread_context(AVCodecContext *dst,
     h->recovery_frame        = h1->recovery_frame;
     h->non_gray              = h1->non_gray;
 
+    av_buffer_replace(&h->lcevc, h1->lcevc);
+
     return err;
 }
 
@@ -519,6 +531,9 @@  static int h264_frame_start(H264Context *h)
     pic->needs_fg = h->sei.common.film_grain_characteristics.present && !h->avctx->hwaccel &&
         !(h->avctx->export_side_data & AV_CODEC_EXPORT_DATA_FILM_GRAIN);
 
+    pic->needs_lcevc = CONFIG_LIBLCEVC_DEC && h->sei.common.lcevc.info && !h->avctx->hwaccel &&
+        !(h->avctx->export_side_data & AV_CODEC_EXPORT_DATA_ENHANCEMENTS);
+
     if ((ret = alloc_picture(h, pic)) < 0)
         return ret;
 
diff --git a/libavcodec/h264dec.c b/libavcodec/h264dec.c
index c77d8f42db..1c3a84f081 100644
--- a/libavcodec/h264dec.c
+++ b/libavcodec/h264dec.c
@@ -38,6 +38,7 @@ 
 #include "libavutil/video_enc_params.h"
 
 #include "codec_internal.h"
+#include "decode.h"
 #include "internal.h"
 #include "error_resilience.h"
 #include "avcodec.h"
@@ -49,6 +50,7 @@ 
 #include "golomb.h"
 #include "hwaccel_internal.h"
 #include "hwconfig.h"
+#include "lcevcdec.h"
 #include "mpegutils.h"
 #include "profiles.h"
 #include "rectangle.h"
@@ -377,6 +379,8 @@  static av_cold int h264_decode_end(AVCodecContext *avctx)
     h264_free_pic(h, &h->cur_pic);
     h264_free_pic(h, &h->last_pic_for_ec);
 
+    av_buffer_unref(&h->lcevc);
+
     return 0;
 }
 
@@ -404,6 +408,9 @@  FF_ENABLE_DEPRECATION_WARNINGS
 #endif
 
     if (!avctx->internal->is_copy) {
+#if CONFIG_LIBLCEVC_DEC
+        FFLCEVCContext *lcevc;
+#endif
         if (avctx->extradata_size > 0 && avctx->extradata) {
             ret = ff_h264_decode_extradata(avctx->extradata, avctx->extradata_size,
                                            &h->ps, &h->is_avc, &h->nal_length_size,
@@ -418,6 +425,21 @@  FF_ENABLE_DEPRECATION_WARNINGS
                ret = 0;
            }
         }
+#if CONFIG_LIBLCEVC_DEC
+        lcevc = av_mallocz(sizeof(FFLCEVCContext));
+        ret = ff_lcevc_init(lcevc, avctx);
+        if (ret < 0) {
+            int explode = avctx->err_recognition & AV_EF_EXPLODE;
+            av_log(avctx, explode ? AV_LOG_ERROR: AV_LOG_WARNING,
+                   "Error initializing LCEVC\n");
+            if (explode) {
+                av_free(lcevc);
+                return ret;
+            }
+            ret = 0;
+        }
+        h->lcevc = av_buffer_create(lcevc, sizeof(FFLCEVCContext), ff_lcevc_free, lcevc, 0);
+#endif
     }
 
     if (h->ps.sps && h->ps.sps->bitstream_restriction_flag &&
@@ -667,9 +689,20 @@  static int decode_nal_units(H264Context *h, const uint8_t *buf, int buf_size)
                     h->setup_finished = 1;
                 }
 
-                if (h->avctx->hwaccel &&
-                    (ret = FF_HW_CALL(h->avctx, start_frame, buf, buf_size)) < 0)
-                    goto end;
+                if (h->avctx->hwaccel) {
+                    ret = FF_HW_CALL(h->avctx, start_frame, buf, buf_size);
+                    if (ret < 0)
+                        goto end;
+                } else if (CONFIG_LIBLCEVC_DEC && h->cur_pic_ptr->needs_lcevc) {
+                    FrameDecodeData *fdd = (FrameDecodeData*)h->cur_pic_ptr->f->private_ref->data;
+                    fdd->post_process_opaque = av_buffer_ref(h->lcevc);
+                    if (!fdd->post_process_opaque) {
+                        ret = -1;
+                        goto end;
+                    }
+                    fdd->post_process_opaque_free = ff_lcevc_unref;
+                    fdd->post_process = ff_lcevc_process;
+                }
             }
 
             max_slice_ctx = avctx->hwaccel ? 1 : h->nb_slice_ctx;
@@ -904,6 +937,8 @@  static int output_frame(H264Context *h, AVFrame *dst, H264Picture *srcp)
 
     if (!(h->avctx->export_side_data & AV_CODEC_EXPORT_DATA_FILM_GRAIN))
         av_frame_remove_side_data(dst, AV_FRAME_DATA_FILM_GRAIN_PARAMS);
+    if (!(h->avctx->export_side_data & AV_CODEC_EXPORT_DATA_ENHANCEMENTS))
+        av_frame_remove_side_data(dst, AV_FRAME_DATA_LCEVC);
 
     return 0;
 fail:
diff --git a/libavcodec/h264dec.h b/libavcodec/h264dec.h
index ccd7583bf4..419d699bab 100644
--- a/libavcodec/h264dec.h
+++ b/libavcodec/h264dec.h
@@ -154,6 +154,7 @@  typedef struct H264Picture {
     int invalid_gap;
     int sei_recovery_frame_cnt;
     int needs_fg;           ///< whether picture needs film grain synthesis (see `f_grain`)
+    int needs_lcevc;        ///< whether picture needs LCEVC enhancement
 
     const PPS   *pps;
 
@@ -580,6 +581,8 @@  typedef struct H264Context {
     int non_gray;                       ///< Did we encounter a intra frame after a gray gap frame
     int noref_gray;
     int skip_gray;
+
+    AVBufferRef *lcevc; // FFLCEVCContext
 } H264Context;
 
 extern const uint16_t ff_h264_mb_sizes[4];