king
/
v811_spc009
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from 'main'
${ noResults }
v811_spc009/external/libhevc/encoder/ihevce_coarse_me_pass.c

1474 lines
54 KiB
Raw Permalink Blame History

/******************************************************************************
*
* Copyright (C) 2018 The Android Open Source Project
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at:
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*
*****************************************************************************
* Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore
*/
/*!
******************************************************************************
* \file ihevce_coarse_me_pass.c
*
* \brief
* Converts the language of the encoder to language of me. This is an i/f
* between the encoder style APIs and ME style APIs. This is basically
* a memoryless glue layer.
*
* \date
* 22/10/2012
*
* \author
* Ittiam
*
*
* List of Functions
*
*
******************************************************************************
*/
/*****************************************************************************/
/* File Includes */
/*****************************************************************************/
/* System include files */
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <assert.h>
#include <stdarg.h>
#include <math.h>
/* User include files */
#include "ihevc_typedefs.h"
#include "itt_video_api.h"
#include "ihevce_api.h"
#include "rc_cntrl_param.h"
#include "rc_frame_info_collector.h"
#include "rc_look_ahead_params.h"
#include "ihevc_defs.h"
#include "ihevc_structs.h"
#include "ihevc_platform_macros.h"
#include "ihevc_deblk.h"
#include "ihevc_itrans_recon.h"
#include "ihevc_chroma_itrans_recon.h"
#include "ihevc_chroma_intra_pred.h"
#include "ihevc_intra_pred.h"
#include "ihevc_inter_pred.h"
#include "ihevc_mem_fns.h"
#include "ihevc_padding.h"
#include "ihevc_weighted_pred.h"
#include "ihevc_sao.h"
#include "ihevc_resi_trans.h"
#include "ihevc_quant_iquant_ssd.h"
#include "ihevc_cabac_tables.h"
#include "ihevce_defs.h"
#include "ihevce_lap_enc_structs.h"
#include "ihevce_multi_thrd_structs.h"
#include "ihevce_me_common_defs.h"
#include "ihevce_had_satd.h"
#include "ihevce_error_codes.h"
#include "ihevce_bitstream.h"
#include "ihevce_cabac.h"
#include "ihevce_rdoq_macros.h"
#include "ihevce_function_selector.h"
#include "ihevce_enc_structs.h"
#include "ihevce_entropy_structs.h"
#include "ihevce_cmn_utils_instr_set_router.h"
#include "ihevce_enc_loop_structs.h"
#include "ihevce_bs_compute_ctb.h"
#include "ihevce_global_tables.h"
#include "ihevce_dep_mngr_interface.h"
#include "hme_datatype.h"
#include "hme_interface.h"
#include "hme_common_defs.h"
#include "hme_defs.h"
#include "ihevce_me_instr_set_router.h"
#include "ihevce_ipe_instr_set_router.h"
#include "ihevce_ipe_structs.h"
#include "hme_globals.h"
#include "hme_utils.h"
#include "hme_coarse.h"
#include "hme_refine.h"
#include "ihevce_me_pass.h"
#include "ihevce_coarse_me_pass.h"
/*****************************************************************************/
/* Function Definitions */
/*****************************************************************************/
/*!
******************************************************************************
* \if Function name : ihevce_coarse_me_get_num_mem_recs \endif
*
* \brief
* Number of memory records are returned for ME module
* Note : Include total mem. req. for HME + Total mem. req. for Dep Mngr for HME
*
* \return
* Number of memory records
*
* \author
* Ittiam
*
*****************************************************************************
*/
WORD32 ihevce_coarse_me_get_num_mem_recs()
{
WORD32 hme_mem_recs = hme_coarse_num_alloc();
WORD32 hme_dep_mngr_mem_recs = hme_coarse_dep_mngr_num_alloc();
return ((hme_mem_recs + hme_dep_mngr_mem_recs));
}
/*!
******************************************************************************
* \if Function name : ihevce_coarse_me_get_mem_recs \endif
*
* \brief
* Memory requirements are returned for coarse ME.
*
* \param[in,out] ps_mem_tab : pointer to memory descriptors table
* \param[in] ps_init_prms : Create time static parameters
* \param[in] i4_num_proc_thrds : Number of processing threads for this module
* \param[in] i4_mem_space : memspace in whihc memory request should be done
*
* \return
* Number of records
*
* \author
* Ittiam
*
*****************************************************************************
*/
WORD32 ihevce_coarse_me_get_mem_recs(
iv_mem_rec_t *ps_mem_tab,
ihevce_static_cfg_params_t *ps_init_prms,
WORD32 i4_num_proc_thrds,
WORD32 i4_mem_space,
WORD32 i4_resolution_id)
{
hme_memtab_t as_memtabs[HME_COARSE_TOT_MEMTABS];
WORD32 n_tabs, i;
/* Init prms structure specific to HME */
hme_init_prms_t s_hme_init_prms;
//return (ihevce_coarse_me_get_num_mem_recs());
/*************************************************************************/
/* code flow: we call hme alloc function and then remap those memtabs */
/* to a different type of memtab structure. */
/*************************************************************************/
ASSERT(HME_COARSE_TOT_MEMTABS >= hme_coarse_num_alloc());
/*************************************************************************/
/* POPULATE THE HME INIT PRMS */
/*************************************************************************/
ihevce_derive_me_init_prms(ps_init_prms, &s_hme_init_prms, i4_num_proc_thrds, i4_resolution_id);
/*************************************************************************/
/* CALL THE ME FUNCTION TO GET MEMTABS */
/*************************************************************************/
n_tabs = hme_coarse_alloc(&as_memtabs[0], &s_hme_init_prms);
ASSERT(n_tabs == hme_coarse_num_alloc());
/*************************************************************************/
/* REMAP RESULTS TO ENCODER MEMTAB STRUCTURE */
/*************************************************************************/
for(i = 0; i < n_tabs; i++)
{
ps_mem_tab[i].i4_mem_size = as_memtabs[i].size;
ps_mem_tab[i].i4_mem_alignment = as_memtabs[i].align;
ps_mem_tab[i].e_mem_type = (IV_MEM_TYPE_T)i4_mem_space;
ps_mem_tab[i].i4_size = sizeof(iv_mem_rec_t);
}
/*************************************************************************/
/* --- HME Coarse sync Dep Mngr Mem requests -- */
/*************************************************************************/
{
WORD32 n_dep_tabs;
ps_mem_tab += n_tabs;
n_dep_tabs = hme_coarse_dep_mngr_alloc(
ps_mem_tab, ps_init_prms, i4_mem_space, i4_num_proc_thrds, i4_resolution_id);
ASSERT(n_dep_tabs == hme_coarse_dep_mngr_num_alloc());
/* Update the total no. of mem tabs */
n_tabs += n_dep_tabs;
}
return (n_tabs);
}
/*!
******************************************************************************
* \if Function name : ihevce_coarse_me_init \endif
*
* \brief
* Intialization for ME context state structure .
*
* \param[in] ps_mem_tab : pointer to memory descriptors table
* \param[in] ps_init_prms : Create time static parameters
* \param[in] pv_osal_handle : Osal handle
*
* \return
* Handle to the ME context
*
* \author
* Ittiam
*
*****************************************************************************
*/
void *ihevce_coarse_me_init(
iv_mem_rec_t *ps_mem_tab,
ihevce_static_cfg_params_t *ps_init_prms,
WORD32 i4_num_proc_thrds,
void *pv_osal_handle,
WORD32 i4_resolution_id,
UWORD8 u1_is_popcnt_available)
{
/* ME handle to be returned */
void *pv_me_ctxt;
WORD32 status;
coarse_me_master_ctxt_t *ps_ctxt;
/* Init prms structure specific to HME */
hme_init_prms_t s_hme_init_prms;
/* memtabs to be passed to hme */
hme_memtab_t as_memtabs[HME_COARSE_TOT_MEMTABS];
WORD32 n_tabs, n_dep_tabs, i;
/*************************************************************************/
/* POPULATE THE HME INIT PRMS */
/*************************************************************************/
ihevce_derive_me_init_prms(ps_init_prms, &s_hme_init_prms, i4_num_proc_thrds, i4_resolution_id);
/*************************************************************************/
/* Ensure local declaration is sufficient */
/*************************************************************************/
n_tabs = hme_coarse_num_alloc();
ASSERT(HME_COARSE_TOT_MEMTABS >= n_tabs);
/*************************************************************************/
/* MAP RESULTS TO HME MEMTAB STRUCTURE */
/*************************************************************************/
for(i = 0; i < n_tabs; i++)
{
as_memtabs[i].size = ps_mem_tab[i].i4_mem_size;
as_memtabs[i].align = ps_mem_tab[i].i4_mem_alignment;
as_memtabs[i].pu1_mem = (U08 *)ps_mem_tab[i].pv_base;
}
/*************************************************************************/
/* CALL THE ME FUNCTION TO GET MEMTABS */
/*************************************************************************/
pv_me_ctxt = (void *)as_memtabs[0].pu1_mem;
status = hme_coarse_init(pv_me_ctxt, &as_memtabs[0], &s_hme_init_prms);
ps_ctxt = (coarse_me_master_ctxt_t *)pv_me_ctxt;
if(status == -1)
return NULL;
/*************************************************************************/
/* --- HME sync Dep Mngr Mem init -- */
/*************************************************************************/
ps_mem_tab += n_tabs;
n_dep_tabs = hme_coarse_dep_mngr_init(
ps_mem_tab, ps_init_prms, pv_me_ctxt, pv_osal_handle, i4_num_proc_thrds, i4_resolution_id);
ASSERT(n_dep_tabs <= hme_coarse_dep_mngr_num_alloc());
n_tabs += n_dep_tabs;
ihevce_me_instr_set_router(
(ihevce_me_optimised_function_list_t *)ps_ctxt->pv_me_optimised_function_list,
ps_init_prms->e_arch_type);
ihevce_cmn_utils_instr_set_router(
&ps_ctxt->s_cmn_opt_func, u1_is_popcnt_available, ps_init_prms->e_arch_type);
return (pv_me_ctxt);
}
/*!
******************************************************************************
* \if Function name : ihevce_coarse_me_reg_thrds_sem \endif
*
* \brief
* Intialization for ME context state structure with semaphores .
*
* \param[in] pv_me_ctxt : pointer to Coarse ME ctxt
* \param[in] ppv_sem_hdls : Array of semaphore handles
* \param[in] i4_num_proc_thrds : Number of processing threads
*
* \return
* none
*
* \author
* Ittiam
*
*****************************************************************************
*/
void ihevce_coarse_me_reg_thrds_sem(void *pv_me_ctxt, void **ppv_sem_hdls, WORD32 i4_num_proc_thrds)
{
hme_coarse_dep_mngr_reg_sem(pv_me_ctxt, ppv_sem_hdls, i4_num_proc_thrds);
return;
}
/*!
******************************************************************************
* \if Function name : ihevce_coarse_me_delete \endif
*
* \brief
* Destroy Coarse ME module
* Note : Only Destroys the resources allocated in the module like
* semaphore,etc. Memory free is done Separately using memtabs
*
* \param[in] pv_me_ctxt : pointer to Coarse ME ctxt
* \param[in] ps_init_prms : Create time static parameters
* \param[in] pv_osal_handle : Osal handle
*
* \return
* None
*
* \author
* Ittiam
*
*****************************************************************************
*/
void ihevce_coarse_me_delete(
void *pv_me_ctxt, ihevce_static_cfg_params_t *ps_init_prms, WORD32 i4_resolution_id)
{
/* --- HME sync Dep Mngr Delete --*/
hme_coarse_dep_mngr_delete(pv_me_ctxt, ps_init_prms, i4_resolution_id);
}
/**
*******************************************************************************
* \if Function name : ihevce_coarse_me_set_resolution \endif
*
* \brief
* Sets the resolution for ME state
*
* \par Description:
* ME requires information of resolution to prime up its layer descriptors
* and contexts. This API is called whenever a control call from application
* causes a change of resolution. Has to be called once initially before
* processing any frame. Again this is just a glue function and calls the
* actual ME API for the same.
*
* \param[in,out] pv_me_ctxt: Handle to the ME context
* \param[in] n_enc_layers: Number of layers getting encoded
* \param[in] p_wd : Pointer containing widths of each layer getting encoded.
* \param[in] p_ht : Pointer containing heights of each layer getting encoded.
*
* \returns
* none
*
* \author
* Ittiam
*
*******************************************************************************
*/
void ihevce_coarse_me_set_resolution(
void *pv_me_ctxt, WORD32 n_enc_layers, WORD32 *p_wd, WORD32 *p_ht)
{
/* local variables */
coarse_me_master_ctxt_t *ps_master_ctxt = (coarse_me_master_ctxt_t *)pv_me_ctxt;
WORD32 thrds;
for(thrds = 0; thrds < ps_master_ctxt->i4_num_proc_thrds; thrds++)
{
coarse_me_ctxt_t *ps_me_thrd_ctxt;
ps_me_thrd_ctxt = ps_master_ctxt->aps_me_ctxt[thrds];
hme_coarse_set_resolution((void *)ps_me_thrd_ctxt, n_enc_layers, p_wd, p_ht);
}
}
void ihevce_coarse_me_get_rc_param(
void *pv_me_ctxt,
LWORD64 *i8_acc_frame_hme_cost,
LWORD64 *i8_acc_frame_hme_sad,
LWORD64 *i8_acc_num_blks_higher_sad,
LWORD64 *i8_total_blks,
WORD32 i4_is_prev_pic_same_scene)
{
coarse_me_master_ctxt_t *ps_master_ctxt = (coarse_me_master_ctxt_t *)pv_me_ctxt;
WORD32 thrds;
coarse_me_ctxt_t *ps_me_thrd_ctxt;
*i8_acc_frame_hme_cost = 0;
*i8_acc_frame_hme_sad = 0;
for(thrds = 0; thrds < ps_master_ctxt->i4_num_proc_thrds; thrds++)
{
ps_me_thrd_ctxt = ps_master_ctxt->aps_me_ctxt[thrds];
*i8_acc_frame_hme_cost += ps_me_thrd_ctxt->i4_L1_hme_best_cost;
/*Calculate me cost wrt. to ref only for P frame */
if(ps_me_thrd_ctxt->s_frm_prms.is_i_pic == ps_me_thrd_ctxt->s_frm_prms.bidir_enabled)
{
*i8_acc_num_blks_higher_sad += ps_me_thrd_ctxt->i4_num_blks_high_sad;
*i8_total_blks += ps_me_thrd_ctxt->i4_num_blks;
}
*i8_acc_frame_hme_sad += ps_me_thrd_ctxt->i4_L1_hme_sad;
}
}
/*!
******************************************************************************
* \if Function name : ihevce_coarse_me_process \endif
*
* \brief
* Frame level ME function
*
* \par Description:
* Processing of all layers starting from coarse and going
* to the refinement layers, except enocde layer
*
* \param[in] pv_ctxt : pointer to ME module
* \param[in] ps_enc_lap_inp : pointer to input yuv buffer (frame buffer)
* \param[in,out] ps_ctb_out : pointer to CTB analyse output structure (frame buffer)
* \param[out] ps_cu_out : pointer to CU analyse output structure (frame buffer)
* \param[in] pd_intra_costs : pointerto intra cost buffer
* \param[in] ps_multi_thrd_ctxt : pointer to multi thread ctxt
* \param[in] thrd_id : Thread id of the current thrd in which function is executed
*
* \return
* None
*
* \author
* Ittiam
*
*****************************************************************************
*/
void ihevce_coarse_me_process(
void *pv_me_ctxt,
ihevce_lap_enc_buf_t *ps_enc_lap_inp,
multi_thrd_ctxt_t *ps_multi_thrd_ctxt,
WORD32 thrd_id,
WORD32 i4_ping_pong)
{
/* local variables */
coarse_me_master_ctxt_t *ps_master_ctxt = (coarse_me_master_ctxt_t *)pv_me_ctxt;
coarse_me_ctxt_t *ps_thrd_ctxt;
/* get the current thread ctxt pointer */
ps_thrd_ctxt = ps_master_ctxt->aps_me_ctxt[thrd_id];
ps_thrd_ctxt->thrd_id = thrd_id;
/* frame level processing function */
hme_coarse_process_frm(
(void *)ps_thrd_ctxt,
&ps_master_ctxt->s_ref_map,
&ps_master_ctxt->s_frm_prms,
ps_multi_thrd_ctxt,
i4_ping_pong,
&ps_master_ctxt->apv_dep_mngr_hme_sync[0]);
return;
}
/*!
******************************************************************************
* \if Function name : ihevce_coarse_me_frame_end \endif
*
* \brief
* End of frame update function performs
* - GMV collation
* - Dynamic Search Range collation
*
* \param[in] pv_ctxt : pointer to ME module
*
* \return
* None
*
* \author
* Ittiam
*
*****************************************************************************
*/
void ihevce_coarse_me_frame_end(void *pv_me_ctxt)
{
/* local variables */
coarse_me_master_ctxt_t *ps_master_ctxt = (coarse_me_master_ctxt_t *)pv_me_ctxt;
coarse_me_ctxt_t *ps_thrd0_ctxt;
layer_ctxt_t *ps_curr_layer;
WORD32 num_ref, num_thrds, cur_poc;
WORD32 coarse_layer_id;
WORD32 i4_num_ref;
ME_QUALITY_PRESETS_T e_me_quality_preset;
/* GMV collation is done for coarse Layer only */
ps_thrd0_ctxt = ps_master_ctxt->aps_me_ctxt[0];
coarse_layer_id = ps_thrd0_ctxt->num_layers - 1;
ps_curr_layer = ps_thrd0_ctxt->ps_curr_descr->aps_layers[coarse_layer_id];
i4_num_ref = ps_master_ctxt->s_ref_map.i4_num_ref;
e_me_quality_preset = ps_thrd0_ctxt->s_init_prms.s_me_coding_tools.e_me_quality_presets;
/* No processing is required if current pic is I pic */
if(1 == ps_master_ctxt->s_frm_prms.is_i_pic)
{
return;
}
/* use thrd 0 ctxt to collate the GMVs histogram and Dynamic Search Range */
/* across all threads */
for(num_ref = 0; num_ref < i4_num_ref; num_ref++)
{
WORD32 i4_offset, i4_lobe_size, i4_layer_id;
mv_hist_t *ps_hist_thrd0;
dyn_range_prms_t *aps_dyn_range_prms_thrd0[MAX_NUM_LAYERS];
ps_hist_thrd0 = ps_thrd0_ctxt->aps_mv_hist[num_ref];
/* Only for P pic. For P, both are 0, I&B has them mut. exclusive */
if(ps_thrd0_ctxt->s_frm_prms.is_i_pic == ps_thrd0_ctxt->s_frm_prms.bidir_enabled)
{
for(i4_layer_id = coarse_layer_id; i4_layer_id > 0; i4_layer_id--)
{
aps_dyn_range_prms_thrd0[i4_layer_id] =
&ps_thrd0_ctxt->s_coarse_dyn_range_prms.as_dyn_range_prms[i4_layer_id][num_ref];
}
}
i4_lobe_size = ps_hist_thrd0->i4_lobe1_size;
i4_offset = i4_lobe_size >> 1;
/* run a loop over all the other threads to add up the histogram */
/* and to update the dynamical search range */
for(num_thrds = 1; num_thrds < ps_master_ctxt->i4_num_proc_thrds; num_thrds++)
{
dyn_range_prms_t *ps_dyn_range_prms;
if(ME_XTREME_SPEED_25 != e_me_quality_preset)
{
mv_hist_t *ps_hist;
WORD32 i4_y, i4_x;
/* get current thrd histogram pointer */
ps_hist = ps_master_ctxt->aps_me_ctxt[num_thrds]->aps_mv_hist[num_ref];
/* Accumalate the Bin count for all the thread */
for(i4_y = 0; i4_y < ps_hist_thrd0->i4_num_rows; i4_y++)
{
for(i4_x = 0; i4_x < ps_hist_thrd0->i4_num_cols; i4_x++)
{
S32 i4_bin_id;
i4_bin_id = i4_x + (i4_y * ps_hist_thrd0->i4_num_cols);
ps_hist_thrd0->ai4_bin_count[i4_bin_id] +=
ps_hist->ai4_bin_count[i4_bin_id];
}
}
}
/* Update the dynamical search range for each Layer */
/* Only for P pic. For P, both are 0, I&B has them mut. exclusive */
if(ps_thrd0_ctxt->s_frm_prms.is_i_pic == ps_thrd0_ctxt->s_frm_prms.bidir_enabled)
{
for(i4_layer_id = coarse_layer_id; i4_layer_id > 0; i4_layer_id--)
{
/* get current thrd, layer dynamical search range param. pointer */
ps_dyn_range_prms =
&ps_master_ctxt->aps_me_ctxt[num_thrds]
->s_coarse_dyn_range_prms.as_dyn_range_prms[i4_layer_id][num_ref];
/* TODO : This calls can be optimized further. No need for min in 1st call and max in 2nd call */
hme_update_dynamic_search_params(
aps_dyn_range_prms_thrd0[i4_layer_id], ps_dyn_range_prms->i2_dyn_max_y);
hme_update_dynamic_search_params(
aps_dyn_range_prms_thrd0[i4_layer_id], ps_dyn_range_prms->i2_dyn_min_y);
}
}
}
}
/*************************************************************************/
/* Get the MAX/MIN per POC distance based on the all the ref. pics */
/*************************************************************************/
/* Only for P pic. For P, both are 0, I&B has them mut. exclusive */
if(ps_thrd0_ctxt->s_frm_prms.is_i_pic == ps_thrd0_ctxt->s_frm_prms.bidir_enabled)
{
WORD32 i4_layer_id;
cur_poc = ps_thrd0_ctxt->i4_curr_poc;
for(i4_layer_id = coarse_layer_id; i4_layer_id > 0; i4_layer_id--)
{
ps_thrd0_ctxt->s_coarse_dyn_range_prms.i2_dyn_max_y_per_poc[i4_layer_id] = 0;
ps_thrd0_ctxt->s_coarse_dyn_range_prms.i2_dyn_min_y_per_poc[i4_layer_id] = 0;
}
for(num_ref = 0; num_ref < i4_num_ref; num_ref++)
{
for(i4_layer_id = coarse_layer_id; i4_layer_id > 0; i4_layer_id--)
{
WORD16 i2_mv_per_poc;
WORD32 ref_poc, poc_diff;
dyn_range_prms_t *ps_dyn_range_prms_thrd0;
ps_dyn_range_prms_thrd0 =
&ps_thrd0_ctxt->s_coarse_dyn_range_prms.as_dyn_range_prms[i4_layer_id][num_ref];
ref_poc = ps_dyn_range_prms_thrd0->i4_poc;
ASSERT(ref_poc < cur_poc);
poc_diff = (cur_poc - ref_poc);
/* cur. ref. pic. max y per POC */
i2_mv_per_poc = (ps_dyn_range_prms_thrd0->i2_dyn_max_y + (poc_diff - 1)) / poc_diff;
/* update the max y per POC */
ps_thrd0_ctxt->s_coarse_dyn_range_prms.i2_dyn_max_y_per_poc[i4_layer_id] =
MAX(ps_thrd0_ctxt->s_coarse_dyn_range_prms.i2_dyn_max_y_per_poc[i4_layer_id],
i2_mv_per_poc);
/* cur. ref. pic. min y per POC */
i2_mv_per_poc = (ps_dyn_range_prms_thrd0->i2_dyn_min_y - (poc_diff - 1)) / poc_diff;
/* update the min y per POC */
ps_thrd0_ctxt->s_coarse_dyn_range_prms.i2_dyn_min_y_per_poc[i4_layer_id] =
MIN(ps_thrd0_ctxt->s_coarse_dyn_range_prms.i2_dyn_min_y_per_poc[i4_layer_id],
i2_mv_per_poc);
}
}
/*************************************************************************/
/* Populate the results to all thread ctxt */
/*************************************************************************/
for(num_thrds = 1; num_thrds < ps_master_ctxt->i4_num_proc_thrds; num_thrds++)
{
for(i4_layer_id = coarse_layer_id; i4_layer_id > 0; i4_layer_id--)
{
ps_master_ctxt->aps_me_ctxt[num_thrds]
->s_coarse_dyn_range_prms.i2_dyn_max_y_per_poc[i4_layer_id] =
ps_thrd0_ctxt->s_coarse_dyn_range_prms.i2_dyn_max_y_per_poc[i4_layer_id];
ps_master_ctxt->aps_me_ctxt[num_thrds]
->s_coarse_dyn_range_prms.i2_dyn_min_y_per_poc[i4_layer_id] =
ps_thrd0_ctxt->s_coarse_dyn_range_prms.i2_dyn_min_y_per_poc[i4_layer_id];
}
}
}
if(ME_XTREME_SPEED_25 != e_me_quality_preset)
{
/* call the function which calcualtes the GMV */
/* layer pointer is shared across all threads */
/* hence all threads will have access to updated */
/* GMVs populated using thread 0 ctxt */
for(num_ref = 0; num_ref < i4_num_ref; num_ref++)
{
hme_calculate_global_mv(
ps_thrd0_ctxt->aps_mv_hist[num_ref],
&ps_curr_layer->s_global_mv[num_ref][GMV_THICK_LOBE],
GMV_THICK_LOBE);
}
}
return;
}
/*!
******************************************************************************
* \if Function name : ihevce_coarse_me_frame_dpb_update \endif
*
* \brief
* Frame level ME initialisation function
*
* \par Description:
* Updation of ME's internal DPB
* based on available ref list information
*
* \param[in] pv_ctxt : pointer to ME module
* \param[in] num_ref_l0 : Number of reference pics in L0 list
* \param[in] num_ref_l1 : Number of reference pics in L1 list
* \param[in] pps_rec_list_l0 : List of recon pics in L0 list
* \param[in] pps_rec_list_l1 : List of recon pics in L1 list
*
* \return
* None
*
* \author
* Ittiam
*
*****************************************************************************
*/
void ihevce_coarse_me_frame_dpb_update(
void *pv_me_ctxt,
WORD32 num_ref_l0,
WORD32 num_ref_l1,
recon_pic_buf_t **pps_rec_list_l0,
recon_pic_buf_t **pps_rec_list_l1)
{
coarse_me_master_ctxt_t *ps_master_ctxt = (coarse_me_master_ctxt_t *)pv_me_ctxt;
coarse_me_ctxt_t *ps_thrd0_ctxt;
WORD32 a_pocs_buffered_in_me[MAX_NUM_REF + 1];
WORD32 a_pocs_to_remove[MAX_NUM_REF + 2];
WORD32 poc_remove_id = 0;
WORD32 i, count;
/* All processing done using shared / common memory across */
/* threads is done using thrd ctxt */
ps_thrd0_ctxt = ps_master_ctxt->aps_me_ctxt[0];
/*************************************************************************/
/* Updation of ME's DPB list. This involves the following steps: */
/* 1. Obtain list of active POCs maintained within ME. */
/* 2. Search each of them in the ref list. Whatever is not found goes to */
/* the list to be removed. Note: a_pocs_buffered_in_me holds the */
/* currently active POC list within ME. a_pocs_to_remove holds the */
/* list of POCs to be removed, terminated by -1. */
/*************************************************************************/
hme_coarse_get_active_pocs_list((void *)ps_thrd0_ctxt, a_pocs_buffered_in_me);
count = 0;
while(a_pocs_buffered_in_me[count] != -1)
{
WORD32 poc_to_search = a_pocs_buffered_in_me[count];
WORD32 match_found_flag = 0;
/*********************************************************************/
/* Search in any one list (L0/L1) since both lists contain all the */
/* active ref pics. */
/*********************************************************************/
for(i = 0; i < num_ref_l0; i++)
{
if(poc_to_search == pps_rec_list_l0[i]->i4_poc)
{
match_found_flag = 1;
break;
}
}
for(i = 0; i < num_ref_l1; i++)
{
if(poc_to_search == pps_rec_list_l1[i]->i4_poc)
{
match_found_flag = 1;
break;
}
}
if(0 == match_found_flag)
{
/*****************************************************************/
/* POC buffered inside ME but not part of ref list given by DPB */
/* Hence this needs to be flagged to ME for removal. */
/*****************************************************************/
a_pocs_to_remove[poc_remove_id] = poc_to_search;
poc_remove_id++;
}
count++;
}
/* List termination */
a_pocs_to_remove[poc_remove_id] = -1;
/* Call the ME API to remove "outdated" POCs */
hme_coarse_discard_frm(ps_thrd0_ctxt, a_pocs_to_remove);
}
/*!
******************************************************************************
* \if Function name : ihevce_coarse_me_frame_init \endif
*
* \brief
* Coarse Frame level ME initialisation function
*
* \par Description:
* The following pre-conditions exist for this function: a. We have the input
* pic ready for encode, b. We have the reference list with POC, L0/L1 IDs
* and ref ptrs ready for this picture and c. ihevce_me_set_resolution has
* been called atleast once. Once these are supplied, the following are
* done here: a. Input pyramid creation, b. Updation of ME's internal DPB
* based on available ref list information
*
* \param[in] pv_ctxt : pointer to ME module
* \param[in] ps_frm_ctb_prms : CTB characteristics parameters
* \param[in] ps_frm_lamda : Frame level Lambda params
* \param[in] num_ref_l0 : Number of reference pics in L0 list
* \param[in] num_ref_l1 : Number of reference pics in L1 list
* \param[in] num_ref_l0_active : Active reference pics in L0 dir for current frame (shall be <= num_ref_l0)
* \param[in] num_ref_l1_active : Active reference pics in L1 dir for current frame (shall be <= num_ref_l1)
* \param[in] pps_rec_list_l0 : List of recon pics in L0 list
* \param[in] pps_rec_list_l1 : List of recon pics in L1 list
* \param[in] ps_enc_lap_inp : pointer to input yuv buffer (frame buffer)
* \param[in] i4_frm_qp : current picture QP
*
* \return
* None
*
* \author
* Ittiam
*
*****************************************************************************
*/
void ihevce_coarse_me_frame_init(
void *pv_me_ctxt,
ihevce_static_cfg_params_t *ps_stat_prms,
frm_ctb_ctxt_t *ps_frm_ctb_prms,
frm_lambda_ctxt_t *ps_frm_lamda,
WORD32 num_ref_l0,
WORD32 num_ref_l1,
WORD32 num_ref_l0_active,
WORD32 num_ref_l1_active,
recon_pic_buf_t **pps_rec_list_l0,
recon_pic_buf_t **pps_rec_list_l1,
ihevce_lap_enc_buf_t *ps_enc_lap_inp,
WORD32 i4_frm_qp,
ihevce_ed_blk_t *ps_layer1_buf, //EIID
ihevce_ed_ctb_l1_t *ps_ed_ctb_l1,
UWORD8 *pu1_me_reverse_map_info,
WORD32 i4_temporal_layer_id)
{
/* local variables */
coarse_me_master_ctxt_t *ps_master_ctxt = (coarse_me_master_ctxt_t *)pv_me_ctxt;
coarse_me_ctxt_t *ps_ctxt;
coarse_me_ctxt_t *ps_thrd0_ctxt;
WORD32 inp_poc, num_ref;
WORD32 i;
/* Input POC is derived from input buffer */
inp_poc = ps_enc_lap_inp->s_lap_out.i4_poc;
num_ref = num_ref_l0 + num_ref_l1;
/* All processing done using shared / common memory across */
/* threads is done using thrd 0 ctxt */
ps_thrd0_ctxt = ps_master_ctxt->aps_me_ctxt[0];
ps_master_ctxt->s_frm_prms.u1_num_active_ref_l0 = num_ref_l0_active;
ps_master_ctxt->s_frm_prms.u1_num_active_ref_l1 = num_ref_l1_active;
/* store the frm ctb ctxt to all the thrd ctxt */
{
WORD32 num_thrds;
/* initialise the parameters for all the threads */
for(num_thrds = 0; num_thrds < ps_master_ctxt->i4_num_proc_thrds; num_thrds++)
{
ps_ctxt = ps_master_ctxt->aps_me_ctxt[num_thrds];
ps_ctxt->pv_ext_frm_prms = (void *)ps_frm_ctb_prms;
/*EIID: early decision buffer pointer */
ps_ctxt->ps_ed_blk = ps_layer1_buf;
ps_ctxt->ps_ed_ctb_l1 = ps_ed_ctb_l1;
/* weighted pred enable flag */
ps_ctxt->i4_wt_pred_enable_flag = ps_enc_lap_inp->s_lap_out.i1_weighted_pred_flag |
ps_enc_lap_inp->s_lap_out.i1_weighted_bipred_flag;
if(1 == ps_ctxt->i4_wt_pred_enable_flag)
{
/* log2 weight denom */
ps_ctxt->s_wt_pred.wpred_log_wdc =
ps_enc_lap_inp->s_lap_out.i4_log2_luma_wght_denom;
}
else
{
/* default value */
ps_ctxt->s_wt_pred.wpred_log_wdc = DENOM_DEFAULT;
}
ps_ctxt->i4_L1_hme_best_cost = 0;
ps_ctxt->i4_L1_hme_sad = 0;
ps_ctxt->i4_num_blks_high_sad = 0;
ps_ctxt->i4_num_blks = 0;
ps_ctxt->pv_me_optimised_function_list = ps_master_ctxt->pv_me_optimised_function_list;
ps_ctxt->ps_cmn_utils_optimised_function_list = &ps_master_ctxt->s_cmn_opt_func;
}
}
/* Create the reference map for ME */
ihevce_me_create_ref_map(
pps_rec_list_l0,
pps_rec_list_l1,
num_ref_l0_active,
num_ref_l1_active,
num_ref,
&ps_master_ctxt->s_ref_map);
/*************************************************************************/
/* Call the ME frame level processing for further actiion. */
/* ToDo: Support Row Level API. */
/*************************************************************************/
ps_master_ctxt->s_frm_prms.i2_mv_range_x = ps_thrd0_ctxt->s_init_prms.max_horz_search_range;
ps_master_ctxt->s_frm_prms.i2_mv_range_y = ps_thrd0_ctxt->s_init_prms.max_vert_search_range;
ps_master_ctxt->s_frm_prms.is_i_pic = 0;
ps_master_ctxt->s_frm_prms.i4_temporal_layer_id = i4_temporal_layer_id;
ps_master_ctxt->s_frm_prms.is_pic_second_field =
(!(ps_enc_lap_inp->s_input_buf.i4_bottom_field ^
ps_enc_lap_inp->s_input_buf.i4_topfield_first));
{
S32 pic_type = ps_enc_lap_inp->s_lap_out.i4_pic_type;
/*********************************************************************/
/* For I Pic, we do not call update fn at ctb level, instead we do */
/* one shot update for entire picture. */
/*********************************************************************/
if((pic_type == IV_I_FRAME) || (pic_type == IV_II_FRAME) || (pic_type == IV_IDR_FRAME))
{
ps_master_ctxt->s_frm_prms.is_i_pic = 1;
ps_master_ctxt->s_frm_prms.bidir_enabled = 0;
}
else if((pic_type == IV_P_FRAME) || (pic_type == IV_PP_FRAME))
{
ps_master_ctxt->s_frm_prms.bidir_enabled = 0;
}
else if((pic_type == IV_B_FRAME) || (pic_type == IV_BB_FRAME))
{
ps_master_ctxt->s_frm_prms.bidir_enabled = 1;
}
else
{
/* not sure whether we need to handle mixed frames like IP, */
/* they should ideally come as single field. */
/* TODO : resolve thsi ambiguity */
ASSERT(0);
}
}
/************************************************************************/
/* Lambda calculations moved outside ME and to one place, so as to have */
/* consistent lambda across ME, IPE, CL RDOPT etc */
/************************************************************************/
{
#define CLIP3_F(min, max, val) (((val) < (min)) ? (min) : (((val) > (max)) ? (max) : (val)))
double q_steps[6] = { 0.625, 0.703, 0.79, 0.889, 1.0, 1.125 };
double d_b_pic_factor;
double d_q_factor;
//double d_lambda;
UWORD8 u1_temp_hier = ps_enc_lap_inp->s_lap_out.i4_temporal_lyr_id;
if(u1_temp_hier)
{
d_b_pic_factor = CLIP3_F(2.0, 4.0, (i4_frm_qp - 12.0) / 6.0);
}
else
d_b_pic_factor = 1.0;
d_q_factor = (1 << (i4_frm_qp / 6)) * q_steps[i4_frm_qp % 6];
ps_master_ctxt->s_frm_prms.qstep = (WORD32)d_q_factor;
ps_master_ctxt->s_frm_prms.i4_frame_qp = i4_frm_qp;
}
/* HME Dependency Manager : Reset the num ctb processed in every row */
/* for ME sync in every layer */
{
WORD32 ctr;
for(ctr = 1; ctr < ps_thrd0_ctxt->num_layers; ctr++)
{
void *pv_dep_mngr_state;
pv_dep_mngr_state = ps_master_ctxt->apv_dep_mngr_hme_sync[ctr - 1];
ihevce_dmgr_rst_row_row_sync(pv_dep_mngr_state);
}
}
/* Frame level init of all threads of ME */
{
WORD32 num_thrds;
/* initialise the parameters for all the threads */
for(num_thrds = 0; num_thrds < ps_master_ctxt->i4_num_proc_thrds; num_thrds++)
{
ps_ctxt = ps_master_ctxt->aps_me_ctxt[num_thrds];
hme_coarse_process_frm_init(
(void *)ps_ctxt, ps_ctxt->ps_hme_ref_map, ps_ctxt->ps_hme_frm_prms);
}
}
ps_master_ctxt->s_frm_prms.i4_cl_sad_lambda_qf = ps_frm_lamda->i4_cl_sad_lambda_qf;
ps_master_ctxt->s_frm_prms.i4_cl_satd_lambda_qf = ps_frm_lamda->i4_cl_satd_lambda_qf;
ps_master_ctxt->s_frm_prms.i4_ol_sad_lambda_qf = ps_frm_lamda->i4_ol_sad_lambda_qf;
ps_master_ctxt->s_frm_prms.i4_ol_satd_lambda_qf = ps_frm_lamda->i4_ol_satd_lambda_qf;
ps_master_ctxt->s_frm_prms.lambda_q_shift = LAMBDA_Q_SHIFT;
ps_master_ctxt->s_frm_prms.pf_interp_fxn = NULL;
/*************************************************************************/
/* If num ref is 0, that means that it has to be coded as I. Do nothing */
/* However mv bank update needs to happen with "intra" mv. */
/*************************************************************************/
if(ps_master_ctxt->s_ref_map.i4_num_ref == 0 || ps_master_ctxt->s_frm_prms.is_i_pic)
{
for(i = 1; i < ps_thrd0_ctxt->num_layers; i++)
{
layer_ctxt_t *ps_layer_ctxt = ps_thrd0_ctxt->ps_curr_descr->aps_layers[i];
BLK_SIZE_T e_blk_size;
S32 use_4x4;
/* The mv bank is filled with "intra" mv */
use_4x4 = hme_get_mv_blk_size(
ps_thrd0_ctxt->s_init_prms.use_4x4,
i,
ps_thrd0_ctxt->num_layers,
ps_thrd0_ctxt->u1_encode[i]);
e_blk_size = use_4x4 ? BLK_4x4 : BLK_8x8;
hme_init_mv_bank(ps_layer_ctxt, e_blk_size, 2, 1, ps_ctxt->u1_encode[i]);
hme_fill_mvbank_intra(ps_layer_ctxt);
/* Clear out the global mvs */
memset(
ps_layer_ctxt->s_global_mv,
0,
sizeof(hme_mv_t) * ps_thrd0_ctxt->max_num_ref * NUM_GMV_LOBES);
}
return;
}
/*************************************************************************/
/* Coarse & refine Layer frm init (layer mem is common across thrds) */
/*************************************************************************/
{
coarse_prms_t s_coarse_prms;
refine_prms_t s_refine_prms;
S16 i2_max;
S32 layer_id;
layer_id = ps_thrd0_ctxt->num_layers - 1;
i2_max = ps_thrd0_ctxt->ps_curr_descr->aps_layers[layer_id]->i2_max_mv_x;
i2_max = MAX(i2_max, ps_thrd0_ctxt->ps_curr_descr->aps_layers[layer_id]->i2_max_mv_y);
s_coarse_prms.i4_layer_id = layer_id;
{
S32 log_start_step;
/* Based on Preset, set the starting step size for Refinement */
if(ME_MEDIUM_SPEED > ps_ctxt->s_init_prms.s_me_coding_tools.e_me_quality_presets)
{
log_start_step = 0;
}
else
{
log_start_step = 1;
}
s_coarse_prms.i4_max_iters = i2_max >> log_start_step;
s_coarse_prms.i4_start_step = 1 << log_start_step;
}
s_coarse_prms.i4_num_ref = ps_master_ctxt->s_ref_map.i4_num_ref;
s_coarse_prms.do_full_search = 1;
s_coarse_prms.num_results = ps_thrd0_ctxt->max_num_results_coarse;
hme_coarse_frm_init(ps_thrd0_ctxt, &s_coarse_prms);
layer_id--;
/*************************************************************************/
/* This loop will run for all refine layers (non- encode layers) */
/*************************************************************************/
while(layer_id > 0)
{
layer_ctxt_t *ps_curr_layer;
layer_ctxt_t *ps_coarse_layer;
ps_coarse_layer = ps_thrd0_ctxt->ps_curr_descr->aps_layers[layer_id + 1];
ps_curr_layer = ps_thrd0_ctxt->ps_curr_descr->aps_layers[layer_id];
hme_set_refine_prms(
&s_refine_prms,
ps_thrd0_ctxt->u1_encode[layer_id],
ps_master_ctxt->s_ref_map.i4_num_ref,
layer_id,
ps_thrd0_ctxt->num_layers,
ps_thrd0_ctxt->num_layers_explicit_search,
ps_thrd0_ctxt->s_init_prms.use_4x4,
&ps_master_ctxt->s_frm_prms,
NULL,
&ps_thrd0_ctxt->s_init_prms.s_me_coding_tools);
hme_refine_frm_init(ps_curr_layer, &s_refine_prms, ps_coarse_layer);
layer_id--;
}
}
return;
}
/*!
******************************************************************************
* \if Function name : ihevce_decomp_pre_intra_frame_init \endif
*
* \brief
* Frame Intialization for Decomp intra pre analysis.
*
* \param[in] pv_ctxt : pointer to module ctxt
* \param[in] ppu1_decomp_lyr_bufs : pointer to array of layer buffer pointers
* \param[in] pi4_lyr_buf_stride : pointer to array of layer buffer strides
*
* \return
* None
*
* \author
* Ittiam
*
*****************************************************************************
*/
WORD32 ihevce_coarse_me_get_lyr_buf_desc(
void *pv_me_ctxt, UWORD8 **ppu1_decomp_lyr_bufs, WORD32 *pi4_lyr_buf_stride)
{
/* local variables */
coarse_me_master_ctxt_t *ps_master_ctxt = (coarse_me_master_ctxt_t *)pv_me_ctxt;
coarse_me_ctxt_t *ps_thrd0_ctxt;
WORD32 lyr_no;
layers_descr_t *ps_curr_descr;
WORD32 i4_free_idx;
/* All processing done using shared / common memory across */
/* threads is done using thrd0 ctxt */
ps_thrd0_ctxt = ps_master_ctxt->aps_me_ctxt[0];
/* Obtain an empty layer descriptor */
i4_free_idx = hme_coarse_find_free_descr_idx((void *)ps_thrd0_ctxt);
ps_curr_descr = &ps_thrd0_ctxt->as_ref_descr[i4_free_idx];
/* export all the layer buffers except Layer 0 (encode layer) */
for(lyr_no = 1; lyr_no < ps_thrd0_ctxt->num_layers; lyr_no++)
{
pi4_lyr_buf_stride[lyr_no - 1] = ps_curr_descr->aps_layers[lyr_no]->i4_inp_stride;
ppu1_decomp_lyr_bufs[lyr_no - 1] = ps_curr_descr->aps_layers[lyr_no]->pu1_inp;
}
return (i4_free_idx);
}
/*!
******************************************************************************
* \if Function name : ihevce_coarse_me_get_lyr_prms_job_que \endif
*
* \brief Returns to the caller key attributes related to dependency between layers
* for multi-thread execution
*
*
* \par Description:
* This function requires the precondition that the width and ht of encode
* layer is known, and ME API ihevce_me_set_resolution() API called with
* this info. Based on this, ME populates useful information for the encoder
* to execute the multi-thread (concurrent across layers) in this API.
* The number of layers, number of vertical units in each layer, and for
* each vertial unit in each layer, its dependency on previous layer's units
* From ME's perspective, a vertical unit is one which is smallest min size
* vertically (and spans the entire row horizontally). This is CTB for encode
* layer, and 8x8 / 4x4 for non encode layers.
*
* \param[in] pv_ctxt : ME handle
* \param[in] ps_curr_inp : Input buffer descriptor
* \param[out] pi4_num_hme_lyrs : Num of HME layers (ME updates)
* \param[out] pi4_num_vert_units_in_lyr : Array of size N (num layers), each
* entry has num vertical units in that particular layer
* \param[in] ps_me_job_q_prms : Array of job queue prms, one for each unit in a
* layer. Note that this is contiguous in order of processing
* All k units of layer N-1 from top to bottom, followed by
* all m units of layer N-2 .... ends with X units of layer 0
*
* \return
* None
*
* \author
* Ittiam
*
*****************************************************************************
*/
void ihevce_coarse_me_get_lyr_prms_job_que(
void *pv_me_ctxt,
ihevce_lap_enc_buf_t *ps_curr_inp,
WORD32 *pi4_num_hme_lyrs,
WORD32 *pi4_num_vert_units_in_lyr,
multi_thrd_me_job_q_prms_t *ps_me_job_q_prms)
{
coarse_me_ctxt_t *ps_ctxt;
coarse_me_master_ctxt_t *ps_master_ctxt = (coarse_me_master_ctxt_t *)pv_me_ctxt;
/* These arrays and ptrs track input dependencies for units of a layer */
/* This is a ping poing design, while using one part, we update other part */
U08 au1_inp_dep[2][MAX_NUM_VERT_UNITS_FRM];
U08 *pu1_inp_dep_c, *pu1_inp_dep_n;
/* Height of current and next layers */
S32 ht_c, ht_n;
/* Blk ht at a given layer and next layer*/
S32 unit_ht_c, unit_ht_n, blk_ht_c, blk_ht_n;
/* Number of vertical units in current and next layer */
S32 num_vert_c, num_vert_n;
S32 ctb_size = 64, num_layers, i, j, k;
/* since same layer desc pointer is stored in all thread ctxt */
/* a free idx is obtained using 0th thread ctxt pointer */
ps_ctxt = ps_master_ctxt->aps_me_ctxt[0];
/* Set the number of layers */
num_layers = ps_ctxt->num_layers;
*pi4_num_hme_lyrs = num_layers;
pu1_inp_dep_c = &au1_inp_dep[0][0];
pu1_inp_dep_n = &au1_inp_dep[1][0];
ASSERT(num_layers >= 2);
ht_n = ps_ctxt->a_ht[num_layers - 2];
ht_c = ps_ctxt->a_ht[num_layers - 1];
/* compute blk ht and unit ht for c and n */
if(ps_ctxt->u1_encode[num_layers - 1])
{
blk_ht_c = 16;
unit_ht_c = ctb_size;
}
else
{
blk_ht_c = hme_get_blk_size(ps_ctxt->s_init_prms.use_4x4, num_layers - 1, num_layers, 0);
unit_ht_c = blk_ht_c;
}
num_vert_c = (ht_c + unit_ht_c - 1) / unit_ht_c;
/* For new design in Coarsest HME layer we need */
/* one additional row extra at the end of frame */
/* hence num_vert_c is incremented by 1 */
num_vert_c++;
/* Dummy initialization outside loop, not used first time */
memset(pu1_inp_dep_c, 0, num_vert_c);
/*************************************************************************/
/* Run through each layer, set the number of vertical units and job queue*/
/* attrs for each vert unit in the layer */
/*************************************************************************/
for(i = num_layers - 1; i > 0; i--)
{
/* 0th entry is actually layer id num_layers - 1 */
/* and entry num_layers-1 equals the biggest layer (id = 0) */
pi4_num_vert_units_in_lyr[num_layers - 1 - i] = num_vert_c;
/* "n" is computed for first time */
ht_n = ps_ctxt->a_ht[i - 1];
blk_ht_n = hme_get_blk_size(ps_ctxt->s_init_prms.use_4x4, i - 1, num_layers, 0);
unit_ht_n = blk_ht_n;
if(ps_ctxt->u1_encode[i - 1])
unit_ht_n = ctb_size;
num_vert_n = (ht_n + unit_ht_n - 1) / unit_ht_n;
/* Initialize all units' inp dep in next layer to 0 */
memset(pu1_inp_dep_n, 0, num_vert_n * sizeof(U08));
/* Evaluate dependencies for this layer */
for(j = 0; j < num_vert_c; j++)
{
S32 v1, v2;
/* Output dependencies. When one unit in current layer finishes, */
/* how many in the next layer it affects?. Assuming that the top */
/* of this vertical unit and bottom of this vertical unit project*/
/* somewhere in the next layer. The top of this vertical unit */
/* becomes the bottom right point for somebody, and the bottom of*/
/* this vertical unit becomes the colocated pt for somebody, this*/
/* is the extremum. */
/* for the initial unit affected by j in "c" layer, take j-1th */
/* unit top and project it. */
v1 = (j - 1) * unit_ht_c * ht_n;
v1 /= (ht_c * unit_ht_n);
v1 -= 1;
/* for the final unit affected by j in "c" layer, take jth unit */
/* bottom and project it. */
v2 = (j + 1) * unit_ht_c * ht_n;
v2 /= (ht_c * unit_ht_n);
v2 += 1;
/* Clip to be within valid limits */
v1 = HME_CLIP(v1, 0, (num_vert_n - 1));
v2 = HME_CLIP(v2, 0, (num_vert_n - 1));
/* In the layer "n", units starting at offset v1, and upto v2 are*/
/* dependent on unit j of layer "c". So for each of these units */
/* increment the dependency by 1 corresponding to "jth" unit in */
/* layer "c" */
ps_me_job_q_prms->i4_num_output_dep = v2 - v1 + 1;
ASSERT(ps_me_job_q_prms->i4_num_output_dep <= MAX_OUT_DEP);
for(k = v1; k <= v2; k++)
pu1_inp_dep_n[k]++;
/* Input dependency would have been calculated in prev run */
ps_me_job_q_prms->i4_num_inp_dep = pu1_inp_dep_c[j];
ASSERT(ps_me_job_q_prms->i4_num_inp_dep <= MAX_OUT_DEP);
/* Offsets */
for(k = v1; k <= v2; k++)
ps_me_job_q_prms->ai4_out_dep_unit_off[k - v1] = k;
ps_me_job_q_prms++;
}
/* Compute the blk size and vert unit size in each layer */
/* "c" denotes curr layer, and "n" denotes the layer to which result */
/* is projected to */
ht_c = ht_n;
blk_ht_c = blk_ht_n;
unit_ht_c = unit_ht_n;
num_vert_c = num_vert_n;
/* Input dep count for next layer was computed this iteration. */
/* Swap so that p_inp_dep_n becomes current for next iteration, */
/* and p_inp_dep_c will become update area during next iteration */
/* for next to next. */
{
U08 *pu1_tmp = pu1_inp_dep_n;
pu1_inp_dep_n = pu1_inp_dep_c;
pu1_inp_dep_c = pu1_tmp;
}
}
/* LAYER 0 OR ENCODE LAYER UPDATE : NO OUTPUT DEPS */
/* set the numebr of vertical units */
pi4_num_vert_units_in_lyr[num_layers - 1] = num_vert_c;
for(j = 0; j < num_vert_c; j++)
{
/* Here there is no output dependency for ME. However this data is used for encode, */
/* and there is a 1-1 correspondence between this and the encode */
/* Hence we set output dependency of 1 */
ps_me_job_q_prms->i4_num_output_dep = 1;
ps_me_job_q_prms->ai4_out_dep_unit_off[0] = j;
ps_me_job_q_prms->i4_num_inp_dep = pu1_inp_dep_c[j];
ASSERT(ps_me_job_q_prms->i4_num_inp_dep <= MAX_OUT_DEP);
ps_me_job_q_prms++;
}
return;
}
/*!
******************************************************************************
* \if Function name : ihevce_coarse_me_set_lyr1_mv_bank \endif
*
* \brief
* Frame level ME initialisation of MV bank of penultimate layer
*
* \par Description:
* Updates the Layer1 context with the given buffers
*
* \param[in] pv_me_ctxt : pointer to ME module
* \param[in] pu1_mv_bank : MV bank buffer pointer
* \param[in] pu1_ref_idx_bank : refrence bank buffer pointer
*
* \return
* None
*
* \author
* Ittiam
*
*****************************************************************************
*/
void ihevce_coarse_me_set_lyr1_mv_bank(
void *pv_me_ctxt,
ihevce_lap_enc_buf_t *ps_enc_lap_inp,
void *pv_mv_bank,
void *pv_ref_idx_bank,
WORD32 i4_curr_idx)
{
coarse_me_ctxt_t *ps_thrd0_ctxt;
coarse_me_master_ctxt_t *ps_master_ctxt = (coarse_me_master_ctxt_t *)pv_me_ctxt;
layer_ctxt_t *ps_lyr1_ctxt;
/* Input descriptor that is updated and passed to ME */
hme_inp_desc_t s_inp_desc;
/*************************************************************************/
/* Add the current input to ME's DPB. This will also create the pyramids */
/* for the HME layers tha are not "encoded". */
/*************************************************************************/
s_inp_desc.i4_poc = ps_enc_lap_inp->s_lap_out.i4_poc;
s_inp_desc.s_layer_desc[0].pu1_y = (UWORD8 *)ps_enc_lap_inp->s_lap_out.s_input_buf.pv_y_buf;
s_inp_desc.s_layer_desc[0].pu1_u = (UWORD8 *)ps_enc_lap_inp->s_lap_out.s_input_buf.pv_u_buf;
s_inp_desc.s_layer_desc[0].pu1_v = (UWORD8 *)ps_enc_lap_inp->s_lap_out.s_input_buf.pv_v_buf;
s_inp_desc.s_layer_desc[0].luma_stride = ps_enc_lap_inp->s_lap_out.s_input_buf.i4_y_strd;
s_inp_desc.s_layer_desc[0].chroma_stride = ps_enc_lap_inp->s_lap_out.s_input_buf.i4_uv_strd;
hme_coarse_add_inp(pv_me_ctxt, &s_inp_desc, i4_curr_idx);
/* All processing done using shared / common memory across */
/* threads is done using thrd 0 ctxt since layer ctxt is shared accross all threads */
ps_thrd0_ctxt = ps_master_ctxt->aps_me_ctxt[0];
ps_lyr1_ctxt = ps_thrd0_ctxt->ps_curr_descr->aps_layers[1];
/* register the mv bank & ref idx bank pointer */
ps_lyr1_ctxt->ps_layer_mvbank->pi1_ref_idx_base = (S08 *)pv_ref_idx_bank;
ps_lyr1_ctxt->ps_layer_mvbank->ps_mv_base = (hme_mv_t *)pv_mv_bank;
return;
}
/*!
******************************************************************************
* \if Function name : ihevce_coarse_me_get_lyr1_ctxt \endif
*
* \brief
* function to get teh Layer 1 properties to be passed on the encode layer
*
* \par Description:
* Ucopies the enitre layer ctxt emory to the destination
*
* \param[in] pv_me_ctxt : pointer to ME module
* \param[in] pu1_mv_bank : MV bank buffer pointer
* \param[in] pu1_ref_idx_bank : refrence bank buffer pointer
*
* \return
* None
*
* \author
* Ittiam
*
*****************************************************************************
*/
void ihevce_coarse_me_get_lyr1_ctxt(
void *pv_me_ctxt, void *pv_layer_ctxt, void *pv_layer_mv_bank_ctxt)
{
coarse_me_ctxt_t *ps_thrd0_ctxt;
coarse_me_master_ctxt_t *ps_master_ctxt = (coarse_me_master_ctxt_t *)pv_me_ctxt;
layer_ctxt_t *ps_lyr1_ctxt;
/* All processing done using shared / common memory across */
/* threads is done using thrd 0 ctxt since layer ctxt is shared accross all threads */
ps_thrd0_ctxt = ps_master_ctxt->aps_me_ctxt[0];
/* get the context of layer 1 */
ps_lyr1_ctxt = ps_thrd0_ctxt->ps_curr_descr->aps_layers[1];
/* copy the layer ctxt eve registerd mv bank & ref idx bank also goes in */
memcpy(pv_layer_ctxt, ps_lyr1_ctxt, sizeof(layer_ctxt_t));
/* copy the layer mv bank contents */
memcpy(pv_layer_mv_bank_ctxt, ps_lyr1_ctxt->ps_layer_mvbank, sizeof(layer_mv_t));
/* register the MV bank pointer in the layer ctxt*/
((layer_ctxt_t *)pv_layer_ctxt)->ps_layer_mvbank = (layer_mv_t *)pv_layer_mv_bank_ctxt;
return;
}
Reference in new issue View Git Blame Copy Permalink