Zynq UltraScale+ MPSoC VCU TRD 2021.1 - Xilinx Low Latency PL DDR XV20 HDMI Video Capture and Display

This page provides all the information related to Design Module 9 - VCU TRD Xilinx low latency(LLP2) PL DDR XV20 HDMI design.

Table of Contents

1 Overview

This module enables capture of video from an HDMI-Rx subsystem implemented in the PL. The video can be displayed through the HDMI-Tx subsystem implemented in the PL. The module can stream-out and stream-in live captured video frames through an Ethernet interface at ultra-low latencies using Sync IP. This module supports four video streams using AXI broadcaster at capture side and mixer at display side for XV20 pixel format. In this design PL_DDR is used for decoding and PS_DDR for encoding so that DDR bandwidth would be enough to support high bandwidth VCU applications requiring simultaneous encoder and decoder operations and transcoding at 4k @60 FPS.

The VCU encoder and decoder operate in slice mode. An input frame is divided into multiple slices (8 or 16) horizontally. The encoder generates a slice_done interrupt at every end of the slice. Generated NAL unit data can be passed to a downstream element immediately without waiting for the frame_done interrupt. The VCU decoder also starts processing data as soon as one slice of data is ready in its circular buffer instead of waiting for complete frame data. The Sync IP does an AXI transaction-level tracking so that the producer and consumer can be synchronized at the granularity of AXI transactions instead of granularity at the video buffer level. Sync IP is responsible for synchronizing buffers between Capture DMA and VCU encoder as both work on same buffer.

The capture element (FB write DMA) writes video buffers in raster-scan order. SyncIP monitors the buffer level while the capture element is writing into DRAM and allows the encoder to read input buffer data if the requested data is already written by DMA, otherwise it blocks the encoder until DMA completes its writes. On the decoder side, the VCU decoder writes decoded video buffer data into DRAM in block-raster scan order and displays reads data in raster-scan order. To avoid display under-run problems, software ensures a phase difference of "~frame_period/2", so that decoder is ahead compare to display.

This design supports the following video interfaces:

Sources:

  • HDMI-Rx capture pipeline implemented in the PL.

  • Stream-In from network or internet.

Sinks:

  • HDMI-Tx display pipeline implemented in the PL.

VCU Codec:

  • Video Encode/Decode capability using VCU hard block in PL.

    • AVC/HEVC encoding

    • Encoder/decoder parameter configuration.

Video format:

  • XV20

Supported Resolution:

The table below provides the supported resolution from command line app only in this design.

Resolution

Command Line

Single Stream

Multi-stream

4kp60

NA

4kp30

√ (Max 2)

1080p60

√ (Max 4 for encoder) (Max 2 for decoder)

√ - Supported
NA – Not applicable
x – Not supported

When using Low Latency mode (LLP1/LLP2), The encoder and decoder are limited by the number of internal cores. The encoder has maximum of four streams and the decoder has maximum of two streams.

The below table gives information about the features supported in this design. 

Pipeline

Input source

Format

Output Type

Resolution

VCU codec

Pipeline

Input source

Format

Output Type

Resolution

VCU codec

Serial pipeline

HDMI-Rx

XV20

HDMI-Tx

4kp60 / 4kp30 / 1080p60

HEVC/AVC

Stream-Out pipeline

HDMI-Rx

XV20

Stream-Out

4kp60 / 4kp30 / 1080p60

HEVC/AVC

Stream-in pipeline

Stream-In

XV20

HDMI-Tx

4kp60 / 4kp30 / 1080p60

HEVC/AVC

The below figure shows the Xilinx Low Latency PL DDR XV20 HDMI design hardware block diagram.

The below figure shows the Xilinx Low Latency PL DDR XV20 HDMI design software block diagram.

1.1 Board Setup

Refer below link for Board Setup

1.2 Run Flow

The TRD package is released with the source code, Vivado project, Petalinux BSP, and SD card image that enables the user to run the demonstration. It also includes the binaries necessary to configure and boot the ZCU106 board. Prior to running the steps mentioned in this wiki page, download the TRD package and extract its contents to a directory referred to as TRD_HOME which is the home directory.

Refer Section 4.1 : Download the TRD of Zynq UltraScale+ MPSoC VCU TRD 2021.1 wiki page to download all TRD contents.

TRD package contents are placed in the following directory structure. The user needs to copy all the files from the $TRD_HOME/images/vcu_llp2_hdmi_xv20/ to FAT32 formatted SD card directory.

rdf0428-zcu106-vcu-trd-2021-1/ ├── apu │   └── vcu_petalinux_bsp ├── images │   ├── vcu_10g │   ├── vcu_audio │   ├── vcu_llp2_hdmi_nv12 │   ├── vcu_llp2_hdmi_nv16 │   ├── vcu_llp2_hdmi_xv20 │   ├── vcu_llp2_sdi_xv20 │   ├── vcu_multistream_nv12 │   ├── vcu_pcie │   ├── vcu_plddrv1_hdr10_hdmi │   ├── vcu_plddrv2_hdr10_hdmi │   └── vcu_sdi_xv20 ├── pcie_host_package │   ├── COPYING │   ├── include │   ├── LICENSE │   ├── readme.txt │   ├── RELEASE │   ├── tests │   ├── tools │   └── xdma ├── pl │   ├── constrs │   ├── designs │   ├── prebuild │   ├── README.md │   └── srcs └── README.txt └── zcu106_vcu_trd_sources_and_licenses.tar.gz

TRD package contents specific to VCU Xilinx Low Latency PL DDR XV20 HDMI design are placed in the following directory structure.

rdf0428-zcu106-vcu-trd-2021-1/ ├── apu │   └── vcu_petalinux_bsp │   └── xilinx-vcu-zcu106-v2021.1-final.bsp ├── images │   ├── vcu_llp2_hdmi_xv20 │   │   ├── autostart.sh │   │   ├── BOOT.BIN │ │ ├── boot.scr │   │   ├── config │   │   ├── Image │   │   ├── rootfs.cpio.gz.u-boot │   │   ├── system.dtb │   │   └── vcu ├── pcie_host_package ├── pl │   ├── constrs │   ├── designs │   │   ├── zcu106_llp2_xv20 │   ├── prebuild │   │   ├── zcu106_llp2_xv20 │   ├── README.md │   └── srcs │   ├── hdl │   └── ip └── README.txt └── zcu106_vcu_trd_sources_and_licenses.tar.gz

Configuration files(input.cfg) for various resolutions are placed in the following directory structure in /media/card.

config ├── 1-4kp60 │   ├── Display │   └── Stream-out ├── 2-1080p60 │   ├── Display │   └── Stream-out ├── 2-4kp30 │   ├── Display │   └── Stream-out └── 4-1080p60 │ ├── Display │ ├── Stream-in │ └── Stream-out └── input.cfg

1.2.1 GStreamer Application (vcu_gst_app)

The vcu_gst_app is a command line multi-threaded linux application. The command line application requires an input configuration file (input.cfg) to be provided in the plain text.

Run below modetest command to set CRTC configurations for 4kp60:

Run below modetest command to set CRTC configurations for 4kp30:

Execution of the application is shown below:

Example:

  • Make sure HDMI-Rx should be configured to 4kp60 mode, while running below example pipelines.

  • Low latency(LLP1/LLP2) stream-in pipelines are not supported in vcu_gst_app.

4kp60 XV20 HEVC_25Mbps ultra low-latency(LLP2) display pipeline execution.

4kp60 XV20 HEVC_25Mbps ultra low-latency(LLP2) stream-out pipeline execution.

4kp60 XV20 HEVC ultra low-latency(LLP2) stream-in pipeline execution.

For LLP1/LLP2 Multistream HEVC serial and stream-out use-cases (2-4kp30, 2-1080p60, 4-1080p60), use ENC_EXTRA_OP_BUFFERS=10 variable before vcu_gst_app command. Below is the sample pipeline:

The above macro is recommended to use for LLP1/LLP2 multi-stream HEVC use-cases only.

To measure the latency of the pipeline, run the below command. The latency data is huge, so dump it to a file.

Refer below link for detailed run flow steps

1.3 Build Flow

Refer below link for detailed build flow steps


2 Other Information

2.1 Known Issues

2.2 Limitations

2.3 Optimum VCU Encoder parameters for use-cases

Video streaming:

  • Video streaming use-case requires very stable bitrate graph for all pictures.

  • It is good to avoid periodic large Intra pictures during the encoding session

  • Low-latency rate control (hardware RC) is the preferred control-rate for video streaming, it tries to maintain equal amount frame sizes for all pictures.

  • Good to avoid periodic Intra frames instead use low-delay-p (IPPPPP…)

  • VBR is not a preferred mode of streaming.

Performance: AVC Encoder settings:

  • It is preferred to use 8 slices only for better AVC encoder performance.

  • AVC standard does not support Tile mode processing which results in the processing of MB rows sequentially for entropy coding.

Quality: Low bitrate AVC encoding:

  • Enable profile=high and use qp-mode=auto for low-bitrate encoding use-cases.

  • The high profile enables 8x8 transform which results in better video quality at low bitrates.

2.4 Max Bit-rate Benchmarking

The following tables summarize the maximum bit rate achievable for 3840x2610p60 resolution, XV20 pixel format at GStreamer level. The maximum supported target bit rate values vary based on what elements and type of input used in the pipeline.

Maximum Bit Rate support for LLP1/LLP2 Streaming Use case with 4kp60 resolution.

The table below provides Encoder/Decoder Maximum Bit Rate Tests with XV20 format (For Streaming).

Video Streaming ( Server: Live video capture → VCU encoder → Parser → rtppay → Stream-out )

(Client: Stream-in → rtpdepay → Decoder → Display )

Format

Codec

 Rate Control Mode

 Latency Mode

B-Frames = 0

      DDR Mode

Max Target Bitrate



4:2:2, 10 bit


H.264 (AVC)



LOW_LATENCY

LLP1



IPPP




Encoder (PS_DDR),

Decoder (PL_DDR)


25 Mb/s

LLP2

25 Mb/s

H.265 (HEVC)

LLP1

25 Mb/s

LLP2

25 Mb/s

Maximum Bit Rate support for LLP1/LLP2 Serial Use case with 4kp60 resolution.

The table below provides Encoder/Decoder Maximum Bit Rate Tests with XV20 format.

Serial ( Live video capture → VCU encoder → VCU decoder → Display )

Format

Codec

 Rate Control Mode

 Latency Mode

B-Frames = 0

      DDR Mode

Max Target Bitrate



4:2:2, 10 bit


H.264 (AVC)



LOW_LATENCY

LLP1



IPPP




Encoder (PS_DDR),

Decoder (PL_DDR)


25 Mb/s

LLP2

25 Mb/s

H.265 (HEVC)

LLP1

25 Mb/s

LLP2

25 Mb/s


3 Appendix A - Input Configuration File (input.cfg)

The example configuration files are stored at /media/card/config/ folder.

Configuration Type

Configuration Name

Description

Available Options

Note

Configuration Type

Configuration Name

Description

Available Options

Note

Common

Common configuration

It is the starting point of common configuration

 

 

Number of Input

 

1,2,3,4

 

Output

Select the video interface.

HDMI

 

Out Type

 

display and stream

 

Display Rate

Pipeline frame rate

30 or 60 FPS

 

Exit

It indicates to the application that the configuration is over

 

 

Input

Input Configuration

It is the starting point of the input configuration

 

 

Input Numbers

Starting Nth input configuration

1, 2, 3, 4

 

Input Type

Input Type

HDMI

 

Raw

To tell the pipeline is processed or pass-through

FALSE

Raw use-case is not supported for both LLP2 and non-LLP2 use-case as mixer is not connected to PS DDR

Width

The width of the live source

3840, 1920

 

Height

The height of the live source

2160,1080

 

Format

The format of input data

XV20

 

Enable LLP2

To enable LLP2 configuration

TRUE, FALSE

Set Enable LLP2 equals to False for non-LLP2 use-case.

Exit

It indicates to the application that the configuration is over

 

 

Encoder

Encoder Configuration

It is the starting point of encoder configuration

 

 

Encoder Number

Starting Nth encoder configuration

1,2,3,4

 

Encoder Name

Name of the encoder

AVC/HEVC

 

Profile

Name of the profile

high for AVC

main for HEVC

 

Rate Control

Rate control options

Low_Latency

 

Filler Data

Filler Data NAL units for CBR rate control

False

 

QP

QP control mode used by the VCU encoder

Uniform, Auto

 

L2 Cache

Enable or Disable L2Cache buffer in encoding process

True, False

 

Latency Mode

Encoder latency mode

sub_frame

 

Low Bandwidth

If enabled, decrease the vertical search range used for P-frame motion estimation to reduce the bandwidth.

True, False

 

Gop Mode

Group of Pictures mode

Basic, low_delay_p, low_delay_b

 

Bitrate

Target bitrate in Kbps

1-25000

 

B Frames

Number of B-frames between two consecutive P-frames

0

 

Slice

The number of slices produced for each frame. Each slice contains one or more complete macroblock/CTU row(s). Slices are distributed over the frame as regularly as possible. If slice-size is defined as well more slices may be produced to fit the slice-size requirement.

  • 4-22 4K resolution with HEVC codec

  • 4-32 4K resolution with AVC codec

  • 4-32 1080p resolution with HEVC codec

  • 4-32 1080p resolution with AVC codec

The recommended slice for LLP2 use-case is 8.

GoP Length

The distance between two consecutive I frames

1-1000

 

GDR Mode

It specifies which Gradual Decoder Refresh(GDR) scheme should be used when gop-mode = low_delay_p

Horizontal/Vertical/Disabled

GDR mode is currently supported with LLP1/LLP2 low-delay-p use-cases only

Entropy Mode

It specifies the entropy mode for H.264 (AVC) encoding process

CAVLC/CABAC/Default

 

Max Picture Size

It is used to curtail instantaneous peak in the bit-stream using this parameter. It works in CBR/VBR rate-control only. When it is enabled, max-picture-size value is calculated and set with 10% of AllowedPeakMargin. i.e. max-picture-size =  (TargetBitrate / FrameRate) * 1.1

TRUE/FALSE

 

Preset

 

Custom

 

Exit

It indicates to the application that the configuration is over

 

 

Streaming

Streaming Configuration

It is the starting point of streaming configuration

 

 

Streaming Number

Starting Nth Streaming configuration

1, 2, 3, 4

 

Host IP

The host to send the packets to

192.168.25.89 or Windows PC IP

 

Port

The port to send the packets to

5004, 5008, 5012 and 5016

 

Exit

It indicates to the application that the configuration is over

 

 

Trace

Trace Configuration

It is the starting point of trace configuration