Qt Creator can be a very good IDE in order to develop great QT applications. This IDE does not only helps with syntax highlighting, access to examples and tutorials, but also helps you to configure different toolchains Qt binary versions and target options. First download the binary installer from: For 32 bits:
$ wget http://releases.qt-project.org/qtcreator/2.6.2/qt-creator-linux-x86-opensource-2.6.2.bin
For 64 bits:
$ wget http://releases.qt-project.org/qtcreator/2.6.2/qt-creator-linux-x86_64-opensource-2.6.2.bin
execute the binary
Follow the Installer GUI and choose a location. Default options should be OK. in my case the installation was done here:
Open Qt Creator (in my case from command line, use "&" to regain control of the terminal)
Open Tools -> Options Choose Build & Run on the menu of the left. and Select the Compilers Tab Here you can add the toolchain GCC compiler of your convenience. It will appear in the "Manual" section. Now click on Qt Version Tab. Here you can add the Qmake that you had created with your Qt installation; for example, the Qt5 installation described here: Building QT for i.MX6 It will appear in the Manual section. In my case I have Qmake for PC and Qmake for i.MX6. Now click on Kits Tab Here you can create combinations of Compilers and Qmake, and also specify where do you want the executables to go. In my case here I combined the i.MX6 toolchain and the Qmake for I.MX6 i had created. I did not set up device configuration since the sysroot is already shared to my device via NFS, but you can configure it so the files are sent via ssh to your device. And that's It! Next time you load a project you can choose which Kit you want to work on, and it will be compiled just as you need.
Watch the Freescale i.MX team boot up Android 5.0 Lollipop in i.mx6 application processors—在线播放—优酷网，视频高清在线观看 The Freescale i.MX Android team has booted up Android 5.0 Lollipop in the SABRE platform for i.mx6 series. Google pushed all of the latest source for its Android release to AOSP on Nov. 5, and the Freescale Android Team started their work. With the previous 6 days to boot Android Lollipop up, the Freescale i.MX Android team enabled the basic features like connectivity, audio/video playback, sensors, inputs and display on day 7! You can see the some changes in the demo video at the beginning of the post. The Freescale i.MX Android team has closely followed almost every version of Android since it is released by AOSP and has good experience on it. Below are some snapshots and pictures for the Android Lollipop.
If you want to use a USB camera (these types of cameras are also called 'Web Cameras') with GStreamer on i.MX6 devices (Linux Kernel version >= 3.035), you need to either load the module dynamically or compile and link statically selecting (Y) the following config on the Kernel configuration Device Drivers -> Multimedia support -> Video capture adapters -> V4L USB devices -> <*> USB Video Class (UVC) After the Kernel image has been built, flash it into the target, plug the web cam, then on a (target) terminal run gst-launch v4l2src ! mfw_v4lsink You should see what the camera is capturing on the display. In case you need to encode the camera src data, you need to place the encoder into the pipeline gst-launch v4l2src num-buffers=100 ! queue ! vpuenc codec=0 ! matroskamux ! filesink location=output.mkv sync=false We are using a certain codec (codec=0 means mpeg4), check options using 'gst-inspect vpuenc'.
In some cases it is desired to directly have progressive content available from a TV-IN interface through the V4L2 capture device. In the BSP, HW accelerated de-interlacing is only supported in the V4L2 output stream. Below is a patch created against a rather old BSP version that adds support for de-interlaced V4L2 capture. The patch might need to be adapted to newer BSPs, However, the logic and functionality is there and should shorten the development time. This patch adds another input device to the V4L2 framework that can be selected to perform the deinterlacing on the way to memory. The selection is done by passing the index “2” as an argument to the VIDIOC_S_INPUT V4L2 ioctl. Attached is also a modified the tvin unit test to give an example of how to use the new driver. An example sequence for running the test is as follows: modprobe mxc_v4l2_capture ./mxc_v4l2_tvin_vdi.out -ow 720 -oh 480 -ol 10 -ot 20 -f YU12 Some key things to note: This driver does not support resize or color space conversion on the way to memory. The requested format and size should match what can be provided directly by the sensor. The driver was tested on a Sabre AI Rev A board running Linux 12.02. This code is not an official delivery and as such no guarantee of support for this code is provided by Freescale.
Multiple-Display means video playback on multiple screens. In case playback needs to be in a unique screen, the mfw_isink element must be used and some pipelines examples can be found on this link: GStreamer iMX6 Multi-Overlay. Number of Displays Display type Kernel parameters Pipelines # Set these shells variables before running the pipelines alias gl=gst-launch SINK_1="\"mfw_v4lsink device=/dev/video17\"" SINK_2="\"mfw_v4lsink device=/dev/video18\"" SINK_3="\"mfw_v4lsink device=/dev/video20\"" media1=file:///root/media1 media2=file:///root/media2 media3=file:///root/media3 2 hdmi + lvds video=mxcfb0:dev=hdmi,1920x1080M@60,if=RGB24 video=mxcfb1:dev=ldb,LDB-XGA,if=RGB666 gl playbin2 uri=$media1 video-sink=$SINK_1 playbin2 uri=$media2 video-sink=$SINK_2 2 lvds + lvds video=mxcfb0:dev=ldb,LDB-XGA,if=RGB666 video=mxcfb1:dev=ldb,LDB-XGA,if=RGB666 gl playbin2 uri=$media1 video-sink=$SINK_1 playbin2 uri=$media2 video-sink=$SINK_2 2 lcd + lvds video=mxcfb0:dev=lcd,800x480M@55,if=RGB565 video=mxcfb1:dev=ldb,LDB-XGA,if=RGB666 gl playbin2 uri=$media1 video-sink=$SINK_1 playbin2 uri=$media2 video-sink=$SINK_2 3 hdmi + lvds + lvds video=mxcfb0:dev=hdmi,1920x1080M@60,if=RGB24 video=mxcfb1:dev=ldb,LDB-XGA,if=RGB6 video=mxcfb2:dev=ldb,LDB-XGA,if=RGB666 gl playbin2 uri=$media1 video-sink=$SINK_1 playbin2 uri=$media2 video-sink=$SINK_2 playbin2 uri=$media3 video-sink=$SINK_3
Audio, from a file gst-launch filesrc location=test.wav ! wavparse ! mfw_mp3encoder ! filesink location=output.mp3 Audio Recording gst-launch alsasrc num-buffers=$NUMBER blocksize=$SIZE ! mfw_mp3encoder ! filesink location=output.mp3 # where # duration = $NUMBER*$SIZE*8 / (samplerate *channel *bitwidth) # Example: 60 seconds recording # gst-launch alsasrc num-buffers=240 blocksize=44100 ! mfw_mp3encoder ! filesink location=output.mp3 # # To verify that is correct, do a normal audio playback gst-launch filesrc location=output.mp3 typefind=true ! beepdec ! audioconvert ! 'audio/x-raw-int,channels=2' ! alsasink Video, from a test source gst-launch videotestsrc ! queue ! vpuenc ! mat roskamux ! filesink location=./test.avi Video, from a file gst-launch filesrc location=sample.yuv blocksize=$BLOCK_SIZE ! 'video/x-raw-yuv,format=(fourcc)I420, width=$WIDTH, height=$HEIGHT, framerate=(fraction)30/1' ! vpuenc codec=$CODEC ! matroskamux ! filesink location=output.mkv sync=false # where # BLOCK_SIZE = WIDTH * HEIGHT * 1.5 # CODEC = 0(MPEG4), 5(H263), 6(H264) or 12(MJPG). # # For example, encoding a CIF raw file gst-launch filesrc location=sample.yuv blocksize=152064 ! 'video/x-raw-yuv,format=(fourcc)I420, width=352, height=288, framerate=(fraction)30/1' ! vpuenc codec=0 ! matroskamux ! filesink location=sample.mkv sync=false Video, from Web camera # when the web cam is connected, the device node /dev/video0 should be present. In order to test the camera, without encoding gst-launch v4l2src ! mfw_v4lsink # in recording, run: # gst-launch v4l2src num-buffers=-1 ! queue max-size-buffers=2 ! vpuenc codec=0 ! matroskamux ! filesink location=output.mkv sync=false # # where sync=false indicates filesink to to use a clock sync # # In case a specific width/height is needed, just add the filter caps gst-launch v4l2src num-buffers=-1 ! 'video/x-raw-yuv,format=(fourcc)I420, width=352, height=288, framerate=(fraction)30/1' ! queue ! vpuenc codec=0 ! matroskamux ! filesink location=output.mkv sync=false # # In case you want to see in the screen what the camera is capturing, add a tee element # gst-launch v4l2src num-buffers=-1 ! tee name=t ! queue ! mfw_v4lsink t. ! queue ! vpuenc codec=0 ! matroskamux ! filesink location=output.mkv sync=false Video, from Parallel/MIPI camera # The camera driver needs to be loaded before executing the pipeline, refer to the BSP document to see which driver to load # MIPI (J5 port): modprobe ov5640_camera_mipi modprobe mxc_v4l2_capture # Parallel (J9 port): modprobe ov5642_camera modprobe mxc_v4l2_capture gst-launch mfw_v4lsrc ! queue ! vpuenc codec=0 ! matroskamux ! filesink location=output.mkv sync=false # Do a 'gst-inspect mfw_v4lsrc' or 'gst-inspect vpuenc' to see other possible settings (resolution, fps, codec, etc.)
Notes: First run the playback pipeline then the streaming pipeline. The above example streams H263 video and AMR audio data. Change codec format to your needs. In case where the iMX is the streaming machine, the audio encoder 'amrnbenc' must be installed before. This scenario has not been tested Shell variables and pipelines Playback machine (receiver) # On playback machine, set either IMX2PC or PC2IMX variables, then run the pipeline ## IMX2PC: Case where PC does the playback AUDIO_DEC_SINK="rtpamrdepay ! amrnbdec ! alsasink " VIDEO_CAPS="\"application/x-rtp,media=(string)video,clock-rate=(int)90000,encoding-name=(string)H263-1998\"" VIDEO_DEC_SINK="rtph263pdepay ! ffdec_h263 ! autovideosink" ## End of IMX2PC Settings ## PC2IMX: Case where iMX does the playback AUDIO_DEC_SINK="rtpamrdepay ! mfw_amrdecoder ! alsasink " VIDEO_CAPS="\"application/x-rtp,media=(string)video,clock-rate=(int)90000,encoding-name=(string)H263-1998\"" VIDEO_DEC_SINK="rtph263pdepay ! vpudec ! mfw_v4lsink " ## End of PC2IMX Settings PLAYBACK_AUDIO="udpsrc caps=\"application/x-rtp,media=(string)audio,clock-rate=(int)8000,encoding-name=(string)AMR,encoding-params=(string)1,octet-align=(string)1\" \ port=5002 ! rtpbin.recv_rtp_sink_1 \ rtpbin. ! $AUDIO_DEC_SINK \ udpsrc port=5003 ! rtpbin.recv_rtcp_sink_1 \ rtpbin.send_rtcp_src_1 ! udpsink port=5007 sync=false async=false" PLAYBACK_VIDEO="udpsrc caps=$VIDEO_CAPS port=5000 ! rtpbin.recv_rtp_sink_0 \ rtpbin. ! $VIDEO_DEC_SINK \ udpsrc port=5001 ! rtpbin.recv_rtcp_sink_0 \ rtpbin.send_rtcp_src_0 ! udpsink port=5005 sync=false async=false" PLAYBACK_AV="$PLAYBACK_VIDEO $PLAYBACK_AUDIO" # Playback pipeline gst-launch -v gstrtpbin name=rtpbin $PLAYBACK_AV Streaming Machine (sender) # On Streaming machine, set either IMX2PC or PC2IMX variables, then run the pipeline ## IMX2PC: Case where iMX does the streaming IP=x.x.x.x # IP address of the playback machine VIDEO_SRC="mfw_v4lsrc" VIDEO_ENC="vpuenc codec=h263 ! rtph263ppay " AUDIO_ENC="audiotestsrc ! amrnbenc ! rtpamrpay " ## END IMX2PC settings ## PC2IMX: Case where PC does the streaming IP=y.y.y.y # IP address of the playback machine VIDEO_SRC="v4l2src" VIDEO_ENC="ffenc_h263 ! rtph263ppay " AUDIO_ENC="audiotestsrc ! amrnbenc ! rtpamrpay " # END PC2PC settings STREAM_AUDIO="$AUDIO_ENC ! rtpbin.send_rtp_sink_1 \ rtpbin.send_rtp_src_1 ! udpsink host=$IP port=5002 \ rtpbin.send_rtcp_src_1 ! udpsink host=$IP port=5003 sync=false async=false \ udpsrc port=5007 ! rtpbin.recv_rtcp_sink_1" STREAM_VIDEO="$VIDEO_SRC ! $VIDEO_ENC ! rtpbin.send_rtp_sink_0 \ rtpbin.send_rtp_src_0 ! queue ! udpsink host=$IP port=5000 \ rtpbin.send_rtcp_src_0 ! udpsink host=$IP port=5001 sync=false async=false \ udpsrc port=5005 ! rtpbin.recv_rtcp_sink_0" STREAM_AV="$STREAM_VIDEO $STREAM_AUDIO" # Stream pipeline gst-launch -v gstrtpbin name=rtpbin $STREAM_AV
(DEPRECATED. Please check this document for Real Time Streaming) A server can be streaming video and a client, in this case a i.MX6 target, is receiving and decoding it. For example, a server with GStreamer and a web camera connected, can be streaming with the following command:
$ # Pipeline 1
$ gst-launch v4l2src ! 'video/x-raw-yuv, format=(fourcc)I420, width=(int)1280, height=(int)800' ! ffenc_mpeg4 ! tcpserversink host=$CLIENT_IP port=$PORT
and on the target, the client receives, decodes and display with
$ # Pipeline 2
$ gst-launch tcpclientsrc host=$SERVER_IP port=$PORT ! 'video/mpeg, width=(int)1280, height=(int)800, framerate=(fraction)10/1, mpegversion=(int)4, systemstream=(boolean)false' ! vpudec ! mfw_isink
The filter caps between the tcpclientsrc and the decoder (vpudec) depend on the sink caps coming from the server encoder (ffenc_mpeg4), so these may change depending on your needs. Running the above pipelines require the environment variables SERVER_IP, CLIENT_IP and PORT. In case you want the i.MX6 to act as a server, just change the video source (either mfw_v4lsrc of v4l2src) and the encoder (vpuenc), so
$ # Pipeline 3
$ gst-launch v4l2src ! 'video/x-raw-yuv, format=(fourcc)I420, width=(int)640, height=(int)480, interlaced=(boolean)false, framerate=(fraction)10/1' ! vpuenc ! tcpserversink host=$CLIENT_IP port=$PORT
For testing purposes, set SERVER_IP=127.0.0.1, CLIENT_IP=127.0.0.1 and PORT=500, and run pipeline 3 and 2 in two different consoles. Check with 'top' the CPU usage and see that VPU is actually doing most of the work.
BlueZ5 provides support for the core Bluetooth layers and protocols. It is flexible, efficient and uses a modular implementation. BlueZ5 has implemented the Bluetooth low level host stack for Bluetooth core specification 4.0 and 3.0+HS which includes GAP, L2CAP, RFCOMM, and SDP. Besides the host stack, BlueZ5 has also supported the following profiles itself or via a third party software. Profiles p rovided by BlueZ: A2DP 1.3 AVRCP 1.5 DI 1.3 HDP 1.0 HID 1.0 PAN 1.0 SPP 1.1 GATT (LE) profiles: PXP 1.0 HTP 1.0 HoG 1.0 TIP 1.0 CSCP 1.0 OBEX based profiles (by obexd): FTP 1.1 OPP 1.1 PBAP 1.1 MAP 1.0 Provided by the oFono project: HFP 1.6 (AG & HF)Supported Profiles BlueZ5 has been supported in the latest Freescale Linux BSP release, so it would be pretty easy to generate the binaries for Bluetooth core stack and its profiles. In order to support A2DP sink on a SabreSD board, the following software should be downloaded and installed onto the target rootfs too. sbc decoder version 1.3 (http://www.kernel.org/pub/linux/bluetooth/sbc-1.3.tar.gz) PulseAudio 5.0 (http://www.freedesktop.org/software/pulseaudio/releases/pulseaudio-5.0.tar.xz) PulseAudio package has some dependencies with bluetooth and sbc packages, and pulseaudio will detect if the two packages have been built and then decide which pulse plugin modules to be generated. So the building order will be 1) bluez5_utils or bluez_utils 2) sbc 3) pulseaudio. After compile and install the above software onto the target rootfs, you should be able to see the following executable under the directory /usr/bin From BlueZ5: bluetoothctl, hciconfig, hciattach (Needed by operating a UART bluetooth module) From PulseAudio: pulseaudio, pactl, paplay If the building dependency has been setup correctly, the following pulse plugin modules should be located under the directory /usr/lib/pulse-5.0/modules module-bluetooth-discover.so module-bluetooth-policy.so module-bluez5-device.so module-bluez5-discover.so Edit the file /etc/dbus-1/system.d/pulseaudio-system.conf, and add the following lines in red:
< allow send_destination="org.bluez"/>
Edit the file /etc/dbus-1/system.d/bluetooth.conf, and add the following lines:
Adding the following settings at the bottom of the pulseaudio system configuration file which locates in /etc/pulse/system.pa
### Automatically load driver modules for Bluetooth hardware
load-module module-alsa-sink device_id=0 tsched=true tsched_buffer_size=1048576 tsched_buffer_watermark=262144
On the system that can automatically detect the alsa cards, the above line #13 should be removed. Also make sure "auth-anonymous=1" is added to the following line, which can resolve the issue: "Denied access to client with invalid authorization data".
load-module module-native-protocol-unix auth-anonymous=1
Selecting a audio re-sampling algorithm and configuring the audio output by adding the following settings to the file daemon.conf locating in /etc/pulse
resample-method = trivial
enable-remixing = no
enable-lfe-remixing = no
default-sample-format = s16le
default-sample-rate = 48000
alternate-sample-rate = 24000
default-sample-channels = 2
Pulseaudio can be started as a daemon or as a system-wide instance. To run PulseAudio in system-wide mode, the program will automatically drop privileges from "root" and change to the "pulse" user and group. In this case, before launching the program, the "pulse" user and group needs to be created on the target system. In the example below, "/var/run/pulse" is the home directory for "pulse" user.
adduser -h /var/run/pulse pulse
adduser pulse pulse-access
Because PulseAudio needs to access the sound devices, add the user "pulse" to the "audio" group too.
adduser pulse audio
Starting bluetoothd and pulseaudio:
/usr/libexec/bluetooth/bluetoothd -d &
pulseaudio --system --realtime &
To verify if the pulseaudio has been set up correctly, you can play a local wave file by using the following command. If you can hear the sound, the system should have been configured correctly.
paplay -vvv audio8k16S.wav
After setting up the pulseaudio, launch bluetoothctl to pair and connect to a mobile phone. After connecting to a mobile phone, you should be able to see the following information in bluetoothctl console:
Name: BlueZ 5.21
Alias: BlueZ 5.21
UUID: PnP Information (00001200-0000-1000-8000-00805f9b34fb)
UUID: Generic Access Profile (00001800-0000-1000-8000-00805f9b34fb)
UUID: Generic Attribute Profile (00001801-0000-1000-8000-00805f9b34fb)
UUID: A/V Remote Control (0000110e-0000-1000-8000-00805f9b34fb)
UUID: A/V Remote Control Target (0000110c-0000-1000-8000-00805f9b34fb)
UUID: Message Notification Se.. (00001133-0000-1000-8000-00805f9b34fb)
UUID: Message Access Server (00001132-0000-1000-8000-00805f9b34fb)
UUID: Phonebook Access Server (0000112f-0000-1000-8000-00805f9b34fb)
UUID: IrMC Sync (00001104-0000-1000-8000-00805f9b34fb)
UUID: OBEX File Transfer (00001106-0000-1000-8000-00805f9b34fb)
UUID: OBEX Object Push (00001105-0000-1000-8000-00805f9b34fb)
UUID: Vendor specific (00005005-0000-1000-8000-0002ee000001)
UUID: Audio Source (0000110a-0000-1000-8000-00805f9b34fb)
UUID: Audio Sink (0000110b-0000-1000-8000-00805f9b34fb)
If you can see the audio sink UUID, you are ready to enjoy the bluetooth music now.
Audio transcoding # Transcode the input file into MP3 gst-launch filesrc location=media_file typefind=true ! beepdec ! mfw_mp3encoder ! matroskamux ! filesink location=output_audio_file.mk Audio transcoding + streaming # Transcode the input file into MP3 and stream it # On host, get the transcoded audio data gst-launch udpsrc port=8880 ! <CAPS_FROM_THE_TARGET> ! queue ! ffdec_mp3 ! alsasink # where <CAPS_FROM_THE_TARGET> can be something like 'audio/mpeg, mpegversion=(int)1, layer=(int)3, rate=(int)44100, channels=(int)2' # run the pipeline and check the caps gst-launch filesrc location=media_file typefind=true ! queue ! beepdec ! mfw_mp3encoder ! udpsink host=10.112.103.77 port=8880 -v Video Transcoding* # Transcode the input file into AVC (H-264) gst-launch filesrc location=media_file typefind=true ! aiurdemux name=demux ! queue ! vpudec ! vpuenc codec= avc ! matroskamux name=mux ! filesink location=output_media_file.mk Video Transcoding + scaling* # Transcode the input file into AVC (H264) and rescale video to 480p gst-launch filesrc location=media_file typefind=true ! aiurdemux name=demux ! queue ! vpudec ! mfw_ipucsc ! 'video/x-raw-yuv, width=(int) 720 , height=(int) 480 ' ! vpuenc codec= avc ! matroskamux name=mux ! filesink location=output_media_file.mk Video transcoding + scaling + streaming* # NOTE: Run the pipelines in the presented order # On host, get the transcoded+scaled video data $ gst-launch udpsrc port=8888 ! <CAPS_FROM_THE_TARGET> ! queue ! ffdec_h264 ! xvimagesink # On target, run the pipeline and check the caps gst-launch filesrc location=media_file typefind=true ! aiurdemux name=demux ! queue ! vpudec ! mfw_ipucsc ! 'video/x-raw-yuv, width=(int) 720 , height=(int) 480 ' ! vpuenc codec= avc ! udpsink host=$HOST_IP port=8888 -v * There is a limit for the number of pipelines which can be run simultaneously, for high resolution input files, at most two for 1080p and four for 720p.
GStreamer has a simple feature to enable tracing, allowing the developer to do basic debugging. These can be done in two ways: Adding the parameter --gst-debug=LIST to the pipeline (a pipeline is a executed gst-launch command) Prepending the environment variable GST_DEBUG=LIST' LIST is a a comma-separated argument, indicating the GStreamer elements to trace. For example, if one needs to trace the sink element $ GST_DEBUG=*sink*:5 gst-launch playbin2 uri=file:///sample.avi or $ gst-launch playbin2 uri=file:///sample.avi --gst-debug=*sink*:5 Both commands produces the same log. In case want to trace for than one element, so can simple add the <element>:5, for example $ GST_DEBUG=mfw_v4lsink:5,vpudec:5 gst-launch playbin2 uri=file:///sample.avi The number 5 indicates the log category, where 5 is the highest (the most verbose log you can get) and 0 produces no output (5=LOG, 4=DEBUG, 3=INFO, 2=WARN, 1=ERROR). Log can be huge in each pipeline run. One way to filter it is using the grep command. Before grepping, one needs to redirect the standard error to the standard output (GStreamer log goes always to stderr), so $ GST_DEBUG=mfw_v4lsink:5,vpudec:5 gst-launch playbin2 uri=file:///sample.avi 2>&1 | grep <filter string> In case the log needs to be shared, it is important to remove the 'color' of the log, again, one just needs to add the parameter --gst-debug-no-color or prepend the env variable GST_DEBUG_NO_COLOR=1 ----- More shell variables that GStreamer react, can be found here https://developer.gnome.org/gstreamer/0.10/gst-running.html
Freescale does not have a specific GStreamer element to do JPEG encoding, so the standard 'jpegenc' should be used. Image Capture With a web camera gst-launch v4l2src num-buffers=1 ! jpegenc ! filesink location=sample.jpeg With an embedded camera gst-launch mfw_v4lsrc num-buffers=1 ! jpegenc ! filesink location=sample.jpeg More pipelines on GStreamer i.MX6 Pipelines
Video decoding gst-launch filesrc location=sample.mp4 ! qtdemux ! ffdec_h264 ! mfw_v4lsink Notes: On LTIB BSP 3.0.35_4.0.0, prep the package and apply the attached patch on top, then build. On Yocto, the easy way to add the gst-ffmpeg package is by adding these two lines on the conf/local.conf file:
IMAGE_INSTALL_append = " gst-ffmpeg "
LICENSE_FLAGS_WHITELIST = 'commercial'
In this article, some experiments are done to verify the capability of i.MX6DQ on video playback under different VPU clocks. 1. Preparation Board: i.MX6DQ SD Bitstream: 1080p sunflower with 40Mbps, it is considered as the toughest H264 clip. The original clip is copied 20 times to generate a new raw video (repeat 20 times of sun-flower clip) and then encapsulate into a mp4 container. This is to remove and minimize the influence of startup workload of gstreamer compared to vpu unit test. Kernels: Generate different kernel with different VPU clock setting: 270MHz, 298MHz, 329MHz, 352MHz, 382MHz. test setting: 1080p content decoding and display with 1080p device. (no resize) 2. Test command for VPU unit test and Gstreamer The tiled format video playback is faster than NV12 format, so in below experiment, we choose tiled format during video playback. Unit test command: (we set the frame rate -a 70, higher than 1080p 60fps HDMI refresh rate) /unit_tests/mxc_vpu_test.out -D "-i /media/65a78bbd-1608-4d49-bca8-4e009cafac5e/sunflower_2B_2ref_WP_40Mbps.264 -f 2 -y 1 -a 70" Gstreamer command: (free run to get the highest playback speed) gst-launch filesrc location=/media/65a78bbd-1608-4d49-bca8-4e009cafac5e/sunflower_2B_2ref_WP_40Mbps.mp4 typefind=true ! aiurdemux ! vpudec framedrop=false ! queue max-size-buffers=3 ! mfw_v4lsink sync=false 3. Video playback framerate measurement During test, we enter command "echo performance > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor" to make sure the CPU always work at highest frequency, so that it can respond to any interrupt quickly. For each testing point with different VPU clock, we do 5 rounds of tests. The max and min values are removed, and the remaining 3 data are averaged to get the final playback framerate. #1 #2 #3 #4 #5 Min Max Avg Dec Playback Dec Playback Dec Playback Dec Playback Dec Playback Playback Playback Playback 270M unit test 57.8 57.3 57.81 57.04 57.78 57.3 57.87 56.15 57.91 55.4 55.4 57.3 56.83 GST 53.76 54.163 54.136 54.273 53.659 53.659 54.273 54.01967 298M unit test 60.97 58.37 60.98 58.55 60.97 57.8 60.94 58.07 60.98 58.65 57.8 58.65 58.33 GST 56.755 49.144 53.271 56.159 56.665 49.144 56.755 55.365 329M unit test 63.8 59.52 63.92 52.63 63.8 58.1 63.82 58.26 63.78 59.34 52.63 59.52 58.56667 GST 57.815 55.857 56.862 58.637 56.703 55.857 58.637 57.12667 352M unit test 65.79 59.63 65.78 59.68 65.78 59.65 66.16 49.21 65.93 57.67 49.21 59.68 58.98333 GST 58.668 59.103 56.419 58.08 58.312 56.419 59.103 58.35333 382M unit test 64.34 56.58 67.8 58.73 67.75 59.68 67.81 59.36 67.77 59.76 56.58 59.76 59.25667 GST 59.753 58.893 58.972 58.273 59.238 58.273 59.753 59.03433 Note: Dec column means the vpu decoding fps, while Playback column means overall playback fps. Some explanation: Why does the Gstreamer performance data still improve while unit test is more flat? On Gstreamer, there is a vpu wrapper which is used to make the vpu api more intuitive to be called. So at first, the overall GST playback performance is constrained by vpu (vpu dec 57.8 fps). And finally, as vpu decoding performance goes to higher than 60fps when vpu clock increases, the constraint becomes the display refresh rate 60fps. The video display overhead of Gstreamer is only about 1 fps, similar to unit test. Based on the test result, we can see that for 352MHz, the overall 1080p video playback on 1080p display can reach ~60fps. Or if time sharing by two pipelines with two displays, we can do 2 x 1080p @ 30fps video playback. However, this experiment is valid for 1080p video playback on 1080p display. If for interlaced clip and display with size not same as 1080p, the overall playback performance is limited by some postprocessing like de-interlacing and resize.
There is no Freescale GStreamer element which does the JPEG decoding, so we must rely on a standard one, like 'jpegdec'. In case your Linux system was built using LTIB, in order to have the jpegdec element included on the gst-plugin-good, follow these steps: On the LTIB menuconfig, make sure the following packages are selected: gstreamer-plugins-good libjpeg libpng Remove the configure parameters '--disbale-libpng' and '--disable-jpeg' on the file './dist/lfs-5.1/gst-plugins-good/gst-plugins-good.spec' Rebuild and flash your board (or SD card) again. Image display VSALPHA=1 gst-launch filesrc location=sample.jpeg ! jpegdec ! imagefreeze ! mfw_isink Important: non 8 pixel aligned width and height is treated as not supported format in isink plugin.