I have custom code for an RTMP server, and trying to add a gstreamer pipeline to transcode the incoming video and supply those to RTMP playback clients. I have the following pipeline so far:
appsrc name=input !
decodebin !
videoscale ! video/x-raw,width=800,height=600 !
x264enc speed-preset=veryfast name=encoder !
appsink name=output
In order for my appsink's h264 samples to be utilized I need to generate a sequence header packet with the codec data. However, I can't find the codec data.
The caps for the x264enc step is:
"video/x-h264, stream-format=(string)byte-stream, alignment=(string)au, level=(string)3.2, profile=(string)high, width=(int)800, height=(int)600, pixel-aspect-ratio=(fraction)4/3, framerate=(fraction)60/1, interlace-mode=(string)progressive, colorimetry=(string)bt709, chroma-site=(string)mpeg2"
There's no codec_data caps there, which is how I passed it to appsrc so my h264 packets could be read. How else do I get the codec data from gstream?
I was able to extract the codec_data by passing the x264enc output into h264parse.
Related
i am trying to share an h264 encoded data from gstreamer to another two processes(both are based on gstreamer).After some research only way i found is to use the shm plugin.
this is what i am trying to do
gstreamer--->h264 encoder--->shmsink
shmrc--->process1
shmrc--->process2
i was able to get raw data from videotestsrc and webcam working. But for h264 encoded data it doesn't.
this is my test pipeline
gst-launch-1.0 videotestsrc ! video/x-raw,width=640,height=480,format=YUY2 !
x264enc ! shmsink socket-path=/tmp/foo sync=true wait-for-
connection=false shm-size=10000000
gst-launch-1.0 shmsrc socket-path=/tmp/foo ! avdec_h264 ! video/x-
raw,width=640,height=480,framerate=25/1,format=YUY2 ! autovideosink
have anyone tried shm plugins with h264 encoded data, please help
Iam not aware of the capabilities of your sink used in "autovideosink", but as per my knowledge you either need to use videoconvert if the format supported by the sink (like kmssink or ximagesink) are different than provided by the source (in your case YUY2) or use videoparse if the camera format is supported by the sink. You may check this using gst-inspect-1.0 for the formats supported.
Anyways I am able to run your pipeline with some modifications using videoconvert in my setup :
./gst-launch-1.0 videotestsrc ! x264enc ! shmsink socket-path=/tmp/foo sync=true wait-for-connection=false shm-size=10000000
./gst-launch-1.0 shmsrc socket-path=/tmp/foo ! h264parse ! avdec_h264 ! videoconvert ! ximagesink
You may modify it as per the resolutions you want.
Kindly let me know if you face any issue with above.
I would like to store a file which has AAC audio frames,
For that i used the below pipeline,
gst-launch-1.0 filesrc location=Test_44100Hz_2ch_s16le.wav ! "audio/x-raw,rate=44100,format=s16le,channels=2" ! audioparse format=raw raw-format=s16le rate=44100 channels=2 ! faac ! aacparse ! queue ! filesink location=a1
While reading that file again to pulsesink using below pipeline,
gst-launch-1.0 filesrc location=a1 ! aacparse ! faad ! audioconvert ! audioresample ! pulsesink
I am Receiving below error, I used GST_DEBUG=3, but i am not able find the solution.
0:00:00.031924804 3379 0x2231d60 WARN basesrc gstbasesrc.c:3483:gst_base_src_start_complete:<filesrc0> pad not activated yet
Pipeline is PREROLLING ...
0:00:00.033044700 3379 0x2231050 WARN baseparse gstbaseparse.c:3255:gst_base_parse_loop:<aacparse0> error: No valid frames found before end of stream
ERROR: from element /GstPipeline:pipeline0/GstAacParse:aacparse0: No valid frames found before end of stream
Additional debug info:
gstbaseparse.c(3255): gst_base_parse_loop (): /GstPipeline:pipeline0/GstAacParse:aacparse0
ERROR: pipeline doesn't want to preroll.
Can anybody help me, To solve this? I need to store AAC audio frames and need to stream that file as AAC audio stream.
This is it, tested working:
gst-launch-1.0 filesrc location=WAV_44_16bit.wav ! decodebin ! audioconvert ! queue ! voaacenc ! aacparse ! queue ! mp4mux ! filesink location=aac.mp4
gst-launch-1.0 filesrc location=aac.mp4 ! decodebin ! audioconvert ! audioresample ! alsasink
In container there are metadata information stored.. without them the decoder does not know how to process the data.
AAC Audio streams require a container in order to be useful within gstreamer
For decoder initialization it is necessary to know sampling frequency and Audio Object. In gstreamer we are unable to pass this metadata directly to the parser or the decoder. The parser collects this data instead from the mp4 header then the encoder inherits the frame structure/size and sample rate. So this is a deficiency in either aacparse(parser) or avdec_aac/faad(decoder), none of which have exposed parameters to specify frame size of a raw file, the afore mentioned metadata. That being said, I haven't found a compelling reason why anyone would need to do this. I found myself trying to do it before I discovered the aac simply needed to be muxed into an MP4(mp4mux) or another container to work and be portable. The container/framing only adds a small amount of data to the stream.
I can successfully stream HD video using following pipelines:
streame server:
gst-launch-1.0 filesrc location="Gravity.2013.720p.BluRay.x264.YIFY.mp4" ! decodebin ! x264enc ! rtph264pay pt=96 ssrc=0 timestamp-offset=0 seqnum-offset=0 pt=96 ! gdppay ! tcpclientsink host=192.168.1.93 port=5000
client:
gst-launch-1.0 tcpserversrc host=192.168.1.93 port=5000 ! gdpdepay ! rtph264depay ! decodebin ! autovideosink
I want to add the audio stream too.
I guess it is possible to use a different port and tcpserver/tcpclient combination to stream audio parallel to video. But I am not certain, how gstreamer would synchronize two streams properly to play the movie in the client end. Apart from this method, are there any other methods? such as muxing two streams before and demuxing it in client end?
I used the following GStreamer pipeline to store my encoded stream in a binary file:
gst-launch v4l2src ! videorate ! video/x-raw-yuv, framerate=\(fraction\)10/1 \
! videoscale ! video/x-raw-yuv, format=\(fourcc\)I420, width=640, height=480\
! ffmpegcolorspace ! x264enc ! fdsink > vid.bin
Now i want to play previously recorded files in GStreamer using the following pipeline:
cat vid.bin | gst-launch fdsrc ! ffdec_h264 ! autovideosink
But then it gives the following error:
Setting pipeline to PAUSED ...
Pipeline is PREROLLING ...
ERROR: from element /GstPipeline:pipeline0/ffdec_h264:ffdec_h2640: Internal GStreamer error: negotiation problem. Please file a bug at http://bugzilla.gnome.org/enter_bug.cgi?product=GStreamer.
Additional debug info:
gstffmpegdec.c(2804): gst_ffmpegdec_chain (): /GstPipeline:pipeline0/ffdec_h264:ffdec_h2640:
ffdec_h264: input format was not set before data start
ERROR: pipeline doesn't want to preroll.
Setting pipeline to NULL ...
Freeing pipeline ...
I know that the best way to capture video is using Muxers but is there any way to play my previous files?
Thanks
Not sure your pipeline is right.
If you want to write to a file why not simply use filesink and filesrc.
fdsink > vid.bin will not work fine because if you see the prints by gstreamer gst-launch will also go into the file. [Just open vid.bin in an text editor and you will see what I mean].
Also for x264 stream to be stored without a muxer you need to use byte-stream=1 in your x264enc to store it in annexb format so that it is decodable.
To play back raw x264 stream you need to have a color space convertor before the video sink
gst-launch filesrc location=inputfile ! legacyh264parse ! ffdec_h264 ! queue ! ffmpegcolorspace ! autovideosink
plays just fine here at my end
Or, to playback a raw h264 file with gstreamer 1.0:
gst-launch-1.0 filesrc location=/tmp/video.h264 ! h264parse ! avdec_h264 ! autovideosink
I am trying to create a GStreamer pipeline (v 1.0) in order to record and play special file format.
For recording purpose I use the following pipeline:
gst-launch-1.0 videotestsrc ! video/x-raw-yuv, format=\(fourcc\)I420, width=640, height=480 ! videoconvert ! x264enc byte-stream=1 ! queue ! appsink
In appsink (using new_sample() callback) I use a compression method to compress H264 stream and finally store in a output file.
I use the following pipeline to play the recorded file:
gst-launch-1.0 appsrc ! video/x-h264 ! avdec_h264 ! autovideosink
In appsrc I decompress H264 stream and send it to appsrc buffer (using push-buffer). The size of each buffer is 4095.
Unfortunately GStreamer after push 2 buffers print the following debug message:
Error: Internal data flow error.
Is there any way to fix the problem?
Add legacyh264parse or h264parse (depending on your version of gst components) before your decoder. You need to be able to send full frames to the decoder.
Post avdec_h264 it would be nice to have a ffmpegcolorspace to be able to convert the video format to your display requirements.