Newest 'ffmpeg' Questions - Stack Overflow

Les articles publiés sur le site

• Execute ffmpeg to Crop Audio in Qt

  4 mars 2013, par aswin

  I want to create an application which can be used to crop audio files using Qt. I've successfully run ffmpeg via command line to do this using the following command:

  ffmpeg -t 30 -i C:\\test.mp3 -acodec copy C:\\test2.mp3
  

  Then I tried to do this using Qt and my code below seems to work, but I can't find the output file (test2.mp3).

  QProcess* process=new QProcess(this);
  process->start("FFmpeg-N-49957-g8c95d17\\ffmpeg.exe",QStringList()<<"-t 30 -i C:\\test.mp3 -acodec copy C:\\test2.mp3");
  

  Is there anything wrong with my code above?

• Cuda Memory Management : re-using device memory from C calls (multithreaded, ffmpeg), but failing on cudaMemcpy

  4 mars 2013, par Nuke Stollak

  I'm trying to CUDA-fy my ffmpeg filter that was taking over 90% of the CPU time, according to gprof. I first went from one core to OpenMP on 4 cores and got a 3.8x increase in frames encoded per second, but it's still too slow. CUDA seemed like the next natural step.

  I've gotten a modest (20%?) increase by replacing one of my filter's functions with a CUDA kernel call, and just to get things up and running, I was cudaMalloc'ing and cudaMemcpy'ing on each frame. I suspected I would get better results if I weren't doing this each frame, so before I go ahead and move the rest of my code to CUDA, I wanted to fix this by allocating the memory before my filter is called and freeing it afterwards, but the device memory isn't having it. I'm only storing the device memory locations outside of code that knows about CUDA; I'm not trying to use the data there, just save it for the next time I call a CUDA-aware function that needs it.

  Here's where I am so far:

  Environment: the last AMI Linux on EC2's GPU Cluster, latest updates installed. Everything is fairly standard.

  My filter is split into two files: vf_myfilter.c (compiled by gcc, like almost every other file in ffmpeg) and vf_myfilter_cu.cu (compiled by nvcc). My Makefile's link step includes -lcudart and both .o files. I build vf_myfilter_cu.o using (as one line)

  nvcc -I. -I./ -I/opt/nvidia/cuda/include $(CPPFLAGS) 
       -Xcompiler "$(CFLAGS)" 
        -c -o libfilter/vf_myfilter_cu.o libfilter/vf_myfilter_cu.cu
  

  When the variables (set by configure) are expanded, here's what I get, again all in one line but split up here for easier reading. I just noticed the duplicate include path directives, but it shouldn't hurt.

  nvcc -I. -I./ -I/opt/nvidia/cuda/include -I. -I./ -D_ISOC99_SOURCE 
      -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -D_POSIX_C_SOURCE=200112
      -D_XOPEN_SOURCE=600 -DHAVE_AV_CONFIG_H 
      -XCompiler "-fopenmp -std=c99 -fomit-frame-pointer -pthread -g 
                  -Wdeclaration-after-statment -Wall -Wno-parentheses 
                  -Wno-switch -Wno-format-zero-length -Wdisabled-optimization  
                  -Wpointer-arith -Wredundant-decls -Wno-pointer-sign 
                  -Wwrite-strings -Wtype-limits -Wundef -Wmissing-prototypes
                  -Wno-pointer-to-int-case -Wstrict-prototypes -O3 -fno-math-errno 
                  -fno-signed-zeros -fno-tree-vectorize 
                  -Werror=implicit-function-declaration -Werror=missing-prototypes 
                  -Werror=vla " 
      -c -o libavfilter/vf_myfilter_cu.o libavfilter/vf_myfilter_cu.cu
  

  vf_myfilter.c calls three functions from vf_myfilter_cu.cu file which handle memory and call the CUDA kernel code. I thought I would be able to save the device pointers from my memory initialization, which runs once per ffmpeg run, and re-use that space each time I called the wrapper for my kernel function, but when I cudaMemcpy from my host memory to my device memory that I stored, it fails with cudaInvalidValue. If I cudaMalloc my device memory on every frame, I'm fine.

  I plan on using pinned host memory, once I have everything up in CUDA code and have minimized the number of times I need to return to the main ffmpeg code.

  Steps taken:

  First sign of trouble: search the web. I found Passing a pointer to device memory between classes in CUDA and printed out the pointers at various places in my execution to ensure that the device memory values were the same everywhere, and they are. FWIW, they seem to start around 0x90010000.

  ffmpeg's configure gave me -pthreads, so I checked to see if my filter was being called from multiple threads according to how can I tell if pthread_self is the main (first) thread in the process? and checking syscall(SYS_gettid) == getpid() to ensure that I'm not calling CUDA from different threads--I'm indeed in the primary thread at every step, according to those two funcs. I am still using OpenMP later around some for loops in the main .c filter function, but the calls to CUDA don't occur in those loops.

  Code Overview:

  ffmpeg provides me a MyfilterContext structure pointer on each frame, as well as on the filter's config_input and uninit routines (called once per file), so I added some *host_var and *dev_var variables (a few of each, float and unsigned char).

  There is a whole lot of code I skipped for this post, but most of it has to do with my algorithm and details involved in writing an ffmpeg filter. I'm actually using about 6 host variables and 7 device variables right now, but for demonstration I limited it to one of each.

  Here is, broadly, what my vf_myfilter.c looks like.

  // declare my functions from vf_myfilter_cu.cu 
  extern void cudaMyInit(unsigned char **dev_var, size_t mysize);
  extern void cudaMyUninit(unsigned char *dev_var);
  extern void cudaMyFunction(unsigned char *host_var, unsigned char *dev_var, size_t mysize);
  
  // part of the MyFilterContext structure, which ffmpeg keeps track of for me.
  typedef struct {
      unsigned char *host_var;
      unsigned char *dev_var;
  } MyFilterContext;
  
  // ffmpeg calls this function once per file, before any frames are processed.
  static int config_input(AVFilterLink *inlink) {
          // how ffmpeg passes me my context, fairly standard.
      MyfilterContext * myContext = inlink->dst->priv; 
          // compute the size one video plane of one frame of video
      size_t mysize = sizeof(unsigned char) * inlink->w * inlink->h;
          // av_mallocz is a malloc wrapper provided and required by ffmpeg
      myContext->host_var = (unsigned char*) av_mallocz(size);
          // Here's where I attempt to allocate my device memory.
      cudaMyInit( & myContext->dev_var, mysize);  
  }
  
  // Called once per frame of video
  static int filter_frame(AVFilterLink *inlink, AVFilterBufferRef *frame) {
      MyFilterContext *myContext = inlink->dst->priv;
  
      // sanity check to make sure that this isn't part of the multithreaded code
      if ( syscall(SYS_gettid) == getpid() ) 
          av_log(.... ); // This line never runs, so it's not threaded?
  
      // ...fill host_var with data from frame, 
      // set mysize to the size of the buffer
  
      // Call my wrapper function defined in the .cu file
      cudaMyFunction(myContext->host_var, myContext->dev_var, mysize);
  
      // ... take the results from host_var and apply them to frame
      // ... and return the processed frame to ffmpeg 
  }
  
  // called after everything else has happened:  free up the memory.
  static av_cold void uninit(AVFilterContext *ctx) {
      MyFilterContext *myContext = ctx->priv;
      // free my host_var
      if(myContext->host_var!=NULL) {
          av_free(myContext->host_var);
          myContext->host_var=NULL;
      }
      // free my dev_var
      cudaMyUninit(myContext->dev_var);
  }
  

  Here is, broadly, what my vf_myfilter_cu.cu looks like:

  // my kernel function that does the work.
  __global__ void myfunc(unsigned char *dev_var, size_t mysize) {
      // find the offset for this particular GPU thread to process
      // exit this function if the block/thread combo points to somewhere
      //     outside the frame
      // make sure we're less than mysize bytes from the beginning of dev_var
      // do things to dev_var[some_offset]
  } 
  // Allocate the device memory
  extern "C" void cudaMyInit(unsigned char **dev_var, size_t mysize) {
      if(cudaMalloc( (void**) dev_var, mysize) != cudaSuccess) {
          printf("Cannot allocate the memory\n");
      }
  }
  
  // Free the device memory.
  extern "C" void cudaMyUninit(unsigned char *dev_var) {
      cudaFree(dev_var);
  }
  
  // Copy data from the host to the device,
  // Call the kernel function, and 
  // Copy data from the device to the host.
  extern "C" void cudaMyFunction(
          unsigned char *host_var, 
          unsigned char *dev_var, 
          size_t mysize         ) 
  {
      cudaError_t cres;
  
      // dev_works is what I want to get rid of, but 
      // to make sure that there's not something more obvious going 
      // on, I made sure that my cudaMemcpy works if I'm allocating
      // the device memory in every frame.
      unsigned char *dev_works;  
      if(cudaMalloc( (void **) &dev_works, mysize)!=cudaSuccess) { 
          // I don't see this message
          printf("failed at per-frame malloc\n");
      }
  
      // THIS PART WORKS, copying host_var to dev_works
      cres=cudaMemcpy( (void *) dev_works, host_var, mysize, cudaMemcpyHostToDevice);
      if(cres!=cudaSuccess) {
          if(cres==cudaErrorInvalidValue) {
              // I don't see this message.
              printf("cudaErrorInvalidValue at per-frame cudaMemcpy\n");
          }
      }
  
      // THIS PART FAILS, copying host_var to dev_var
      cres=cudaMemcpy( (void *) dev_var, host_var, mysize, cudaMemcpyHostToDevice);
      if(cres!=cudaSuccess) {
          if(cres==cudaErrorInvalidValue) {
              // this is the error code that prints.
              printf("cudaErrorInvalidValue at per-frame cudaMemcpy\n");
          }
          // I check for other error codes, but they're not being hit.
      }
  
      // and this works with dev_works
      myfunc<<>>(dev_works, mysize);
  
      if(cudaMemcpy(host_var, dev_works, mysize, cudaMemcpyDeviceToHost)!=cudaSuccess) {
          // I don't see this message.
          printf("Failed to copy post-kernel func\n");
      }
  
      cudaFree(dev_works);
  
  }
  

  Any ideas?

• ffmpeg - connection refused [migrated]

  4 mars 2013, par Fibericon

  I'm attempting to stream from a file, using the following command:

  ffmpeg -re -i video.webm -c copy -f webm rtmp://localhost:8090/stream
  

  However, I get the following error:

  TCP connection to localhost:8090 failed: Connection refused
  

  This is the config file I'm using, which has the port, BindAddress, and ACL allow 127.0.0.1 already set. What's missing for this to be able to work?

  http://ffmpeg.org/sample.html

• How to build SDL libraries for Android

  4 mars 2013, par Harish

  I am planning to use SDL (Simple DirectMedia Layer) to display video output in my Android application that uses ffmpeg libraries. I have downloaded the sources from http://www.libsdl.org/download-1.2.php and built (./configure, make & make install) on my Ubuntu. But when I use these .so files the Android ndk-build complains that "Could not read symbols. File in wrong Format".

  Can I use the .so files that are built on Ubuntu on Android or do I need to build the SDL for Android in a different way?

• ffmpeg Live Input MP4 Error [migrated]

  3 mars 2013, par Brianjs

  Currently I have a mic and a webcam connected to my computer. I am running ffmpeg on CentOS 6.3.

  When I try to record a video without audio by:

  ffmpeg -y -f video4linux2 -t 15 -s 640x480 -r 25 -i /dev/video0 /home/irdb/Desktop/out2.mp4
  

  it runs perfectly and I get a nice video. However when I try to run with audio included by:

  ffmpeg -y -f video4linux2 -t 15 -s 640x480 -r 25 -i /dev/video0 -f alsa -ar 22050 -ab 64k -ac 2 -i default /home/irdb/Desktop/out2.mp4
  

  It errors out and prints:

  [NULL @ 0x1e33fc0] Codec is experimental but experimental codecs are not enabled,      see -strict -2
  Output #0, mp4, to '/home/irdb/Desktop/out2.mp4':
  Stream #0:0: Video: h264, yuv420p, 640x480, q=-1--1, 90k tbn, 25 tbc
  Stream #0:1: Audio: none, 22050 Hz, 2 channels, flt, 128 kb/s
  Stream mapping:
  Stream #0:0 -> #0:0 (rawvideo -> libx264)
  Stream #1:0 -> #0:1 (pcm_s16le -> aac)
  Error while opening encoder for output stream #0:1 - maybe incorrect parameters such    as bit_rate, rate, width or height
  

  I assume this has to do with the first error as when I use something like mpg it works just fine. However I plan on streaming this live and want mp4 format as that is pretty much supported by all browsers (Firefox with flash fallback).

  Does anyone know how to get the audio to work without additional processing (as I want to stream live and not write to a file eventually).