# SDl 2.0 and ffmpeg

**URL:** <https://discourse.libsdl.org/t/sdl-2-0-and-ffmpeg/18809>\
**Category:** SDL Development\
**Created:** [March 9, 2012, 10:09am UTC](https://discourse.libsdl.org/t/sdl-2-0-and-ffmpeg/18809 "2012-03-09T10:09:05Z")\
**Posts on this page:** 10\
**Page:** 1

<div class="post-metadata">

**Author:** ![David\_martin1](https://discourse.libsdl.org/letter_avatar_proxy/v4/letter/d/ed8c4c/32.png) [@David\_martin1](https://discourse.libsdl.org/u/David_martin1)\
**Post date:** [March 9, 2012, 10:09am UTC](https://discourse.libsdl.org/t/sdl-2-0-and-ffmpeg/18809/1 "2012-03-09T10:09:05Z")

</div>

Hi,  
I’m new to SDL2, I used to play with danger ffmpeg tutorials to stream  
video to an SDL surface.

I’m updating my program incorporating SDL2.

Here is a piece of snippet and would appreciate your help to actually  
draw the video image into the surface using the SDL2 api.  
thanks

…  
AVFrame\* frame ;  
AVPacket avpacket;  
int frame\_finished;  
SDL\_Event event;  
int w = 640;  
int h = 480;  
int done = 0, color = 0;  
SDL\_Window \* window = SDL\_CreateWindow(“SDL”,  
SDL\_WINDOWPOS\_UNDEFINED, SDL\_WINDOWPOS\_UNDEFINED, w, h, 0);  
SDL\_Renderer \* renderer = SDL\_CreateRenderer(window, -1, 0);  
SDL\_Log("+++++ INIT DONE +++++");

```
 frame = avcodec_alloc_frame();

     	while (av_read_frame(format_context, &avpacket) >= 0) {
     		if (avpacket.stream_index == videostream) {

     			avcodec_decode_video2(codec_context, frame, &frame_finished, 

```

&avpacket);

```
     		}

 while (!done) {
 	SDL_WaitEvent(&event);

 	 switch(event.type) {
 			//case FF_ALLOC_EVENT:
                  //alloc_picture(event.user.data1);
              // break;
              //case FF_REFRESH_EVENT:
                  //video_refresh_timer(event.user.data1);
                // break;
        case SDL_KEYDOWN:
        case SDL_KEYUP:
        //case SDL_MOUSEBUTTONDOWN:
        case SDL_QUIT:
        	done=1;
         SDL_Log("Unhandled event type=%d", event.type);

     }

     //Draw Image into renderer.
     //We used to do it through SDL_DisplayYUVOverlay. How to do 

```

that with SDL2.0. Any help for this part ?

```
     SDL_RenderClear(renderer);
     SDL_RenderPresent(renderer);
     SDL_Delay(20);
 }

  }//while av_read_frame

 SDL_Log("+++++ FINISHED +++++");
 SDL_Quit();

```

}

---

<div class="post-metadata">

**Author:** ![Dimitris\_Zenios](https://discourse.libsdl.org/letter_avatar_proxy/v4/letter/d/f14d63/32.png) [@Dimitris\_Zenios](https://discourse.libsdl.org/u/Dimitris_Zenios)\
**Post date:** [March 9, 2012, 10:27am UTC](https://discourse.libsdl.org/t/sdl-2-0-and-ffmpeg/18809/2 "2012-03-09T10:27:35Z")

</div>

Have a look at this.

> **[MPEG acceleration with GLSL](http://slouken.blogspot.com/2011/02/mpeg-acceleration-with-glsl.html)**
>
> Video decoding is something that people are always trying to find ways to accelerate.  Whether it's making HD video more HD or dynamically s...

It might help you

ZeniosOn Fri, Mar 9, 2012 at 12:09 PM, David martin wrote:

> Hi,  
> I’m new to SDL2, I used to play with danger ffmpeg tutorials to stream video  
> to an SDL surface.
> 
> I’m updating my program incorporating SDL2.
> 
> Here is a piece of snippet and would appreciate your help to actually draw  
> the video image into the surface using the SDL2 api.  
> thanks
> 
> …  
> ? ?AVFrame\* frame ;  
> ? ?AVPacket avpacket;  
> ? ?int frame\_finished;  
> ? ?SDL\_Event event;  
> ? ?int w = 640;  
> ? ?int h = 480;  
> ? ?int done = 0, color = 0;  
> ? ?SDL\_Window \* window = SDL\_CreateWindow(“SDL”, SDL\_WINDOWPOS\_UNDEFINED,  
> SDL\_WINDOWPOS\_UNDEFINED, w, h, 0);  
> ? ?SDL\_Renderer \* renderer = SDL\_CreateRenderer(window, -1, 0);  
> ? ?SDL\_Log(“+++++ INIT DONE +++++”);
> 
> ? ?frame = avcodec\_alloc\_frame();
> 
> ? ? ? ? ? ? ? ?while (av\_read\_frame(format\_context, &avpacket) \>= 0) {  
> ? ? ? ? ? ? ? ? ? ? ? ?if (avpacket.stream\_index == videostream) {
> 
> ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?avcodec\_decode\_video2(codec\_context, frame,  
> &frame\_finished, &avpacket);
> 
> ? ? ? ? ? ? ? ? ? ? ? ?}
> 
> ? ?while (!done) {  
> ? ? ? ?SDL\_WaitEvent(&event);
> 
> ? ? ? ? switch(event.type) {  
> ? ? ? ? ? ? ? ? ? ? ? ?//case FF\_ALLOC\_EVENT:  
> ? ? ? ? ? ? ? ? ? ? //alloc\_picture(event.user.data1);  
> ? ? ? ? ? ? ? ? // ? ?break;  
> ? ? ? ? ? ? ? ? //case FF\_REFRESH\_EVENT:  
> ? ? ? ? ? ? ? ? ? ? //video\_refresh\_timer(event.user.data1);  
> ? ? ? ? ? ? ? ? ? // ?break;  
> ? ? ? ? ? case SDL\_KEYDOWN:  
> ? ? ? ? ? case SDL\_KEYUP:  
> ? ? ? ? ? //case SDL\_MOUSEBUTTONDOWN:  
> ? ? ? ? ? case SDL\_QUIT:  
> ? ? ? ? ? ? ? ?done=1;  
> ? ? ? ? ? ?SDL\_Log(“Unhandled event type=%d”, event.type);
> 
> ? ? ? ?}
> 
> ? ? ? ?//Draw Image into renderer.  
> ? ? ? ?//We used to do it through SDL\_DisplayYUVOverlay. How to do that with  
> SDL2.0. ? ? ? ?Any help for this part ?
> 
> ? ? ? ?SDL\_RenderClear(renderer);  
> ? ? ? ?SDL\_RenderPresent(renderer);  
> ? ? ? ?SDL\_Delay(20);  
> ? ?}
> 
> ? ? }//while av\_read\_frame
> 
> ? ?SDL\_Log(“+++++ FINISHED +++++”);  
> ? ?SDL\_Quit();  
> }
> 
> * * *
> 
> SDL mailing list  
> SDL at [lists.libsdl.org](http://lists.libsdl.org)  
> [http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org](http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org)

---

<div class="post-metadata">

**Author:** ![David\_martin1](https://discourse.libsdl.org/letter_avatar_proxy/v4/letter/d/ed8c4c/32.png) [@David\_martin1](https://discourse.libsdl.org/u/David_martin1)\
**Post date:** [March 9, 2012, 2:02pm UTC](https://discourse.libsdl.org/t/sdl-2-0-and-ffmpeg/18809/3 "2012-03-09T14:02:05Z")

</div>

Thanks for the link.  
I’m reading a video streaming (not a single BMp image) so not sure how  
this applies to my code.  
ffmpeg reads in in yuvp42 format (see output from ffmpeg dump\_format)

> ~/workspace/test1/Release $ ./test2 toto.mpeg  
> Input #0, mpeg, from ‘toto.mpeg’:  
> Duration: 00:00:04.83, start: 1.000000, bitrate: 606 kb/s  
> Stream #0:0[0x1e0]: Video: mpeg1video, yuv420p, 352x288 [SAR  
> 178:163 DAR 1958:1467], 104857 kb/s, 60 fps, 60 tbr, 90k tbn, 60 tbc

So the question is do i need to create a surface from each image read  
frame ? How would you do that ?

thanks,On 03/09/2012 11:27 AM, Dimitris Zenios wrote:

> Have a look at this.  
> [Exploring the Galaxy: MPEG acceleration with GLSL](http://slouken.blogspot.com/2011/02/mpeg-acceleration-with-glsl.html)
> 
> It might help you
> 
> Zenios
> 
> On Fri, Mar 9, 2012 at 12:09 PM, David martin\<@David\_martin1\> wrote:
> 
> > Hi,  
> > I’m new to SDL2, I used to play with danger ffmpeg tutorials to stream video  
> > to an SDL surface.
> > 
> > I’m updating my program incorporating SDL2.
> > 
> > Here is a piece of snippet and would appreciate your help to actually draw  
> > the video image into the surface using the SDL2 api.  
> > thanks
> > 
> > …  
> > AVFrame\* frame ;  
> > AVPacket avpacket;  
> > int frame\_finished;  
> > SDL\_Event event;  
> > int w = 640;  
> > int h = 480;  
> > int done = 0, color = 0;  
> > SDL\_Window \* window = SDL\_CreateWindow(“SDL”, SDL\_WINDOWPOS\_UNDEFINED,  
> > SDL\_WINDOWPOS\_UNDEFINED, w, h, 0);  
> > SDL\_Renderer \* renderer = SDL\_CreateRenderer(window, -1, 0);  
> > SDL\_Log(“+++++ INIT DONE +++++”);
> > 
> > ```
> > frame = avcodec_alloc_frame();
> > 
> > while (av_read_frame(format_context,&avpacket)>= 0) {
> > if (avpacket.stream_index == videostream) {
> > 
> > ```

> > ```
> > avcodec_decode_video2(codec_context, frame,
> > 
> > ```
> > 
> > &frame\_finished,&avpacket);
> > 
> > ```
> > }
> > 
> > while (!done) {
> > SDL_WaitEvent(&event);
> > 
> > switch(event.type) {
> > //case FF_ALLOC_EVENT:
> > //alloc_picture(event.user.data1);
> > // break;
> > //case FF_REFRESH_EVENT:
> > //video_refresh_timer(event.user.data1);
> > // break;
> > case SDL_KEYDOWN:
> > case SDL_KEYUP:
> > //case SDL_MOUSEBUTTONDOWN:
> > case SDL_QUIT:
> > done=1;
> > SDL_Log("Unhandled event type=%d", event.type);
> > 
> > }
> > 
> > //Draw Image into renderer.
> > //We used to do it through SDL_DisplayYUVOverlay. How to do that with
> > 
> > ```
> > 
> > SDL2.0. Any help for this part ?
> > 
> > ```
> > SDL_RenderClear(renderer);
> > SDL_RenderPresent(renderer);
> > SDL_Delay(20);
> > }
> > 
> > }//while av_read_frame
> > 
> > SDL_Log("+++++ FINISHED +++++");
> > SDL_Quit();
> > 
> > ```
> > 
> > }
> > 
> > * * *
> > 
> > SDL mailing list  
> > SDL at [lists.libsdl.org](http://lists.libsdl.org)  
> > [http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org](http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org)

---

<div class="post-metadata">

**Author:** ![Gabriele\_Greco](https://discourse.libsdl.org/letter_avatar_proxy/v4/letter/g/8edcca/32.png) [@Gabriele\_Greco](https://discourse.libsdl.org/u/Gabriele_Greco)\
**Post date:** [March 9, 2012, 3:15pm UTC](https://discourse.libsdl.org/t/sdl-2-0-and-ffmpeg/18809/4 "2012-03-09T15:15:41Z")

</div>

> Thanks for the link.  
> I’m reading a video streaming (not a single BMp image) so not sure how  
> this applies to my code.  
> ffmpeg reads in in yuvp42 format (see output from ffmpeg dump\_format)
> 
> > ~/workspace/test1/Release $ ./test2 toto.mpeg  
> > Input #0, mpeg, from ‘toto.mpeg’:  
> > Duration: 00:00:04.83, start: 1.000000, bitrate: 606 kb/s  
> > Stream #0:0[0x1e0]: Video: mpeg1video, yuv420p, 352x288 [SAR 178:163  
> > DAR 1958:1467], 104857 kb/s, 60 fps, 60 tbr, 90k tbn, 60 tbc
> 
> So the question is do i need to create a surface from each image read  
> frame ? How would you do that ?

You should create only a streaming texture, with a command like this:

texture = SDL\_CreateTexture(renderer, SDL\_PIXELFORMAT\_YV12,  
SDL\_TEXTUREACCESS\_STREAMING, frame\_width, frame\_height))

Then you should copy the avpicture you get from ffmpeg with a function  
similat to this one (was a test with an older 1.3 version of SDL I don’t  
know if it still works with recent 2.0 builds):

void display\_picture(AVPicture \*picture)  
{  
unsigned char \*pixels;  
int pitch;

```
if(texture && !SDL_LockTexture(texture, NULL, (void **)&pixels, &pitch)

```

) {

```
    if(pitch == picture->linesize[0]) {
        int size = pitch * frame_height;

        memcpy(pixels, picture->data[0], size);
        memcpy(pixels + size, picture->data[2], size / 4);
        memcpy(pixels + size * 5 / 4, picture->data[1], size / 4);
    }
    else {
        register unsigned char *y1,*y2,*y3,*i1,*i2,*i3;
        int i;
        y1 = pixels;
        y3 = pixels + pitch * frame_height; // invertiti xche' avevo i

```

colori sballati!  
y2 = pixels + pitch \* frame\_height \* 5 / 4;

```
        i1=picture->data[0];
        i2=picture->data[1];
        i3=picture->data[2];

        for (i = 0; i<(frame_height/2); i++) {
            memcpy(y1,i1,pitch);
            i1+=picture->linesize[0];
            y1+=pitch;
            memcpy(y1,i1,pitch);

            memcpy(y2,i2,pitch / 2);
            memcpy(y3,i3,pitch / 2);

            y1+=pitch;
            y2+=pitch / 2;
            y3+=pitch / 2;
            i1+=picture->linesize[0];
            i2+=picture->linesize[1];
            i3+=picture->linesize[2];
        }
    }
    SDL_UnlockTexture(texture);
}

SDL_RenderCopy(renderer, texture, NULL, &rect);
SDL_RenderPresent(renderer);

```

}On Fri, Mar 9, 2012 at 3:02 PM, David martin wrote:

–  
Bye,  
Gabry

---

<div class="post-metadata">

**Author:** ![David\_martin1](https://discourse.libsdl.org/letter_avatar_proxy/v4/letter/d/ed8c4c/32.png) [@David\_martin1](https://discourse.libsdl.org/u/David_martin1)\
**Post date:** [March 9, 2012, 3:55pm UTC](https://discourse.libsdl.org/t/sdl-2-0-and-ffmpeg/18809/5 "2012-03-09T15:55:19Z")

</div>

The problem is that in SDL1.2 picture relies on the function  
SDL\_LockYUVOverlay() that is not implemented in 2.0.  
How to cope with that ?On 03/09/2012 04:15 PM, Gabriele Greco wrote:

> On Fri, Mar 9, 2012 at 3:02 PM, David martin \<@David\_martin1 [mailto:David\_martin1](mailto:David_martin1)\> wrote:
> 
> ```
> Thanks for the link.
> I'm reading a video streaming (not a single BMp image) so not sure
> how this applies to my code.
> ffmpeg reads in in yuvp42 format (see output from ffmpeg dump_format)
> 
> > ~/workspace/test1/Release $ ./test2 toto.mpeg
> Input #0, mpeg, from 'toto.mpeg':
> Duration: 00:00:04.83, start: 1.000000, bitrate: 606 kb/s
> Stream #0:0[0x1e0]: Video: mpeg1video, yuv420p, 352x288 [SAR
> 178:163 DAR 1958:1467], 104857 kb/s, 60 fps, 60 tbr, 90k tbn, 60 tbc
> 
> So the question is do i need to create a surface from each image
> read frame ? How would you do that ?
> 
> ```
> 
> You should create only a streaming texture, with a command like this:
> 
> texture = SDL\_CreateTexture(renderer, SDL\_PIXELFORMAT\_YV12,  
> SDL\_TEXTUREACCESS\_STREAMING, frame\_width, frame\_height))
> 
> Then you should copy the avpicture you get from ffmpeg with a function  
> similat to this one (was a test with an older 1.3 version of SDL I don’t  
> know if it still works with recent 2.0 builds):  
> void display\_picture(AVPicture \*picture)  
> {  
> unsigned char \*pixels;  
> int pitch;
> 
> ```
> if(texture && !SDL_LockTexture(texture, NULL, (void **)&pixels,
> 
> ```
> 
> &pitch) ) {
> 
> ```
> if(pitch == picture->linesize[0]) {
> int size = pitch * frame_height;
> 
> memcpy(pixels, picture->data[0], size);
> memcpy(pixels + size, picture->data[2], size / 4);
> memcpy(pixels + size * 5 / 4, picture->data[1], size / 4);
> }
> else {
> register unsigned char *y1,*y2,*y3,*i1,*i2,*i3;
> int i;
> y1 = pixels;
> y3 = pixels + pitch * frame_height; // invertiti xche'
> 
> ```
> 
> avevo i colori sballati!  
> y2 = pixels + pitch \* frame\_height \* 5 / 4;
> 
> ```
> i1=picture->data[0];
> i2=picture->data[1];
> i3=picture->data[2];
> 
> for (i = 0; i<(frame_height/2); i++) {
> memcpy(y1,i1,pitch);
> i1+=picture->linesize[0];
> y1+=pitch;
> memcpy(y1,i1,pitch);
> 
> memcpy(y2,i2,pitch / 2);
> memcpy(y3,i3,pitch / 2);
> 
> y1+=pitch;
> y2+=pitch / 2;
> y3+=pitch / 2;
> i1+=picture->linesize[0];
> i2+=picture->linesize[1];
> i3+=picture->linesize[2];
> }
> }
> SDL_UnlockTexture(texture);
> }
> 
> SDL_RenderCopy(renderer, texture, NULL, &rect);
> SDL_RenderPresent(renderer);
> 
> ```
> 
> }
> 
> –  
> Bye,  
> Gabry
> 
> * * *
> 
> SDL mailing list  
> SDL at [lists.libsdl.org](http://lists.libsdl.org)  
> [http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org](http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org)

---

<div class="post-metadata">

**Author:** ![David\_martin1](https://discourse.libsdl.org/letter_avatar_proxy/v4/letter/d/ed8c4c/32.png) [@David\_martin1](https://discourse.libsdl.org/u/David_martin1)\
**Post date:** [March 9, 2012, 4:00pm UTC](https://discourse.libsdl.org/t/sdl-2-0-and-ffmpeg/18809/6 "2012-03-09T16:00:42Z")

</div>

From the link that dimitris posted it seems that the problems of  
SDL\_LockYUVOverlay(bmp) can be fixed by using a surface implementation.

image = SDL\_LoadBMP(argv[1]);  
if (!image) {  
fprintf(stderr, “Couldn’t load BMP file %s: %s\n”, argv[1],  
SDL\_GetError());  
quit(3);  
}  
if (image-\>format-\>BytesPerPixel != 3) {  
fprintf(stderr, “BMP must be 24-bit\n”);  
quit(4);  
}

```
 /* wxh for the V plane, and then w/2xh/2 for the U and V planes */
 imageYUV = SDL_malloc(image->w*image->h+(image->w*image->h)/2);
 ConvertRGBtoYV12(image->pixels, imageYUV, image->w, image->h, 0, 100);

```

Check his link for the details on the function.  
Actually SDL\_LoadBMP reads from an image file, in case of ffmpeg i don’t  
know how to create a surface from an image stream.

Any idea ?On 03/09/2012 04:15 PM, Gabriele Greco wrote:

> On Fri, Mar 9, 2012 at 3:02 PM, David martin \<@David\_martin1 [mailto:David\_martin1](mailto:David_martin1)\> wrote:
> 
> ```
> Thanks for the link.
> I'm reading a video streaming (not a single BMp image) so not sure
> how this applies to my code.
> ffmpeg reads in in yuvp42 format (see output from ffmpeg dump_format)
> 
> > ~/workspace/test1/Release $ ./test2 toto.mpeg
> Input #0, mpeg, from 'toto.mpeg':
> Duration: 00:00:04.83, start: 1.000000, bitrate: 606 kb/s
> Stream #0:0[0x1e0]: Video: mpeg1video, yuv420p, 352x288 [SAR
> 178:163 DAR 1958:1467], 104857 kb/s, 60 fps, 60 tbr, 90k tbn, 60 tbc
> 
> So the question is do i need to create a surface from each image
> read frame ? How would you do that ?
> 
> ```
> 
> You should create only a streaming texture, with a command like this:
> 
> texture = SDL\_CreateTexture(renderer, SDL\_PIXELFORMAT\_YV12,  
> SDL\_TEXTUREACCESS\_STREAMING, frame\_width, frame\_height))
> 
> Then you should copy the avpicture you get from ffmpeg with a function  
> similat to this one (was a test with an older 1.3 version of SDL I don’t  
> know if it still works with recent 2.0 builds):  
> void display\_picture(AVPicture \*picture)  
> {  
> unsigned char \*pixels;  
> int pitch;
> 
> ```
> if(texture && !SDL_LockTexture(texture, NULL, (void **)&pixels,
> 
> ```
> 
> &pitch) ) {
> 
> ```
> if(pitch == picture->linesize[0]) {
> int size = pitch * frame_height;
> 
> memcpy(pixels, picture->data[0], size);
> memcpy(pixels + size, picture->data[2], size / 4);
> memcpy(pixels + size * 5 / 4, picture->data[1], size / 4);
> }
> else {
> register unsigned char *y1,*y2,*y3,*i1,*i2,*i3;
> int i;
> y1 = pixels;
> y3 = pixels + pitch * frame_height; // invertiti xche'
> 
> ```
> 
> avevo i colori sballati!  
> y2 = pixels + pitch \* frame\_height \* 5 / 4;
> 
> ```
> i1=picture->data[0];
> i2=picture->data[1];
> i3=picture->data[2];
> 
> for (i = 0; i<(frame_height/2); i++) {
> memcpy(y1,i1,pitch);
> i1+=picture->linesize[0];
> y1+=pitch;
> memcpy(y1,i1,pitch);
> 
> memcpy(y2,i2,pitch / 2);
> memcpy(y3,i3,pitch / 2);
> 
> y1+=pitch;
> y2+=pitch / 2;
> y3+=pitch / 2;
> i1+=picture->linesize[0];
> i2+=picture->linesize[1];
> i3+=picture->linesize[2];
> }
> }
> SDL_UnlockTexture(texture);
> }
> 
> SDL_RenderCopy(renderer, texture, NULL, &rect);
> SDL_RenderPresent(renderer);
> 
> ```
> 
> }
> 
> –  
> Bye,  
> Gabry
> 
> * * *
> 
> SDL mailing list  
> SDL at [lists.libsdl.org](http://lists.libsdl.org)  
> [http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org](http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org)

---

<div class="post-metadata">

**Author:** ![David\_martin1](https://discourse.libsdl.org/letter_avatar_proxy/v4/letter/d/ed8c4c/32.png) [@David\_martin1](https://discourse.libsdl.org/u/David_martin1)\
**Post date:** [March 12, 2012, 1:39pm UTC](https://discourse.libsdl.org/t/sdl-2-0-and-ffmpeg/18809/7 "2012-03-12T13:39:30Z")

</div>

I have managed to make some progress on the ffmpeg with SDL2.0.  
Here is the code that is compiling and running, although can’t display  
the image for some reason that i don’t know. i was wondering if you  
could run it on your machines and let me know what the image is not  
displayed.  
You can copy the whole code and run it. For some reason the renderer  
does not display the image. I would appreciate if somebody could have a  
look as it is almost done. Note that the code is not optimized yet, just  
wanted to have a functional code first…

Runit with a mpeg video for example like:  
./test2 file.mpeg

thanks

/\*

- test2.c\*

\*/  
#include \<libavcodec/avcodec.h\>  
#include \<libavformat/avformat.h\>  
#include \<libswscale/swscale.h\>  
#include \<SDL.h\>  
#include \<stdio.h\>

void sdl\_init(AVFormatContext\* format\_context, AVCodecContext\*  
codec\_context, int videostream ) {  
SDL\_Init(SDL\_INIT\_EVERYTHING);  
AVFrame\* frame;  
AVFrame \*pFrameRGB;  
AVPacket avpacket;  
AVPicture picture;

```
 // Convert the image from its native format to RGB
 struct SwsContext *img_convert_ctx =NULL;

 SDL_Event event;
 SDL_RendererInfo info;
 SDL_Texture *texture;
 SDL_Surface *image;
 Uint32 then, now , frames;
 int w = 640;
 int h = 480;
 int done = 0;
 SDL_Window * window = SDL_CreateWindow("SDL", 

```

SDL\_WINDOWPOS\_UNDEFINED, SDL\_WINDOWPOS\_UNDEFINED, w, h, 0);  
SDL\_Renderer \* renderer = SDL\_CreateRenderer(window, -1, 0);

```
 SDL_GetRendererInfo(renderer, &info);
 printf("Using %s rendering\n", info.name);

 SDL_Log("+++++ INIT DONE +++++");

 frame = avcodec_alloc_frame();

 pFrameRGB = avcodec_alloc_frame();
     if (pFrameRGB == NULL)
     {
      printf("Cannot allocate pFrame\n");
      exit(-1);
     }

     unsigned char* pixels;
     int pitch;
     uint8_t *buffer;
     int numBytes;
     // Determine required buffer size and allocate buffer
     numBytes=avpicture_get_size(PIX_FMT_RGB24, codec_context->width,
     		codec_context->height);
     buffer=(uint8_t *)av_malloc(numBytes*sizeof(uint8_t));

     // Assign appropriate parts of buffer to image planes in pFrameRGB
     // Note that pFrameRGB is an AVFrame, but AVFrame is a superset
     // of AVPicture
     avpicture_fill((AVPicture *)pFrameRGB, buffer, PIX_FMT_RGB24,
     		codec_context->width, codec_context->height);

     while (av_read_frame(format_context, &avpacket) >= 0) {
     	if (avpacket.stream_index == videostream) {
     		// Video stream packet
     		int frame_finished;

     		avcodec_decode_video2(codec_context, frame, &frame_finished, 

```

&avpacket);

```
     		if(frame_finished)
     		{

     			picture.data[0] = frame->data[0];
     			picture.data[1] = frame->data[1];
     			picture.data[2] = frame->data[2];
     			picture.linesize[0] = frame->linesize[0];
     			picture.linesize[1] = frame->linesize[1];
     			picture.linesize[2] = frame->linesize[2];

     			img_convert_ctx = sws_getCachedContext(img_convert_ctx,
     					codec_context->width,
     					codec_context->height,
     					PIX_FMT_YUV420P,
     					frame->width,
     					frame->height,
     					PIX_FMT_RGB24,
     					SWS_BICUBIC,
     					NULL,
     					NULL,
     					NULL);

     			if (img_convert_ctx == NULL)
     			{
     				fprintf(stderr, "Cannot initialize the conversion context!\n");
     				exit(1);
     			}

     			int s = sws_scale(
     					img_convert_ctx,
     					frame->data,
     					frame->linesize,
     					0,
     					codec_context->height,
     					picture.data,
     					picture.linesize);

     			texture = SDL_CreateTexture(renderer, SDL_PIXELFORMAT_YV12, 

```

SDL\_TEXTUREACCESS\_STREAMING, frame-\>width, frame-\>height);//3 plane texture

```
     			if (!texture) {
     				fprintf(stderr, "Couldn't set create texture: %s\n", 

```

SDL\_GetError());  
exit(-1);  
}

```
						if(texture && !SDL_LockTexture(texture, NULL, (void **)&pixels, 

```

&pitch) ) {

```
							if(pitch == picture.linesize[0]) {
								int size = pitch * frame->height;
								memcpy(pixels, picture.data[0], size);
								memcpy(pixels + size, picture.data[2], size / 4);
								memcpy(pixels + size * 5 / 4, picture.data[1], size / 4);
							}

							else {
								register unsigned char *y1,*y2,*y3,*i1,*i2,*i3;
								int i;
								y1 = pixels;
								y3 = pixels + pitch * frame->height; // invertiti xche' avevo i 

```

colori sballati!  
y2 = pixels + pitch \* frame-\>height \* 5 / 4;

```
								i1=picture.data[0];
								i2=picture.data[1];
								i3=picture.data[2];

								for (i = 0; i<(frame->height/2); i++) {
									memcpy(y1,i1,pitch);
									i1+=picture.linesize[0];
									y1+=pitch;
									memcpy(y1,i1,pitch);

									memcpy(y2,i2,pitch / 2);
									memcpy(y3,i3,pitch / 2);

									y1+=pitch;
									y2+=pitch / 2;
									y3+=pitch / 2;
									i1+=picture.linesize[0];
									i2+=picture.linesize[1];
									i3+=picture.linesize[2];
								}
							}
						}//texture

     		}//frame
     av_free_packet(&avpacket);
     	}//avpacket
 frames = 0;
 then = SDL_GetTicks();

// while (!done) {
 	SDL_PollEvent(&event);

 	 switch(event.type) {

            if (event.key.keysym.sym == SDLK_ESCAPE) {
           // done = 1;
             }
            break;

        case SDL_QUIT:
        	//done=1;
         SDL_Log("Unhandled event type=%d", event.type);
         break;
    // }
 	 /* Print out some timing information */
 	   // printf("%d\n",avpacket.dts);
     if (!SDL_UpdateTexture(texture, NULL, frame, frame->width)) {
         //SDL_UpdateTexture(texture, NULL, imageYUV, image->w);

             fprintf(stderr, "Couldn't update texture: %s\n", 

```

SDL\_GetError());  
exit(-1);  
}

```
 	      SDL_RenderCopy(renderer, texture, NULL, NULL);
 	      SDL_RenderPresent(renderer);
 	      SDL_Delay(20);
     		}

  }//while av_read_frame

 SDL_Log("+++++ FINISHED +++++");
 SDL_Quit();

```

}

int main(int argc, char \* argv) {

```
AVCodecContext* codec_context;
int videostream;

if (argc < 2) {
		printf("Usage: %s filename\n", argv[0]);
		return 0;
	}

	// Register all available file formats and codecs
	av_register_all();

	int err;
	// Init SDL with video support
	err = SDL_Init(SDL_INIT_VIDEO);
	if (err < 0) {
		fprintf(stderr, "Unable to init SDL: %s\n", SDL_GetError());
		return -1;
	}

	// Open video file
	const char* filename = argv[1];
	AVFormatContext* format_context = NULL;
	err = avformat_open_input(&format_context, filename, NULL, NULL);
	if (err < 0) {
		fprintf(stderr, "ffmpeg: Unable to open input\n");
		return -1;
	}

	// Retrieve stream information
	err = avformat_find_stream_info(format_context, NULL);
	if (err < 0) {
		fprintf(stderr, "ffmpeg: Unable to find stream info\n");
		return -1;
	}

	// Dump information about file onto standard error
	av_dump_format(format_context, 0, argv[1], 0);

	// Find the first video stream

	for (videostream = 0; videostream < format_context->nb_streams; 

```

++videostream) {  
if (format\_context-\>streams[videostream]-\>codec-\>codec\_type ==  
AVMEDIA\_TYPE\_VIDEO) {  
break;  
}  
}  
if (videostream == format\_context-\>nb\_streams) {  
fprintf(stderr, “ffmpeg: Unable to find video stream\n”);  
return -1;  
}

```
	codec_context = format_context->streams[videostream]->codec;
	AVCodec* codec = avcodec_find_decoder(codec_context->codec_id);

	avcodec_alloc_context3(codec);//FIXME
	//codec_context = avcodec_alloc_context3(codec);//FIXME

	 if (avcodec_open2(codec_context, codec, NULL) < 0)
	 {
		 fprintf(stderr, "ffmpeg: Unable to allocate codec context\n");
	 }

		else {
			printf("Codec initialized\n");

	}

	 /*
	 Initializing display
	 */
	

 sdl_init(format_context, codec_context,videostream);

 return 0;

```

}

// END OF MYSCRIPT

On 03/09/2012 11:09 AM, David martin wrote:

> Hi,  
> I’m new to SDL2, I used to play with danger ffmpeg tutorials to stream  
> video to an SDL surface.
> 
> I’m updating my program incorporating SDL2.
> 
> Here is a piece of snippet and would appreciate your help to actually  
> draw the video image into the surface using the SDL2 api.  
> thanks
> 
> …  
> AVFrame\* frame ;  
> AVPacket avpacket;  
> int frame\_finished;  
> SDL\_Event event;  
> int w = 640;  
> int h = 480;  
> int done = 0, color = 0;  
> SDL\_Window \* window = SDL\_CreateWindow(“SDL”, SDL\_WINDOWPOS\_UNDEFINED,  
> SDL\_WINDOWPOS\_UNDEFINED, w, h, 0);  
> SDL\_Renderer \* renderer = SDL\_CreateRenderer(window, -1, 0);  
> SDL\_Log(“+++++ INIT DONE +++++”);
> 
> frame = avcodec\_alloc\_frame();
> 
> while (av\_read\_frame(format\_context, &avpacket) \>= 0) {  
> if (avpacket.stream\_index == videostream) {
> 
> avcodec\_decode\_video2(codec\_context, frame, &frame\_finished, &avpacket);
> 
> }
> 
> while (!done) {  
> SDL\_WaitEvent(&event);
> 
> switch(event.type) {  
> //case FF\_ALLOC\_EVENT:  
> //alloc\_picture(event.user.data1);  
> // break;  
> //case FF\_REFRESH\_EVENT:  
> //video\_refresh\_timer(event.user.data1);  
> // break;  
> case SDL\_KEYDOWN:  
> case SDL\_KEYUP:  
> //case SDL\_MOUSEBUTTONDOWN:  
> case SDL\_QUIT:  
> done=1;  
> SDL\_Log(“Unhandled event type=%d”, event.type);
> 
> }
> 
> //Draw Image into renderer.  
> //We used to do it through SDL\_DisplayYUVOverlay. How to do that with  
> SDL2.0. Any help for this part ?
> 
> SDL\_RenderClear(renderer);  
> SDL\_RenderPresent(renderer);  
> SDL\_Delay(20);  
> }
> 
> }//while av\_read\_frame
> 
> SDL\_Log(“+++++ FINISHED +++++”);  
> SDL\_Quit();  
> }

---

<div class="post-metadata">

**Author:** ![David\_martin1](https://discourse.libsdl.org/letter_avatar_proxy/v4/letter/d/ed8c4c/32.png) [@David\_martin1](https://discourse.libsdl.org/u/David_martin1)\
**Post date:** [March 12, 2012, 2:13pm UTC](https://discourse.libsdl.org/t/sdl-2-0-and-ffmpeg/18809/8 "2012-03-12T14:13:47Z")

</div>

Sorry had a wrong loop on previous file. This should do it.  
thanks  
/\*

- test2.c\*
- Created on: Mar 9, 2012
- 

```
 Author: dvi

```

\*/  
#include \<libavcodec/avcodec.h\>  
#include \<libavformat/avformat.h\>  
#include \<libswscale/swscale.h\>  
#include \<SDL.h\>  
#include \<stdio.h\>

void sdl\_init(AVFormatContext\* format\_context, AVCodecContext\*  
codec\_context, int videostream ) {  
SDL\_Init(SDL\_INIT\_EVERYTHING);  
AVFrame\* frame;  
AVFrame \*pFrameRGB;  
AVPacket avpacket;  
AVPicture picture;

```
 // Convert the image from its native format to RGB
 struct SwsContext *img_convert_ctx =NULL;

 SDL_Event event;
 SDL_RendererInfo info;
 SDL_Texture *texture;
 SDL_Surface *image;
 Uint32 then, now , frames;
 int w = 640;
 int h = 480;
 int done = 0;
 SDL_Window * window = SDL_CreateWindow("SDL", 

```

SDL\_WINDOWPOS\_UNDEFINED, SDL\_WINDOWPOS\_UNDEFINED, w, h, 0);  
SDL\_Renderer \* renderer = SDL\_CreateRenderer(window, -1, 0);

```
 SDL_GetRendererInfo(renderer, &info);
 printf("Using %s rendering\n", info.name);

 SDL_Log("+++++ INIT DONE +++++");

 frame = avcodec_alloc_frame();

 pFrameRGB = avcodec_alloc_frame();
     if (pFrameRGB == NULL)
     {
      printf("Cannot allocate pFrame\n");
      exit(-1);
     }

     unsigned char* pixels;
     int pitch;
     uint8_t *buffer;
     int numBytes;
     // Determine required buffer size and allocate buffer
     numBytes=avpicture_get_size(PIX_FMT_RGB24, codec_context->width,
     		codec_context->height);
     buffer=(uint8_t *)av_malloc(numBytes*sizeof(uint8_t));

     // Assign appropriate parts of buffer to image planes in pFrameRGB
     // Note that pFrameRGB is an AVFrame, but AVFrame is a superset
     // of AVPicture
     avpicture_fill((AVPicture *)pFrameRGB, buffer, PIX_FMT_RGB24,
     		codec_context->width, codec_context->height);

     while (av_read_frame(format_context, &avpacket) >= 0) {
     	if (avpacket.stream_index == videostream) {
     		// Video stream packet
     		int frame_finished;

     		avcodec_decode_video2(codec_context, frame, &frame_finished, 

```

&avpacket);

```
     		if(frame_finished)
     		{

     			picture.data[0] = frame->data[0];
     			picture.data[1] = frame->data[1];
     			picture.data[2] = frame->data[2];
     			picture.linesize[0] = frame->linesize[0];
     			picture.linesize[1] = frame->linesize[1];
     			picture.linesize[2] = frame->linesize[2];

     			img_convert_ctx = sws_getCachedContext(img_convert_ctx,
     					codec_context->width,
     					codec_context->height,
     					PIX_FMT_YUV420P,
     					frame->width,
     					frame->height,
     					PIX_FMT_RGB24,
     					SWS_BICUBIC,
     					NULL,
     					NULL,
     					NULL);

     			if (img_convert_ctx == NULL)
     			{
     				fprintf(stderr, "Cannot initialize the conversion context!\n");
     				exit(1);
     			}

     			int s = sws_scale(
     					img_convert_ctx,
     					frame->data,
     					frame->linesize,
     					0,
     					codec_context->height,
     					picture.data,
     					picture.linesize);

     			texture = SDL_CreateTexture(renderer, SDL_PIXELFORMAT_YV12, 

```

SDL\_TEXTUREACCESS\_STREAMING, codec\_context-\>width,  
codec\_context-\>height);//3 plane texture

```
     			if (!texture) {
     				fprintf(stderr, "Couldn't set create texture: %s\n", 

```

SDL\_GetError());  
exit(-1);  
}

```
						if(texture && !SDL_LockTexture(texture, NULL, (void **)&pixels, 

```

&pitch) ) {

```
							if(pitch == picture.linesize[0]) {
								int size = pitch * frame->height;
								memcpy(pixels, picture.data[0], size);
								memcpy(pixels + size, picture.data[2], size / 4);
								memcpy(pixels + size * 5 / 4, picture.data[1], size / 4);
							}

							else {
								register unsigned char *y1,*y2,*y3,*i1,*i2,*i3;
								int i;
								y1 = pixels;
								y3 = pixels + pitch * frame->height; // invertiti xche' avevo i 

```

colori sballati!  
y2 = pixels + pitch \* frame-\>height \* 5 / 4;

```
								i1=picture.data[0];
								i2=picture.data[1];
								i3=picture.data[2];

								for (i = 0; i<(frame->height/2); i++) {
									memcpy(y1,i1,pitch);
									i1+=picture.linesize[0];
									y1+=pitch;
									memcpy(y1,i1,pitch);

									memcpy(y2,i2,pitch / 2);
									memcpy(y3,i3,pitch / 2);

									y1+=pitch;
									y2+=pitch / 2;
									y3+=pitch / 2;
									i1+=picture.linesize[0];
									i2+=picture.linesize[1];
									i3+=picture.linesize[2];
								}
							}
						}//texture
						SDL_UnlockTexture(texture);
     		}//frame
     av_free_packet(&avpacket);
     	}//avpacket

 //while (!done) {
 	SDL_PollEvent(&event);

 	 switch(event.type) {

            if (event.key.keysym.sym == SDLK_ESCAPE) {
             done = 1;
            }
            break;

        case SDL_QUIT:
        	//done=1;
         SDL_Log("Unhandled event type=%d", event.type);
         done=1;
         break;
 	 }
 	 /* Print out some timing information */
 	  // printf("TIMING %d\n",avpacket.dts);
 	 //SDL_UpdateTexture(texture, NULL, imageYUV, image->w)
     if (!SDL_UpdateTexture(texture, NULL, frame, frame->width)) {
         //SDL_UpdateTexture(texture, NULL, imageYUV, image->w);

             fprintf(stderr, "Couldn't update texture: %s\n", 

```

SDL\_GetError());  
//exit(-1);  
}

```
     SDL_RenderClear(renderer);
     SDL_RenderCopy(renderer, texture, NULL, NULL);
     SDL_RenderPresent(renderer);
     SDL_Delay(20);
 //}

```

// }

```
  }//while av_read_frame

 SDL_Log("+++++ FINISHED +++++");
 SDL_Quit();

```

}

int main(int argc, char \* argv) {

```
AVCodecContext* codec_context;
int videostream;

if (argc < 2) {
		printf("Usage: %s filename\n", argv[0]);
		return 0;
	}

	// Register all available file formats and codecs
	av_register_all();

	int err;
	// Init SDL with video support
	err = SDL_Init(SDL_INIT_VIDEO);
	if (err < 0) {
		fprintf(stderr, "Unable to init SDL: %s\n", SDL_GetError());
		return -1;
	}

	// Open video file
	const char* filename = argv[1];
	AVFormatContext* format_context = NULL;
	err = avformat_open_input(&format_context, filename, NULL, NULL);
	if (err < 0) {
		fprintf(stderr, "ffmpeg: Unable to open input\n");
		return -1;
	}

	// Retrieve stream information
	err = avformat_find_stream_info(format_context, NULL);
	if (err < 0) {
		fprintf(stderr, "ffmpeg: Unable to find stream info\n");
		return -1;
	}

	// Dump information about file onto standard error
	av_dump_format(format_context, 0, argv[1], 0);

	// Find the first video stream

	for (videostream = 0; videostream < format_context->nb_streams; 

```

++videostream) {  
if (format\_context-\>streams[videostream]-\>codec-\>codec\_type ==  
AVMEDIA\_TYPE\_VIDEO) {  
break;  
}  
}  
if (videostream == format\_context-\>nb\_streams) {  
fprintf(stderr, “ffmpeg: Unable to find video stream\n”);  
return -1;  
}

```
	codec_context = format_context->streams[videostream]->codec;
	AVCodec* codec = avcodec_find_decoder(codec_context->codec_id);

	avcodec_alloc_context3(codec);

	 if (avcodec_open2(codec_context, codec, NULL) < 0)
	 {
		 fprintf(stderr, "ffmpeg: Unable to allocate codec context\n");
	 }

		else {
			printf("Codec initialized\n");

	}

	 /*
	 Initializing display
	 */
	 printf("Width:%d\n",codec_context->width);
	 printf("height:%d\n",codec_context->height);
	 //exit(0);

 sdl_init(format_context, codec_context,videostream);

 return 0;

```

}

On 03/09/2012 11:09 AM, David martin wrote:

> Hi,  
> I’m new to SDL2, I used to play with danger ffmpeg tutorials to stream  
> video to an SDL surface.
> 
> I’m updating my program incorporating SDL2.
> 
> Here is a piece of snippet and would appreciate your help to actually  
> draw the video image into the surface using the SDL2 api.  
> thanks
> 
> …  
> AVFrame\* frame ;  
> AVPacket avpacket;  
> int frame\_finished;  
> SDL\_Event event;  
> int w = 640;  
> int h = 480;  
> int done = 0, color = 0;  
> SDL\_Window \* window = SDL\_CreateWindow(“SDL”, SDL\_WINDOWPOS\_UNDEFINED,  
> SDL\_WINDOWPOS\_UNDEFINED, w, h, 0);  
> SDL\_Renderer \* renderer = SDL\_CreateRenderer(window, -1, 0);  
> SDL\_Log(“+++++ INIT DONE +++++”);
> 
> frame = avcodec\_alloc\_frame();
> 
> while (av\_read\_frame(format\_context, &avpacket) \>= 0) {  
> if (avpacket.stream\_index == videostream) {
> 
> avcodec\_decode\_video2(codec\_context, frame, &frame\_finished, &avpacket);
> 
> }
> 
> while (!done) {  
> SDL\_WaitEvent(&event);
> 
> switch(event.type) {  
> //case FF\_ALLOC\_EVENT:  
> //alloc\_picture(event.user.data1);  
> // break;  
> //case FF\_REFRESH\_EVENT:  
> //video\_refresh\_timer(event.user.data1);  
> // break;  
> case SDL\_KEYDOWN:  
> case SDL\_KEYUP:  
> //case SDL\_MOUSEBUTTONDOWN:  
> case SDL\_QUIT:  
> done=1;  
> SDL\_Log(“Unhandled event type=%d”, event.type);
> 
> }
> 
> //Draw Image into renderer.  
> //We used to do it through SDL\_DisplayYUVOverlay. How to do that with  
> SDL2.0. Any help for this part ?
> 
> SDL\_RenderClear(renderer);  
> SDL\_RenderPresent(renderer);  
> SDL\_Delay(20);  
> }
> 
> }//while av\_read\_frame
> 
> SDL\_Log(“+++++ FINISHED +++++”);  
> SDL\_Quit();  
> }

---

<div class="post-metadata">

**Author:** ![michelleC](https://discourse.libsdl.org/letter_avatar_proxy/v4/letter/m/dec6dc/32.png) [@michelleC](https://discourse.libsdl.org/u/michelleC)\
**Post date:** [May 20, 2012, 10:30pm UTC](https://discourse.libsdl.org/t/sdl-2-0-and-ffmpeg/18809/9 "2012-05-20T22:30:57Z")

</div>

I haven’t got around to trying to make this work with newer versions of sdl yet, more interested right now in getting the ffmpeg wrapper framework working for both sdl players and non-sdl players.

But here’s a start that might interest some of you.

I suspect I will eventually make this work with sdl 2.0

right now it works with sdl 1.3 revision 5517.

> **[mooncatventures-group/RTSPPlay](https://github.com/mooncatventures-group/RTSPPlay)**
>
> RTSPPlay - Test RTSP Player framework ffplay.framework

---

<div class="post-metadata">

**Author:** ![Simone\_Cercignani](https://discourse.libsdl.org/letter_avatar_proxy/v4/letter/s/dc4da7/32.png) [@Simone\_Cercignani](https://discourse.libsdl.org/u/Simone_Cercignani)\
**Post date:** [October 9, 2013, 9:50am UTC](https://discourse.libsdl.org/t/sdl-2-0-and-ffmpeg/18809/10 "2013-10-09T09:50:47Z")

</div>

Hi,

anyone has solved this issue?

[SDL] SDl 2.0 and ffmpeg

_David martin_ vilanew at [gmail.com](http://gmail.com)  
[mailto:sdl%40lists.libsdl.org?Subject=Re%3A%20[SDL]%20SDl%202.0%20and%20ffmpeg&In-Reply-To=\<jjl0ar%24bsp%242%40dough.gmane.org\>](mailto:sdl%40lists.libsdl.org?Subject=Re%3A%20%5BSDL%5D%20SDl%202.0%20and%20ffmpeg&In-Reply-To=%3Cjjl0ar%24bsp%242%40dough.gmane.org%3E)  
/Mon Mar 12 07:13:47 PDT 2012/

- Previous message: [SDL] SDl 2.0 and ffmpeg  
[http://lists.libsdl.org/pipermail/sdl-libsdl.org/2012-March/084108.html](http://lists.libsdl.org/pipermail/sdl-libsdl.org/2012-March/084108.html)
- Next message: [SDL] SDL2 Graphics API, was Re: compiling with  
SDL\_NO\_COMPAT  
[http://lists.libsdl.org/pipermail/sdl-libsdl.org/2012-March/084091.html](http://lists.libsdl.org/pipermail/sdl-libsdl.org/2012-March/084091.html)
- _Messages sorted by:_ [date]  
[http://lists.libsdl.org/pipermail/sdl-libsdl.org/2012-March/date.html#84110](http://lists.libsdl.org/pipermail/sdl-libsdl.org/2012-March/date.html#84110)  
[thread]  
[http://lists.libsdl.org/pipermail/sdl-libsdl.org/2012-March/thread.html#84110](http://lists.libsdl.org/pipermail/sdl-libsdl.org/2012-March/thread.html#84110)  
[subject]  
[http://lists.libsdl.org/pipermail/sdl-libsdl.org/2012-March/subject.html#84110](http://lists.libsdl.org/pipermail/sdl-libsdl.org/2012-March/subject.html#84110)  
[author]  
[http://lists.libsdl.org/pipermail/sdl-libsdl.org/2012-March/author.html#84110](http://lists.libsdl.org/pipermail/sdl-libsdl.org/2012-March/author.html#84110)------------------------------------------------------------------------

Sorry had a wrong loop on previous file. This should do it.  
thanks  
/\*

- test2.c
- 
- Created on: Mar 9, 2012
- 

```
 Author: dvi

```

\*/  
#include \<libavcodec/avcodec.h\>  
#include \<libavformat/avformat.h\>  
#include \<libswscale/swscale.h\>  
#include \<SDL.h\>  
#include \<stdio.h\>

void sdl\_init(AVFormatContext\* format\_context, AVCodecContext\*  
codec\_context, int videostream ) {  
SDL\_Init(SDL\_INIT\_EVERYTHING);  
AVFrame\* frame;  
AVFrame \*pFrameRGB;  
AVPacket avpacket;  
AVPicture picture;

```
  // Convert the image from its native format to RGB
  struct SwsContext *img_convert_ctx =NULL;

  SDL_Event event;
  SDL_RendererInfo info;
  SDL_Texture *texture;
  SDL_Surface *image;
  Uint32 then, now , frames;
  int w = 640;
  int h = 480;
  int done = 0;
  SDL_Window * window = SDL_CreateWindow("SDL",

```

SDL\_WINDOWPOS\_UNDEFINED, SDL\_WINDOWPOS\_UNDEFINED, w, h, 0);  
SDL\_Renderer \* renderer = SDL\_CreateRenderer(window, -1, 0);

```
  SDL_GetRendererInfo(renderer, &info);
  printf("Using %s rendering\n", info.name);

  SDL_Log("+++++ INIT DONE +++++");

  frame = avcodec_alloc_frame();

  pFrameRGB = avcodec_alloc_frame();
      if (pFrameRGB == NULL)
      {
       printf("Cannot allocate pFrame\n");
       exit(-1);
      }

      unsigned char* pixels;
      int pitch;
      uint8_t *buffer;
      int numBytes;
      // Determine required buffer size and allocate buffer
      numBytes=avpicture_get_size(PIX_FMT_RGB24, codec_context->width,
      		codec_context->height);
      buffer=(uint8_t *)av_malloc(numBytes*sizeof(uint8_t));

      // Assign appropriate parts of buffer to image planes in pFrameRGB
      // Note that pFrameRGB is an AVFrame, but AVFrame is a superset
      // of AVPicture
      avpicture_fill((AVPicture *)pFrameRGB, buffer, PIX_FMT_RGB24,
      		codec_context->width, codec_context->height);

      while (av_read_frame(format_context, &avpacket) >= 0) {
      	if (avpacket.stream_index == videostream) {
      		// Video stream packet
      		int frame_finished;

      		avcodec_decode_video2(codec_context, frame, &frame_finished,

```

&avpacket);

```
      		if(frame_finished)
      		{

      			picture.data[0] = frame->data[0];
      			picture.data[1] = frame->data[1];
      			picture.data[2] = frame->data[2];
      			picture.linesize[0] = frame->linesize[0];
      			picture.linesize[1] = frame->linesize[1];
      			picture.linesize[2] = frame->linesize[2];

      			img_convert_ctx = sws_getCachedContext(img_convert_ctx,
      					codec_context->width,
      					codec_context->height,
      					PIX_FMT_YUV420P,
      					frame->width,
      					frame->height,
      					PIX_FMT_RGB24,
      					SWS_BICUBIC,
      					NULL,
      					NULL,
      					NULL);

      			if (img_convert_ctx == NULL)
      			{
      				fprintf(stderr, "Cannot initialize the conversion context!\n");
      				exit(1);
      			}

      			int s = sws_scale(
      					img_convert_ctx,
      					frame->data,
      					frame->linesize,
      					0,
      					codec_context->height,
      					picture.data,
      					picture.linesize);

      			texture = SDL_CreateTexture(renderer, SDL_PIXELFORMAT_YV12,

```

SDL\_TEXTUREACCESS\_STREAMING, codec\_context-\>width,  
codec\_context-\>height);//3 plane texture

```
      			if (!texture) {
      				fprintf(stderr, "Couldn't set create texture: %s\n",

```

SDL\_GetError());  
exit(-1);  
}

```
						if(texture && !SDL_LockTexture(texture, NULL, (void **)&pixels,

```

&pitch) ) {

```
							if(pitch == picture.linesize[0]) {
								int size = pitch * frame->height;
								memcpy(pixels, picture.data[0], size);
								memcpy(pixels + size, picture.data[2], size / 4);
								memcpy(pixels + size * 5 / 4, picture.data[1], size / 4);
							}

							else {
								register unsigned char *y1,*y2,*y3,*i1,*i2,*i3;
								int i;
								y1 = pixels;
								y3 = pixels + pitch * frame->height; // invertiti xche' avevo i

```

colori sballati!  
y2 = pixels + pitch \* frame-\>height \* 5 / 4;

```
								i1=picture.data[0];
								i2=picture.data[1];
								i3=picture.data[2];

								for (i = 0; i<(frame->height/2); i++) {
									memcpy(y1,i1,pitch);
									i1+=picture.linesize[0];
									y1+=pitch;
									memcpy(y1,i1,pitch);

									memcpy(y2,i2,pitch / 2);
									memcpy(y3,i3,pitch / 2);

									y1+=pitch;
									y2+=pitch / 2;
									y3+=pitch / 2;
									i1+=picture.linesize[0];
									i2+=picture.linesize[1];
									i3+=picture.linesize[2];
								}
							}
						}//texture
						SDL_UnlockTexture(texture);
      		}//frame
      av_free_packet(&avpacket);
      	}//avpacket

  //while (!done) {
  	SDL_PollEvent(&event);

  	 switch(event.type) {

             if (event.key.keysym.sym == SDLK_ESCAPE) {
              done = 1;
             }
             break;

         case SDL_QUIT:
         	//done=1;
          SDL_Log("Unhandled event type=%d", event.type);
          done=1;
          break;
  	 }
  	 /* Print out some timing information */
  	  // printf("TIMING %d\n",avpacket.dts);
  	 //SDL_UpdateTexture(texture, NULL, imageYUV, image->w)
      if (!SDL_UpdateTexture(texture, NULL, frame, frame->width)) {
          //SDL_UpdateTexture(texture, NULL, imageYUV, image->w);

              fprintf(stderr, "Couldn't update texture: %s\n",

```

SDL\_GetError());  
//exit(-1);  
}

```
      SDL_RenderClear(renderer);
      SDL_RenderCopy(renderer, texture, NULL, NULL);
      SDL_RenderPresent(renderer);
      SDL_Delay(20);
  //}

```

// }

```
   }//while av_read_frame

  SDL_Log("+++++ FINISHED +++++");
  SDL_Quit();

```

}

int main(int argc, char \* argv) {

```
AVCodecContext* codec_context;
int videostream;

if (argc < 2) {
		printf("Usage: %s filename\n", argv[0]);
		return 0;
	}

	// Register all available file formats and codecs
	av_register_all();

	int err;
	// Init SDL with video support
	err = SDL_Init(SDL_INIT_VIDEO);
	if (err < 0) {
		fprintf(stderr, "Unable to init SDL: %s\n", SDL_GetError());
		return -1;
	}

	// Open video file
	const char* filename = argv[1];
	AVFormatContext* format_context = NULL;
	err = avformat_open_input(&format_context, filename, NULL, NULL);
	if (err < 0) {
		fprintf(stderr, "ffmpeg: Unable to open input\n");
		return -1;
	}

	// Retrieve stream information
	err = avformat_find_stream_info(format_context, NULL);
	if (err < 0) {
		fprintf(stderr, "ffmpeg: Unable to find stream info\n");
		return -1;
	}

	// Dump information about file onto standard error
	av_dump_format(format_context, 0, argv[1], 0);

	// Find the first video stream

	for (videostream = 0; videostream < format_context->nb_streams;

```

++videostream) {  
if (format\_context-\>streams[videostream]-\>codec-\>codec\_type ==  
AVMEDIA\_TYPE\_VIDEO) {  
break;  
}  
}  
if (videostream == format\_context-\>nb\_streams) {  
fprintf(stderr, “ffmpeg: Unable to find video stream\n”);  
return -1;  
}

```
	codec_context = format_context->streams[videostream]->codec;
	AVCodec* codec = avcodec_find_decoder(codec_context->codec_id);

	avcodec_alloc_context3(codec);

	 if (avcodec_open2(codec_context, codec, NULL) < 0)
	 {
		 fprintf(stderr, "ffmpeg: Unable to allocate codec context\n");
	 }

		else {
			printf("Codec initialized\n");

	}

	 /*
	 Initializing display
	 */
	 printf("Width:%d\n",codec_context->width);
	 printf("height:%d\n",codec_context->height);
	 //exit(0);

  sdl_init(format_context, codec_context,videostream);

  return 0;

```

}

On 03/09/2012 11:09 AM, David martin wrote:

> / Hi,  
> /\>/ I’m new to SDL2, I used to play with danger ffmpeg tutorials to stream  
> /\>/ video to an SDL surface.  
> /\>/  
> /\>/ I’m updating my program incorporating SDL2.  
> /\>/  
> /\>/  
> /\>/ Here is a piece of snippet and would appreciate your help to actually  
> /\>/ draw the video image into the surface using the SDL2 api.  
> /\>/ thanks  
> /\>/  
> /\>/  
> /\>/  
> /\>/ …  
> /\>/ AVFrame\* frame ;  
> /\>/ AVPacket avpacket;  
> /\>/ int frame\_finished;  
> /\>/ SDL\_Event event;  
> /\>/ int w = 640;  
> /\>/ int h = 480;  
> /\>/ int done = 0, color = 0;  
> /\>/ SDL\_Window \* window = SDL\_CreateWindow(“SDL”, SDL\_WINDOWPOS\_UNDEFINED,  
> /\>/ SDL\_WINDOWPOS\_UNDEFINED, w, h, 0);  
> /\>/ SDL\_Renderer \* renderer = SDL\_CreateRenderer(window, -1, 0);  
> /\>/ SDL\_Log(“+++++ INIT DONE +++++”);  
> /\>/  
> /\>/ frame = avcodec\_alloc\_frame();  
> /\>/  
> /\>/ while (av\_read\_frame(format\_context, &avpacket) \>= 0) {  
> /\>/ if (avpacket.stream\_index == videostream) {  
> /\>/  
> /\>/ avcodec\_decode\_video2(codec\_context, frame, &frame\_finished, &avpacket);  
> /\>/  
> /\>/  
> /\>/  
> /\>/ }  
> /\>/  
> /\>/  
> /\>/  
> /\>/ while (!done) {  
> /\>/ SDL\_WaitEvent(&event);  
> /\>/  
> /\>/ switch(event.type) {  
> /\>/ //case FF\_ALLOC\_EVENT:  
> /\>/ //alloc\_picture(event.user.data1);  
> /\>/ // break;  
> /\>/ //case FF\_REFRESH\_EVENT:  
> /\>/ //video\_refresh\_timer(event.user.data1);  
> /\>/ // break;  
> /\>/ case SDL\_KEYDOWN:  
> /\>/ case SDL\_KEYUP:  
> /\>/ //case SDL\_MOUSEBUTTONDOWN:  
> /\>/ case SDL\_QUIT:  
> /\>/ done=1;  
> /\>/ SDL\_Log(“Unhandled event type=%d”, event.type);  
> /\>/  
> /\>/ }  
> /\>/  
> /\>/ //Draw Image into renderer.  
> /\>/ //We used to do it through SDL\_DisplayYUVOverlay. How to do that with  
> /\>/ SDL2.0. Any help for this part ?  
> /\>/  
> /\>/ SDL\_RenderClear(renderer);  
> /\>/ SDL\_RenderPresent(renderer);  
> /\>/ SDL\_Delay(20);  
> /\>/ }  
> /\>/  
> /\>/ }//while av\_read\_frame  
> /\>/  
> /\>/ SDL\_Log(“+++++ FINISHED +++++”);  
> /\>/ SDL\_Quit();  
> /\>/ }  
> /

Best

Simone