TDA4VM: App_srv_camera output

qian yue

Part Number: TDA4VM

Dear

static vx_status app_run_graph(AppObj *obj)
{
    vx_status status = VX_SUCCESS;
    uint32_t num_refs, buf_id;
    int graph_parameter_num = 0;
    uint32_t iteration = 0;

    if(NULL == obj->sensor_name)
    {
        printf("sensor name is NULL \n");
        status = VX_FAILURE;
    }
    if(status == VX_SUCCESS)
    {
        status = appStartImageSensor(obj->sensor_name, 0xF);
        APP_PRINTF("Sensor start status = %d\n", status);
    }

    /* Enqueue buf for pipe up but don't trigger graph execution */
    for(buf_id=0; buf_id<NUM_BUFS; buf_id++)
    {
        graph_parameter_num = 0;
        if(status == VX_SUCCESS)
        {
            status = vxGraphParameterEnqueueReadyRef(obj->graph, graph_parameter_num,
                       (vx_reference*)&obj->capt_frames[buf_id], 1);
        }
        graph_parameter_num++;
    }
    if(obj->test_mode == 1)
    {
        for(buf_id=0; buf_id<NUM_BUFS; buf_id++)
        {
            if(status == VX_SUCCESS)
            {
                status = vxGraphParameterEnqueueReadyRef(obj->graph, obj->displayNodeGraphParamNum,
                                (vx_reference*)&obj->output_image[buf_id], 1);
            }
        }
    }

    /* wait for graph instances to complete, compare output and
     * recycle data buffers, schedule again */
    vx_uint32 actual_checksum = 0;
    while(status == VX_SUCCESS)
    {
        vx_object_array out_capture_frames;
        vx_image test_image;
        graph_parameter_num = 0;

        appPerfPointBegin(&obj->total_perf);

        /* Get output reference, waits until a frame is available */
        if(status == VX_SUCCESS)
        {
            status = vxGraphParameterDequeueDoneRef(obj->graph, graph_parameter_num,
                            (vx_reference*)&out_capture_frames, 1, &num_refs);
        }
        graph_parameter_num++;

        if(obj->test_mode == 1)
        {
            /* Get output reference, waits until a frame is available */
            if(status == VX_SUCCESS)
            {
                status = vxGraphParameterDequeueDoneRef(obj->graph, obj->displayNodeGraphParamNum,
                                                        (vx_reference*)&test_image, 1, &num_refs);
            }
            printf("test iteration: %d of %d\n", iteration, TEST_BUFFER+1);
            if( (status == VX_SUCCESS) && (iteration > TEST_BUFFER) )
            {
                if(app_test_check_image(test_image, checksums_expected[0][0], &actual_checksum) == vx_false_e)
                {
                    test_result = vx_false_e;
                }
                populate_gatherer(0, 0, actual_checksum);
                obj->stop_task = 1;
            }
            /* Get output reference, waits until a frame is available */
            if(status == VX_SUCCESS)
            {
                status = vxGraphParameterEnqueueReadyRef(obj->graph, obj->displayNodeGraphParamNum,
                                (vx_reference*)&test_image, 1);
            }
        }

        graph_parameter_num = 0;
        if(status == VX_SUCCESS)
        {
            status = vxGraphParameterEnqueueReadyRef(obj->graph, graph_parameter_num, (vx_reference*)&out_capture_frames, 1);
        }
        graph_parameter_num++;

        appPerfPointEnd(&obj->total_perf);

        if(iteration==100)
        {
            /* after first 'n' iteration reset performance stats */
            appPerfStatsResetAll();
        }

        iteration++;

        if((obj->stop_task) || (status != VX_SUCCESS))
        {
            break;
        }
    }
    /* ensure all graph processing is complete */
    vxWaitGraph(obj->graph);

    /* Dequeue buf for pipe down */
    num_refs = 0xFF;
    graph_parameter_num = 0;
    while((num_refs > 0) && (status == VX_SUCCESS))
    {
        vx_object_array out_capture_frames;
        if(status == VX_SUCCESS)
        {
            status = vxGraphParameterCheckDoneRef(obj->graph, graph_parameter_num, &num_refs);
        }
        if(num_refs > 0)
        {
            APP_PRINTF("Dequeue capture\n");
            if(status == VX_SUCCESS)
            {
                status = vxGraphParameterDequeueDoneRef(
                                            obj->graph,
                                            graph_parameter_num,
                                            (vx_reference*)&out_capture_frames,
                                            1,
                                            &num_refs);
            }
        }
    }
    num_refs = 0xFF;
    while((num_refs > 0) && (obj->test_mode == 1) && (status == VX_SUCCESS))
    {
        vx_image out_image;
        if(status == VX_SUCCESS)
        {
            status = vxGraphParameterCheckDoneRef(obj->graph, obj->displayNodeGraphParamNum, &num_refs);
        }
        if(num_refs > 0)
        {
            APP_PRINTF("Dequeue display \n");
            if(status == VX_SUCCESS)
            {
                status = vxGraphParameterDequeueDoneRef(
                                            obj->graph,
                                            obj->displayNodeGraphParamNum,
                                            (vx_reference*)&out_image,
                                            1,
                                            &num_refs);
            }
        }
    }
    return status;
}

How can I find the output stream in "app_srv_camera" in this code and convert it (RGB) to H264 output?

over 2 years ago

0 Brijesh Jadav over 2 years ago

TI__Guru**** 484365 points

Hi,

This is coming below code. Here the output image is created in RGBx format and that is used as an output image for the SRV node. The same output image is passed as input to the display node.

/* Creating objects for OpenGL node */
for(int i=0; i < NUM_BUFS; i++)
{
if(status == VX_SUCCESS)
{
obj->output_image[i] = vxCreateImage(obj->context, obj->outWidth, obj->outHeight, VX_DF_IMAGE_RGBX);
status = vxGetStatus((vx_reference) obj->output_image[i]);
}
}

obj->srv_node = tivxGlSrvNode(obj->graph, obj->param_obj, obj->viss_out_frames, obj->srv_views_array, obj->out_gpulut_array, obj->output_image[0]);
if (vxGetStatus((vx_reference)obj->srv_node) != VX_SUCCESS)
{
APP_PRINTF("srv_node create failed\n");
status = VX_FAILURE;
}

If you want to encode the SRV output, you need to first convert this input to YUV420 format. You can use DSS M2M path for this conversion. Once converted, it can be encoded on A72 using gStreamer.

Regards,

Brijesh

0 qian yue over 2 years ago in reply to Brijesh Jadav

Intellectual 981 points

Thank you very much for your reply. Can you provide a specific implementation method for using gStreamer encoding as H264 on "app_srv_camera"? We are not very clear about this aspect, we are extremely grateful!

0 qian yue over 2 years ago in reply to Brijesh Jadav

Intellectual 981 points

Dear,

Can you help me with how to perform this conversion on DSS M2M path? I want to obtain surround video stream output in "h264" format.

0 Brijesh Jadav over 2 years ago in reply to qian yue

TI__Guru**** 484365 points

Hi,

These two are different requests.

- For the first request, i would recommend referring to the existing example, app_multi_cam_codec example. This example in vision apps captures the images from the camera and then encodes it using codec. Please refer to this example for integrating codec in your application.

- For the second request, can you please refer to below faq?

https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1007456/faq-tda4vm-how-to-use-dss-m2m-for-format-conversion/4539758#4539758

Regards,

Brijesh

0 qian yue over 2 years ago in reply to Brijesh Jadav

Intellectual 981 points

Thank you for your answer. Could you please kindly inform us of the specific location of the integrated codec in "app_multi_cam_codec"? If you could provide this assistance, we would greatly appreciate it.

0 Brijesh Jadav over 2 years ago in reply to qian yue

TI__Guru**** 484365 points

You can find this example in ti-processor-sdk-rtos-j721e-evm-08_06_00_12\vision_apps\apps\basic_demos\app_multi_cam_codec\main.c file.

Regards,

Brijesh

0 qian yue over 2 years ago in reply to Brijesh Jadav

Intellectual 981 points

OK, thank you!

0 qian yue over 2 years ago in reply to Brijesh Jadav

Intellectual 981 points

Can you please guide me on how to apply the codec from " vision_apps apps basic_demos app_multi_cam_codec main. c file." to "app_srv_camera" and convert the output video stream to H264? I have been struggling in this area， Please!

0 qian yue over 2 years ago in reply to qian yue

Intellectual 981 points

I am now in a difficult situation. Can you help me make appropriate modifications in the demo regarding this aspect?

0 Brijesh Jadav over 2 years ago in reply to qian yue

TI__Guru**** 484365 points

Sorry, i can only refer to the existing application. You would require to make the changes in SRV application.

Regards,

Brijesh

0 Praveen Rao over 2 years ago in reply to Brijesh Jadav

TI__Mastermind 48833 points

Hi Qian,

We suggest you to look into the below PSDK RTOS documentation link --> https://software-dl.ti.com/jacinto7/esd/processor-sdk-rtos-jacinto7/09_00_00_02/exports/docs/vision_apps/docs/user_guide/group_apps_basic_demos_app_multi_cam_codec.html which show how to integrate encoder with the camera.

Please note that this is a reference demo application that can be modified for your need. Please note that the required demo code is available in the vision_apps/apps/basic_demos directory of PSDK RTOS.

Note that TI does not plan to integrate encoder for the SRV camera use-case and we suggest you to make the required updates refer other demo application provided.

Thanks.

Processors

Processors forum

TDA4VM: App_srv_camera output