How to change the frequency of payload generation interval in deepstream 6.2?

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

NVIDIA-SMI 525.105.17
Driver Version: 525.105.17
CUDA Version: 12.0
deepstream-6.2

±----------------------------------------------------------------------------+
| NVIDIA-SMI 525.125.06 Driver Version: 525.125.06 CUDA Version: 12.0 |
|-------------------------------±---------------------±---------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 NVIDIA GeForce … Off | 00000000:01:00.0 On | N/A |
| 0% 48C P8 35W / 370W | 3825MiB / 12288MiB | 4% Default |
| | | N/A |
±------------------------------±---------------------±---------------------+
| 1 NVIDIA GeForce … Off | 00000000:02:00.0 Off | N/A |
| 0% 46C P8 20W / 370W | 8MiB / 12288MiB | 0% Default |
| | | N/A |

| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |

As the default setting, the payload is generated and sent every 30 frames. However I want to reduce the frequency such as every 200 frames or more.

How to implement this ?

3

What do you mean payload generation interval?

Which demo did you run?

4

I m working with deepstream test5

as the default setting, the payload is generated and sent every 30 frames.

i want to increase this frame number for which payload is sent frame.

bascially i want to increase this frame number -30

how can i achieve this ?

5

You can refer to the link below to set the frame-interval for Gst-nvmsgconv plugin in the config file: https://docs.nvidia.com/metropolis/deepstream/dev-guide/text/DS_plugin_gst-nvmsgconv.html#id2.

6

do you mean i need to add frame-interval key in deepstream config file?

image

7

Yes. You can try that.

8

I added frame-interval in config file.
I tried keeping frame interval =200 , 967 etc
But there in no drop in frequency of payload generation.

[sink6]
enable=1
#Type - 1=FakeSink 2=EglSink 3=File 4=UDPSink 5=nvoverlaysink 6=MsgConvBroker
type=6
msg-conv-config=dstest5_msgconv_sample_config.txt
#(0): PAYLOAD_DEEPSTREAM - Deepstream schema payload
#(1): PAYLOAD_DEEPSTREAM_MINIMAL - Deepstream schema payload minimal
#(256): PAYLOAD_RESERVED - Reserved type
#(257): PAYLOAD_CUSTOM - Custom schema payload
msg-conv-payload-type=1
msg2p-newapi=1
frame-interval=9670

msg-broker-proto-lib=/opt/nvidia/deepstream/deepstream-6.2/lib/libnvds_kafka_proto.so
#msg-broker-proto-lib=/opt/nvidia/deepstream/deepstream-6.2/lib/libnvds_msgconv.so
#Provide your msg-broker-conn-str here
msg-broker-conn-str=192.168.0.118;9092
topic=quickstart-events
#Optional:
#msg-broker-config=…/…/deepstream-test4/cfg_kafka.txt

9

The parameter for sink filed is msg-conv-msg2p-new-api and msg-conv-frame-interval.

11

i can control the frequency of payloads generated using this parameter
msg-conv-msg2p-new-api and msg-conv-frame-interval .

But i’m getting 10 times repetiton of same payloads message with same message id.

why this is happening repetition of message with same message id ?

12

Do you mean that all the information is the same? You can also debug that by yourself, the open source path is : opt\nvidia\deepstream\deepstream\sources\libs\nvmsgconv.

13

its like 1o message in sequence comes with same messsage id , and after 10 message , message id will changed , and then this changed message id will also get 10 times repeat

basically there is a pattern of repetition of same message id 10 times

14

Could you debug that on your end? The open source path is : opt\nvidia\deepstream\deepstream\sources\libs\nvmsgconv .
Or you can provide the detailed steps for reproducing the problems.

15

i did debug this code and found a bug , that below mention function is returning duplicate payloads
generate_dsmeta_message_minimal(ctx->privData, frame_meta);

i m able to resolve the problem , by adding a custom code logic in this function to avoid duplicates

16

OK. Glad to hear that. Could you attach the patch for your senario for others to reference? Thanks

17

else if (ctx->payloadType == NVDS_PAYLOAD_CUSTOM) {

gchar *message = generate_dsmeta_message_minimal(ctx->privData, frame_meta);

if (message) {
  // Parse the JSON message to extract the 'id' field
  JsonParser *parser = json_parser_new();
  GError *error = NULL;
  if (!json_parser_load_from_data(parser, message, -1, &error)) {
    g_error("Error parsing JSON: %s", error->message);
    g_error_free(error);
  } else {
    JsonObject *root = json_node_get_object(json_parser_get_root(parser));
    const gchar *id = json_object_get_string_member(root, "id");

    // Check if the id is different from the previous id
    if (lastCustomId == NULL || g_strcmp0(id, lastCustomId) != 0) {
      g_free(lastCustomMessage); // Free the previous message if exists

      // Store the new id and message to avoid duplicates
      lastCustomId = g_strdup(id);
      lastCustomMessage = g_strdup(message);

      len = strlen(message);
      // Remove '\0' character at the end of string and just copy the content.
      payload->payload = g_memdup(message, len);
      payload->payloadSize = len;
    }
  }

  g_object_unref(parser);
  g_free(message);
}

attaching the patch of code use to resolve issue

1 Like
18

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.