OBS Websocket with Python (Scene Switch: Current scene: {}

Question

I am trying to switch between OBS scenes automatically based on whether a face is detected in a webcam feed. I am using the OBS WebSocket API to connect to OBS and switch between scenes. However, I am encountering an issue where the scene does not switch even though a face is detected. The program runs without any errors, but the debug output shows that the current scene is always an empty dictionary {}. Here is a sample of my code:

import cv2
import time
import numpy as np
from obswebsocket import obsws, requests, events

# OBS WebSocket connection settings
host = "localhost"
port = 4455
password = "PLACEHOLDER"

# Scene names
live_scene = "LIVE"
afk_scene = "AFK"

# Global variable to store the current scene name
current_scene_name = ""

# Connect to OBS
ws = obsws(host, port, password)
print("Connecting to OBS WebSocket at {}:{}".format(host, port))
ws.connect()

try:
    version_info = ws.call(requests.GetVersion())
    print("Connected to OBS WebSocket version: {}.{}.{}".format(
        version_info.data().get("majorVersion", "Unknown"),
        version_info.data().get("minorVersion", "Unknown"),
        version_info.data().get("patchVersion", "Unknown")
    ))
    scenes = ws.call(requests.GetSceneList())
    print("Available scenes:", [scene["name"] for scene in scenes.data()["scenes"]])  # Added debug output
except Exception as e:
    print("Error connecting to OBS WebSocket:")
    print(type(e))
    print(e.args)
    print(e)

# Initialize webcam
cap = cv2.VideoCapture(0)

# Initialize face detection
face_cascade = cv2.CascadeClassifier(cv2.data.haarcascades + 'haarcascade_frontalface_default.xml')

while True:
    ret, frame = cap.read()

    if ret:
        gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)

        faces = face_cascade.detectMultiScale(gray, 1.1, 4)

        if len(faces) > 0:
            print("Face detected")
            current_scene = ws.call(requests.GetCurrentScene())
            print("Current scene:", current_scene.data())  # Added debug output
            current_scene_name = current_scene.data().get("name", "Unknown")
            ws.call(requests.SetCurrentScene(scene_name=live_scene))
            print("Switching to scene:", live_scene)
        else:
            print("No face detected")
            current_scene = ws.call(requests.GetCurrentScene())
            print("Current scene:", current_scene.data())  # Added debug output
            current_scene_name = current_scene.data().get("name", "Unknown")
            ws.call(requests.SetCurrentScene(scene_name=afk_scene))
            print("Switching to scene:", afk_scene)

        time.sleep(1)

    else:
        print("Error reading frame from webcam")
        break

cap.release()
cv2.destroyAllWindows()

I am getting this reply from the OBS WebSocket:

Connected to OBS WebSocket version: Unknown.Unknown.Unknown Current scene: {} Switching to scene: LIVE

It seems that the GetCurrentScene() function is not returning the correct scene. Can someone please help me understand what is causing this issue and how I can fix it?

Answer 1

Here's something I found that might help:

ws.call(requests.SetCurrentScene(**{'scene-name': name}))