Why does setting a slightly smaller resolution(e.g., 1900×1080 instead of 1920×1080) drastically improve USB camera performance in OpenCV on Windows [closed]

I'm using OpenCV-Python on Windows 11 to capture video from a USB camera for object detection tasks. However, I noticed that the camera’s frame rate is extremely low — maybe only about 1 FPS, even though the camera officially supports MJPG 1920×1080 @ 60FPS. On Ubuntu 20.04, the same camera works perfectly with the expected high frame rate.

While searching online, I found an odd workaround: Instead of setting the resolution directly to 1920×1080, if I set it to something like 1900×1080, the program automatically picks the nearest valid resolution (still outputs 1920×1080 frames) — but suddenly, the delay drops dramatically and the frame rate becomes normal.
I don’t understand why this happens on Windows but not on Linux.
Could someone explain the underlying reason for this behavior?

import cv2

cap = cv2.VideoCapture(0, cv2.CAP_DSHOW)
cap.set(cv2.CAP_PROP_FOURCC, cv2.VideoWriter_fourcc(*"MJPG"))
cap.set(cv2.CAP_PROP_FRAME_WIDTH, 1900)  # if set 1920, it's very slow 
cap.set(cv2.CAP_PROP_FRAME_HEIGHT, 1080)
cap.set(cv2.CAP_PROP_FPS, 60)

while True:
    ret, frame = cap.read()
    if not ret:
        break
    cv2.imshow("camera", frame)
    if cv2.waitKey(1) == 27:
        break

cap.release()
cv2.destroyAllWindows()