Arducam ov64a40 (64MP) vs imx708 (120 degree 12MP) Autofocus speed

Raspberry Pi Cameras
1

  1. Where did you get the camera module(s)?
    Amazon

  2. Model number of the product(s)?
    ov64a40 (64MP) and imx708 (120 degree 12MP)

  3. What hardware/platform were you working on?
    Raspberry pi 5 with latest OS

  4. Instructions you have followed. (link/manual/etc.)
    Arducam Picamera2 setup with autofocus

  5. Problems you were having?
    The ov64a40 focuses much faster than the imx708 in multiple situation. Sometimes, the imx708 would “give up” on focusing until I adjust the camera a bit. I don’t see the same problem on the ov64a40.

  6. What help do you need?
    I want to know why the ov64a40 is better than the imx708 at autofocusing even though both uses PDAF technology. Is it purely a hardware different where the ov64a40 has better hardware for autofocusing than the imx708 or it there a way I can tune the imx708 to match the autofocusing speed of the ov64a40?

2

Bumping the thread up.

3

Hi there,
Technically speaking, your observation contradicts the fundamental hardware specifications of these two sensors.

The Technical Reality:

  1. OV64A40 (64MP) uses CDAF (Contrast Detection Auto Focus) .
  • This method works by moving the lens back and forth to find the sharpest point (highest contrast). By definition, CDAF must “hunt” and iterate, which makes it inherently slower .
  1. IMX708 (12MP) uses PDAF (Phase Detection Auto Focus) .
  • This method uses special pixels on the sensor to calculate the exact distance of the object instantly. It typically jumps to the correct focus point without hunting. Therefore, the IMX708 should be significantly faster than the OV64A40.

Why is your IMX708 slower?
The fact that your PDAF camera (IMX708) is “giving up” or slower than the CDAF camera (OV64A40) suggests an issue with the specific setup:

  • The Lens Factor: You mentioned using the 120-degree Wide Angle version of the IMX708. Wide-angle lenses have a naturally deep Depth of Field . This often makes it difficult for PDAF systems to detect a sharp “peak,” causing the motor to hesitate.

  • Lighting: PDAF requires good lighting to work effectively.

Request for Video Evidence:
Since your experience is the opposite of expected performance, we suspect a hardware anomaly or a lens-specific issue.
Could you please provide a side-by-side video comparing the two?
We need to see exactly how the IMX708 fails to lock focus. This will help us confirm if it’s a defective motor or a limitation of the wide-angle optics.