[PATCH] libcamera: camera_sensor_properties: ov5675: Set correct delays
Naushir Patuck
naush at raspberrypi.com
Thu Dec 19 10:33:56 CET 2024
Hi Kieran,
On Thu, 19 Dec 2024 at 01:00, Kieran Bingham
<kieran.bingham at ideasonboard.com> wrote:
>
> The OV5675 uses different delays for gain and exposure than are configured
> in the default sensorDelays utilised by the CameraSensorLegacy.
>
> Empirical testing using a Lenovo X13s shows that the exposure delay is
> only a single frame, and the current setting of 2 frame delay produces
> exceedingly frequent oscillations in the image exposure.
>
> Update the OV5675 sensor delays table accordingly.
>
> Signed-off-by: Kieran Bingham <kieran.bingham at ideasonboard.com>
> ---
> src/libcamera/sensor/camera_sensor_properties.cpp | 7 ++++++-
> 1 file changed, 6 insertions(+), 1 deletion(-)
>
> diff --git a/src/libcamera/sensor/camera_sensor_properties.cpp b/src/libcamera/sensor/camera_sensor_properties.cpp
> index bd1fc86977ce..813878386a35 100644
> --- a/src/libcamera/sensor/camera_sensor_properties.cpp
> +++ b/src/libcamera/sensor/camera_sensor_properties.cpp
> @@ -360,7 +360,12 @@ const CameraSensorProperties *CameraSensorProperties::get(const std::string &sen
> { controls::draft::TestPatternModeOff, 0 },
> { controls::draft::TestPatternModeColorBars, 1 },
> },
> - .sensorDelays = { },
> + .sensorDelays = {
> + .exposureDelay = 1,
>From my experience, an exposure delay of 1 is very unlikely with a
rolling shutter sensor because of the integration overlap between
successive frames. Of course, I could be wrong since I've not
encounter this sensor before. Perhaps it's worth looking at other
delays to see if the general interaction between vblank/exposure/gain
is causing the oscillations?
Regards,
Naush
> + .gainDelay = 1,
> + .vblankDelay = 2,
> + .hblankDelay = 2
> + },
> } },
> { "ov5693", {
> .unitCellSize = { 1400, 1400 },
> --
> 2.47.1
>
More information about the libcamera-devel
mailing list