LibArgus Exposure Time Off by 1 µs on Jetson Orin NX/Nano (JP6.2)

[kishore.kumar1]

Dear NVIDIA team,
We are working on Jetson Orin NX/Nano with JP6.2.

We are observing that the exposure time configured through the Argus camera application or the GStreamer nvarguscamerasrc plugin is always one microsecond less than the requested value.

For example, when we configure the exposure time to 10000 µs using the Argus camera application, the sensor driver receives 9999 µs instead. This can also be verified by reading the value using:

v4l2-ctl -C exposure

We have further verified this behavior in the nvidia-oot

driver and confirmed that the value received from LibArgus itself is already one microsecond lower than the configured value.

We have attached an image for your reference. Could you please let us know whether this behavior is expected? If so, could you also explain the reason behind this one-microsecond difference?

Thanks.

[carolyuu]

— 🤖 This is an automated AI response. I’m here to help, but please verify important details! —

I could not find clear support for that in the provided Jetson Orin Nano documentation. The provided context does not mention the specific issue of exposure time being off by 1 µs on Jetson Orin NX/Nano (JP6.2) when using LibArgus or the GStreamer nvarguscamerasrc plugin. The context discusses various topics related to Jetson Orin Nano, such as camera settings, GStreamer pipelines, and V4L2 API, but does not address this particular issue. Therefore, I cannot provide a definitive answer to the question.

— We support team will also look into your issue to have further suggestions soon. please stay tuned. —

[ShaneCCC]

Can v4l2-ctl set to 10000000 by command v4l2-ctl -c exposure 10000000?

How about change the gain range to check again?

[kishore.kumar1]

When the exposure is set using the v4l2-ctl command, it is configured as expected. The issue is observed only when using LibArgus. However, with v4l2-ctl, the exposure time is specified in microseconds, whereas in LibArgus, it is specified in nanoseconds.

I have attached image for your reference.

[ShaneCCC]

Set the AE antibanding mode to off should be able set to correct value.

Thanks

[kishore.kumar1]

It is still not resolving when setting AE antibanding to OFF.
This issue is not only with 10ms also occurs in lower exposure time. Sharing another example for your reference.

Thanks

[ShaneCCC]

AE should unlocked.

[kishore.kumar1]

AE unlocked. Still issue persists.

Thanks

[ShaneCCC]

I don’t see the problem on IMX219 reference sensor.

Maybe check the device tree and driver config.

[kishore.kumar1]

I am sharing the device tree configuration of that specific mode. Can you please point me out if any misconfiguration?

mode0 { // 2048X1536 12bit 2376Mbps
mclk_khz = "74250";
num_lanes = CAM_LANES_STRING;
tegra_sinterface = TEGRA_S_INTERFACE;
phy_mode = "DPHY";
discontinuous_clk = "no";
dpcm_enable = "false";
cil_settletime = "0";
dynamic_pixel_bit_depth = "12";
csi_pixel_bit_depth = "12";
mode_type = "bayer";
pixel_phase = "rggb";
lane_polarity = LANE_POLARITY;

active_w = "2048";
active_h = "1536";
readout_orientation = "0";
line_length = "2064";
inherent_gain = "1";
mclk_multiplier = "5.33";
    pix_clk_hz = "396000000";

framerate_factor = "1000000";
min_framerate = "1000000";
max_framerate = "72000000";
step_framerate = "1";
default_framerate = "72000000";

gain_factor = "10";
min_gain_val = "0";
max_gain_val = "240";
step_gain_val = "1";
default_gain = "0";

min_hdr_ratio = "1";
max_hdr_ratio = "1";
exposure_factor = "1000000";
min_exp_time = "10";
max_exp_time = "1000000";
step_exp_time = "1";
default_exp_time = "13889";
embedded_metadata_height = "1";

};

Thanks

[kishore.kumar1]

Adding to the above comment, I also checked the value received in the tegracam_set_ctrls() API within the tegracam_ctrls.c file.

There, the exposure time is shown as 9999 µs instead of the expected 10000 µs.

Thanks.

[ShaneCCC]

What’s the version?

I am verify on 36.5 without problem.

 R36 (release), REVISION: 5.0, GCID: 43688277, BOARD: generic, EABI: aarch64, DATE: Fri Jan 16 03:50:45 UTC 2026
# KERNEL_VARIANT: oot
TARGET_USERSPACE_LIB_DIR=nvidia
TARGET_USERSPACE_LIB_DIR_PATH=usr/lib/aarch64-linux-gnu/nvidia

[kishore.kumar1]

I was testing it in the L4T36.4.3 JP6.2.
Can you please confirm the behavior in this release version?

Thanks.

[ShaneCCC]

Replace the /usr/lib/aarch64-linux-gnu/tegra/libnvodm_imager.so by attached file to verify.

libnvodm_imager.so.36.4.3_accuracy (5.4 MB)

[kishore.kumar1]

After replacing the shared library file. It works as expected.

Thanks for the support.