@nian 不好意思，我们暂时没有计划做raw14bit输出的camera模组。

@乔康

@乔康 said in MV-MIPI-SC130M抓图报错:

驱动正常加载。虽然设备树是4lane的。

问题就在这里，MV-MIPI-SC130M不支持4lane模式配置。 你需要更换设备树为2lane模式。

@乔康
这个问题我们没有遇到。
你可以试试直接用v4l2-ctl抓图，看看效果如何？
https://wiki.veye.cc/index.php/VEYE_CS_Camera_on_Firfly_Boards/zh#.E6.8A.93.E5.8F.96.E5.9B.BE.E7.89.87

@alex_ai First, let me explain the gain configuration logic for the GX-MIPI-IMX662. The IMX662 features 30dB of Analog Gain and supports High Conversion Gain (HCG). The HCG function acts as an amplification during the sensor's analog signal readout, equivalent to 15.3dB.
Therefore, when we configure the "Again" parameter, the actual range is [0, 45.3dB]. This means you can set the Analog Gain up to 45dB without enabling Digital Gain (Dgain).
Currently, our firmware integrates HCG into the "Again" setting, which may have caused some confusion during your testing.
As for the feature to read sensor registers, I have added it to the development roadmap. It will be provided later via a firmware update package.

@ouyangxin 你好，你理解的没错，但是我们现在已经取消了对sda模式的支持，所以这上面的讨论已经没有意义了。

@d_101 Sorry for the late reply. Based on the current situation, I can only suspect a hardware issue that is causing the I2C path to be unavailable. This could be a problem with your mainboard or with the camera itself. Please return the camera to us, and we will perform testing and inspection.

@ffgemanof
The veye_viewer client currently does not support the mono10 and mono12 formats. If you need to use these formats, please configure the data format using mv_mipi_i2c_new.sh, and then acquire the data using v4l2-ctl or the V4L2 API.

Hi @veye_xumm, thank you again for all your help. I found that the VEYE camera you made available with I2C address 0x22 does work on the J401B. And now, after a long series of follow-ups with Seeed Tech Support, I found that i2cdump generates I2C transfer timeouts at address 0x3b (but only 0x3b!).

Then Seeed noted that applying the VEYE "compatibility fix" updating the USB Hub firmware to 08C3 removed the timeouts. So I tried that with the VEYE camera and it works! (This with Jetpack 6.1).

A weird puzzle remains, I was sure the J401B had USB Hub Firmware version 08C3 as supplied, although with one of the "devices" giving different details to 08C3 on the J401. Hence I did not expect to need the "compatibility fix", and had some concerns that reapplying it could cause a problem. But reapplying the update just now the firmware version was initially 08B3!?

It has updated to 08C3, and has fixed camera access to 0x3b.

This has been time consuming, but I might now test JetPack 6.2

@veye_xumm
Sorry for the late reply.
Thank you for the new code!
It seems to work partially in the rpi5 standard OS. I was able to compile the read and write files, and some commands work while others don't.

The new mgain is stuck at zero, so if i use new exp mode i can set the shutter speed with new mshutter, but not the new mgain, despite changing it returns zero when i probe with the -r parameter. so using the new exp mode i only get a black image.

the old mshutter works and i was able go get up to 1 second exposure per frame in streaming mode but the capture mode and capture command don't seem to work.

Also i would like to have it work in the rpi archlinux OS that i use for astronomy. The veyecam2m driver works and i get video in the archlinux OS, but i wasn't able to compile the i2c read and write in that system.

@nimadibaj
You have implemented the function of writing 0xff to 0x000b. Similarly, proceed to write the subsequent registers as well.
The method for writing the subsequent registers is the same as that for writing the 0x0b register, but they must be written in sequence.
Addr--->data
0x10 0xDE
0x11 0xC2
0x12 0x00
0x13 0x00

@veye_xumm

Hello my friend.
We are sending I2C commands via a FPGA.

1_ Writing 0xff to 0x000b solved our problem. Thank you. Please give us a command to save this change.

2_ We are trying to change video frame rate (now is 30fps and we need 25fps).
by these commands we cannot do this :
0x10 0xDE
0x11 0xC2
0x12 0x00
0x13 0x00
(sending Two byte, then sending next two byte and ... ).

Please guide me. Thanks a lot.

@deepcamera

Only ISP-related parameters can be persistently saved after power-off through the paramsave command. Therefore, I recommend that you reconfigure your desired parameters in the startup script during program design.

As for which commands can be saved, you can refer to the veye_mipi_i2c.sh source code. If the corresponding read/write functions contain operations like './i2c_write $I2C_DEV $I2C_ADDR 0x10 0xDA', it indicates that the function's register belongs to the ISP category and can be saved using paramsave.

@siderat I think this one will work.

@veye_xumm sorry for necro-posting, but this method stopped working. on fresh system I have Failed to set I2C address after running make.sh under mv_tools_rpi/sources on rpi5. do you know if something has changed there?

@veye_xumm 好的 一定

@goooofy 好的。

Your current image is Jetpack 5.1.2 with L4T version r35.4.1. The mvcam driver included in it is relatively old, so the mv_probe.sh script cannot work properly. — For more details, you can refer to the comments at the top of the mv_probe.sh script.

081b1526-6747-417e-af94-d33ffbbc08cb-image.png

In addition, please try running the following script and see if it shows continuous messages indicating that data is being received. If it still fails to receive data, please run sudo dmesg to check whether there are any error messages.

export WIDTH=2560 export HEIGHT=1544 export FPS=50 v4l2-ctl --set-ctrl roi_x=0 v4l2-ctl --set-ctrl roi_y=0 v4l2-ctl --set-fmt-video=width=$WIDTH,height=$HEIGHT,pixelformat=GREY v4l2-ctl --set-ctrl frame_rate=$FPS v4l2-ctl --set-fmt-video=width=$WIDTH,height=$HEIGHT,pixelformat=GREY --stream-mmap --stream-count=-1 --stream-to=/dev/null

@veye_xumm 这样啊，好的麻烦您了，我再跟公司里沟通一下

@linoooooo
你好，fps参数是可以修改的。但是需要先停止取图。曝光时间是被限制到1/fps的。 你可以先设置合适的fps，然后在范围内进行曝光时间的修改。

在这个脚本中，带星号的命令，都是取图中写保护的。
你也可以参考寄存器清单中的part1.3来理解这个设置规则。

@linoooooo 你好，我们的相机内部isp流程是没有整帧缓存的，相机本身引入的delay是毫秒之内。
我觉得你这个延时主要是v4l2取图缓存和显示缓存的缘故。不过说实话SDL库我们不太懂，怎样优化需要您自己研究一下。
至于相机采集后，您这边程序直接导入算法的话，延迟肯定要比这种测量方法小的多。