We need to process the signal from the AHD cameras.
In the description (http://www.lindeni.org/lindenis-v5.html) in the section "Various interfaces" it is indicated: "4 x AHD"
The diagram (http://files.lindeni.org/lindenis-v5/dat...c-v1.1.pdf) on the second sheet shows the NCSI1 connection interface (AHD NVP6134c decoder chip). But in the scheme of this chip is not.
As I understand it, we just have the opportunity to connect to this interface through the AHD decoder. If this is the case, then I would ask you to consider the possibility of creating and placing an expansion card for sale (input of the AHD video signal)

Regards, Ihar

In this sample: sample_cve_BDII, it builds a depth picture from left and right YUV format file. 
While if we want to create yuv files from cmos sensor, that is mpi_ise work as input, we have to calibration both left and right camera to get inner and external parameters. Without these paraameters, the Binocular Vision is actually unvaliable.
Is there any document can be refer to do this?

Hi,
In the document 《EVE应用API参考手册.pdf》page 6， the feature 10 said: “EVE提供的算法可以定制分类器，支持任意小变形刚性目标检测(人脸、车辆、车牌、行人、人头、头肩等其他重点目标)“, and "分类器可视化设计，提供一整套完整训练、调试、测试与评估工具".
where can we download these tools? We need design a special target classifer based on V5 platform.

When using a single sensor, what is the difference between these three modes?

WARP_PANO180
WARP_NORMAL
WARP_UNDISTORT

WARP_NORMAL is especially confusing since it requires a 2:1 reduction in resolution and only supports 720P output.

If I try to encode a 3840x2160 image to a 640x360 image I get this message:

error  : cedarc <h264CheckCapability:2235>: error: scaler range is 0.25-8.0, but the request scaler[0.166667,0.170370]

Is there a way to increase the range?

I am setting up the encoder like this:
  VENC_CHN_ATTR_S ve_attr_;
  memset(&ve_attr_, 0, sizeof(VENC_CHN_ATTR_S));
  ve_attr_.VeAttr.Type = PT_H264;
  ve_attr_.VeAttr.MaxKeyInterval = 30;
  ve_attr_.VeAttr.SrcPicWidth = 3840;
  ve_attr_.VeAttr.SrcPicHeight = 2160;
  ve_attr_.VeAttr.Field = VIDEO_FIELD_FRAME;
  ve_attr_.VeAttr.PixelFormat = MM_PIXEL_FORMAT_YVU_SEMIPLANAR_420;
  ve_attr_.VeAttr.Rotate = ROTATE_NONE;
  ve_attr_.VeAttr.AttrH264e.bByFrame = TRUE;
  ve_attr_.VeAttr.AttrH264e.Profile = 2;
  ve_attr_.VeAttr.AttrH264e.MaxPicWidth =3840;
  ve_attr_.VeAttr.AttrH264e.MaxPicHeight = 2160;
  ve_attr_.VeAttr.AttrH264e.PicWidth = 640;
  ve_attr_.VeAttr.AttrH264e.PicHeight = 360;

ve_attr_.RcAttr.mRcMode = VENC_RC_MODE_H264CBR;
ve_attr_.RcAttr.mAttrH264Cbr.mGop = 30;
ve_attr_.RcAttr.mAttrH264Cbr.mSrcFrmRate = 30;
ve_attr_.RcAttr.mAttrH264Cbr.fr32DstFrmRate = 30;
ve_attr_.RcAttr.mAttrH264Cbr.mBitRate = 8 * 1024 * 1024;

I would like to encode 3840x2160 video.  But there seems to be a limit of 3264x2160.  Is that true?

[VideoEnc_Component.c:433]       <setVbvBufferConfig>          Be careful! threshSize[8294400]bytes too large, reduce to 7MB
[VideoEnc_Component.c:478]       <setVbvBufferConfig>          bit rate is 8192Kb, set encode vbv size 12441600, frame length threshold 7340032


I am using the IMX317 sensor.
It is plugged into MIPI0.
I am using ISP_DEV_1 and VI_DEV_0.

Is there a datasheet for the IMX317 available in the wiki?

When I change the encoder framerate using AW_MPI_VENC_SetFrameRate() it tells me I have to 'enable soft frame rate control'.  How do I do that?

I am working with code that optimizes the bandwidth used on a network connection by varying the output framerate from the encoder.  Admittedly, it sets some 'weird' combinations.  My source is always 30fps but the optimization code might ask for 17fps from the encoder, for instance.  Is this allowed?

Here is an example of the warning message:
<VideoEncSetFrameRate>        Low framerate mode, dst:17, src:30, enable soft frame rate control!

• Linbian OS Image support ISP tuning tool
  

• ISP tuning tool
  

• SDK
  

Currently all Mpp_sample show unpleasant results, that when you start to run any sample, you have no way to terminate it and back to desktop because when each sample create vi it transfors the desktop video layer to an outter layer 8 and then close the layer 8. It 's so confused why allwinner do that. 
I do some experiments to try:
1.  insert a line to pthread of display: signal(SIGINT,destroy_vo) .  this SIGINT can capture the ESC key, and terminate the process.
2. resize the VO create size to 640*480

then it works, the 640*480 window stop to run, while the shell shows still active and can input commands.
But,but, the  area of 640*480 window can not be recylcle. I donot find a way to solve this issue.
Does anyone notice this question? How to solve?

提示：部分应用在使用alsa-lib库中接口snd_pcm_hw_params_set_period_size_near和snd_pcm_hw_params_set_buffer_size_near时需要alsa-lib库编译时把alsa-devdir、pcm-plugins、ctl-plugins打开
debian系统默认没有打开，如需实用请重新编译alsa-lib源码，通过configure打开此功能如下：

• 卸载原有alsa
  

• 安装编译alsa-lib
  

• 安装编译alsa-utils