How to use the vignette correction of Intel Media Accelerator

[chae_h_]

Hi, I am Chae.

I am reading the sample_camera code of Intel Media SDK sample.

I want to operate vignette correction of Raw Accelerator.

but when I use vignette correction of sample_camera.cpp, VignetteMaskFile is required for using the correction.

So, I want to know what is Foramt of VignetteMaskFile and receive the sample MaskFile.

thank u.

FILE *maskFile;

  /* Load vignette mask from file */
    MSDK_FOPEN(maskFile, pParams->strVignetteMaskFile, MSDK_STRING("rb"));

[Mark_L_Intel1]

Hi Chae,

I checked the samples and document and I didn't find anywhere it specify this file and I will ask.

In the meantime, you can check the following code to reverse engineer the file format, it should be simple and I think it should use the common Vignette Mask file by the popular photo tools.

https://github.com/Intel-Media-SDK/samples/blob/master/samples/sample_camera/src/pipeline_camera.cpp#L1078

At the line 1078 the function CCameraPipeline::AllocAndInitVignetteCorrection() is the Vignette Mask file parsing logic.

Mark

[chae_h_]

Hi, Liu. 
​Thank u for your answer.

​I have to create mask file myself?
and then, I must create a mask file by using popular photo tools? (ex. photoshop etc..)

 

[Mark_L_Intel1]

Hi Chae,

I think you have to create your own, but I can ask.

Here is the file format I got:

For sample_camera input is bayer format (rg16, bg16, ..), output is .bmp (by default)

Let me know if you have questions.

Mark

[chae_h_]

Hi Liu.

​Vignette correction mask must have 4channel(R, Gr, Gb, B) and 2Bytes(integer, mantissa) each channel..

​So mask file is RAW format?