questions of VLIB_packMask32 and VLIB_unpackMask32

bskyfish bskyfish

My video frame size is 320*320, it is grayscale.

code:

------------------------------------------------------

VLIB_packMask32(frame,frame32,320*320);

VLIB_unpackMask32(frame32,frame,320*320); //VLIB_unpackMask32(frame32,frame,0); I make some change. But I get the 8-bit Y data display.

display(frame);

------------------------------------------------------

refer to the examples of VLIB, but I cannnot get the 8-bit Y data display.

where the question?

over 14 years ago

0 David Friedland over 14 years ago

TI__Mastermind 18320 points

Can you fill in what device you are using, and what version of the various software tools and components you are using as well?

0 bskyfish bskyfish over 14 years ago in reply to David Friedland

Prodigy 60 points

sorry!

I using DM6446 and vlib2.1,

Dubeg under Linux.

0 Senthil Kumar Yogamani over 14 years ago in reply to bskyfish bskyfish

TI__Intellectual 1450 points

Hello,

Your syntax seems right.

I am assuming you are using the right data-types too - frame should be unsigned char and frame32 should be unsigned int.

I am also assuming you have linearly arranged the input image as a 1D array.

VLIB_unpackMask32(frame32,frame,0); works because the loop counter is zero and nothing is written in frame and frame already has the data.

There are two assumptions used in these VLIB functions: (from documentation)

• The buffer maskImage need to be double-word aligned in memory.
• The pixelCount must be a multiple of 8.

The second criteria is satisfied as 320*320 is a multiple of 8.

Sometimes alignment may be a problem. You can use the pragma #DATA_ALIGN to explicitly make the array double-word aligned.

If you can share the input image and the output image, I might be able to to help you better.

Regards,

Senthil

0 bskyfish bskyfish over 14 years ago in reply to Senthil Kumar Yogamani

Prodigy 60 points

Senthil, Thanks!

-----------------------------

VLIB_unpackMask32(frame32,frame,0); works because the loop counter is zero and nothing is written in frame and frame already has the data.

-----------------------------

Yes, I think the reason is right. I rewrite the pragram code:

--------------------------------------------------------------------------------

void bpack(void)
{
char sprintBuffer[200];
int i=0,j=0,k=0;

int *inputImage;

Uint8 *y;

int sHEI=320,sWID=320;
y=(Uint8*)malloc(sHEI*sWID);

inputImage=(int *)malloc(sizeof(int)*sHEI*sWID/32); // I try define the inputImage as unsigned int, the result is same.

memset(y,0,sHEI*sWID);

//make a Y image
for(i=4;i<8;i++)
{
      for(j=15;j<18;j++)
   {
          *(y+i*sWID+j)=1;
     }
}
for(i=10;i<15;i++)
{
          for(j=5;j<18;j++)
             {
            *(y+i*sWID+j)=1;
           }
}

memset(inputImage,0,sizeof(int)*sHEI*sWID/32);
VLIB_packMask32(y,inputImage,sHEI*sWID);

VLIB_unpackMask32(inputImage,y,0);

free(inputImage);
free(y);

}

--------------------------------------------------------------------------------

0 Senthil Kumar Yogamani over 14 years ago in reply to bskyfish bskyfish

TI__Intellectual 1450 points

Hello,

This seems to be wrong. Use the appropriate array size instead of 0.

VLIB_unpackMask32(inputImage,y,0);

Please try to debug step by step. It is difficult to debug several functions at a time.

Run VLIB_packMask32 and see if the output is correct by inspect. Only after fixing this, you should move to the unpack function.

Regards,

Senthil

Processors

Processors forum

questions of VLIB_packMask32 and VLIB_unpackMask32