How to do the multiply of Ipp16s and float?

pango · ‎07-13-2005

my program multiply the RGB vector of a image by a float value,the data type of image is Ipp16s,I write below code to do the multiply:

float ratio=0.5f;//the float value

Ipp16s mulValue[]={IppMax16S*ratio,IppMax16S*ratio,IppMax16S*ratio};

ippiMulC_16s_AC4RSfs(pTempSrc,lineBytesEx,mulValue,pTempDst,lineBytesEx,roiSize,16);//wrong

the result of above code is wrong,and I think the key is that the input of "mulValue" and "scaleFactor" param is wrong,but I can't understand the explaination in SDK,so where is the problem of my code?How to rewrite it?

Vladimir_Dudnik · ‎07-27-2005

Hi,

seems you need to do shift by 15 not by 16:

ippiMulC_16s_AC4RSfs(pTempSrc,lineBytesEx,mulValue,pTempDst,lineBytesEx,roiSize,15);

Regards,

Vladimir