Scale from Ipp32f to Ipp16u

[rodym]

Hello,

I'm using IPP for Image processing, and I notice thatsome functions are not available.

For example, I cound't find a function to scale from Ipp32f toIpp16u, so I guess that there should be another way to do that using others functions. If this is the case, whathappen with the overhead?

Well, thank you.

[Vladimir_Dudnik]

Hi,

you can use these functions

IPPAPI ( IppStatus, ippiConvert_32f16u_C1R,( const Ipp32f* pSrc, int srcStep, Ipp16u* pDst, int dstStep,IppiSize roiSize, IppRoundMode roundMode ))

IPPAPI ( IppStatus, ippiConvert_32f16u_C3R,( const Ipp32f* pSrc, int srcStep, Ipp16u* pDst, int dstStep,IppiSize roiSize, IppRoundMode roundMode ))

IPPAPI ( IppStatus, ippiConvert_32f16u_AC4R,( const Ipp32f* pSrc, int srcStep, Ipp16u* pDst, int dstStep,IppiSize roiSize, IppRoundMode roundMode ))

IPPAPI ( IppStatus, ippiConvert_32f16u_C4R,( const Ipp32f* pSrc, int srcStep, Ipp16u* pDst, int dstStep,IppiSize roiSize, IppRoundMode roundMode ))

Regards,
Vladimir

Message Edited by vdudnik on 08-23-2005 01:02 PM

[rodym]

Thanks for your answer Vladimir...

It's not only the convertion, because the result must be a 12-bits image. So I need to scale it.

Also, AddWeight function works only with float or 8 bit inputs. What is the reason to discard the 16 bits inputs?

[Vladimir_Dudnik]

Hi,

the reason is simple - it's all about resources and priorities. If you feel this functionality will be useful - please submit your feature request trough Intel Premier Support channel, we track such requests and use them at next version planning stage

Regards,

Vladimir