Intel® Integrated Performance Primitives
Deliberate problems developing high-performance vision, signal, security, and storage applications.
6704 Discussions

ippiYCbCr411ToYCbCr420_8u_P3R vs cc_YUV411_to_YUV420

anonymous210
Novice
242 Views
Hi,


The file umc_color_space_conversion.cpp contains the following code:

----------------
switch (srcFormat) {
case YUV411:
switch (dstFormat) {
case YUV420:
//status = ippiYCbCr411ToYCbCr420_8u_P3R(pSrc, pSrcStep, pDst, pDstStep, srcSize);
status = cc_YUV411_to_YUV420(pSrc, pSrcStep, pDst, pDstStep, srcSize);
break;

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

Why the ippiYCbCr411ToYCbCr420_8u_P3R is commented out?
Is there any difference between ippiYCbCr411ToYCbCr420_8u_P3R and cc_YUV411_to_YUV420?
I don't see any difference between these functions and just wondering why UMC is not using the optimized ippiYCbCr411ToYCbCr420_8u_P3R?


Thanks
0 Kudos
2 Replies
Chao_Y_Intel
Moderator
242 Views

Hello,

I just checked the code , ippiYCbCr411ToYCbCr420_8u_P3R should be the right optimized function for such color conversion. I submitted a request into engineer database, and the owner will clear up the code to avoid any confusion.

Regards,
Chao

0 Kudos
Vladimir_Dudnik
Employee
242 Views
Hello,

IPP function was designedto not use filtering during transform which cause to visually worse results. It is acceptable in some cases and was done intenionally for performance reasons.Socounterpart functionality was developed in UMC level.

Regards,
Vladimir
0 Kudos
Reply