12-bit JPEG support

12-bit JPEG support

Can someone explain to me which functions in IPP 5.2 deal with 12-bit JPEG compression?

Regards,
Igor Levicki
6 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.

Hello,

I recommend you to take a look on IPP JPEGView sample which demonstrates implementation of JPEG codec with support for lossy 12-bit compression mode.

Regards,
Vladimir

I recommend you to take a look on IPP JPEGView sample which demonstrates implementation of JPEG codec with support for lossy 12-bit compression mode.

Vladimir, if I wanted to troll over 592.8KB of source code to find the info I asked for, I wouldn't bother to ask here.

This page says:

JPEG processing: New functions for JPEG Extended Baseline mode (12-bit per sample)...

I underlined the part which I consider important and misleading. It implies that there are some new functions implemented as part of IPP, not part of the JPEGView sample.

I have read IPP PDF documentation and I couldn't find a single reference to 12-bit JPEG functionality whatsoever. That is why I am asking about it here.

Regards,
Igor Levicki

Hi Igor,

sorry for uncertancy,

there are two kind of functions which can be used to compress 12-bit images:

1. in lossless mode you can compress 2..16-bit images, and set of functions to that is
ippiEncodeHuffmanOne_JPEG_16s1u_C1
ippiDecodeHuffmanOne_JPEG_1u16s_C1

new functions, which improve overall performance of lossless compression mode
ippiEncodeHuffmanRow_JPEG_16s1u_P4C1
ippiDecodeHuffmanRow_JPEG_1u16s_C1P4

ippiDiffPredFirstRow_JPEG_16s_C1
ippiDiffPredRow_JPEG_16s_C1
ippiReconstructPredFirstRow_JPEG_16s_C1
ippiReconstructPredRow_JPEG_16s_C1

2. extended baseline mode for lossy compression of 12-bit images
ippiDCTQuantInv8x8LS_JPEG_16s16u_C1R
ippiDCTQuantFwd8x8LS_JPEG_16u16s_C1R

To not spend time looking through 500+ Kb you can look in only jpegenc.cpp/jpegdec.cpp files, where the actual JPEG codec engine is implemented. Let us know if you have additional questions regarding this.

Regards,
Vladimir

Now that was the kind of answer I was looking for, thanks!

I will look into it.

Regards,
Igor Levicki

Thanks for this thread. I cannot find a way to add this to my favorites.

Leave a Comment

Please sign in to add a comment. Not a member? Join today