jpeg2000 help

[cesarcastro]

Hi,

I read the documentation, but I'm still having a hard time understanding how the jpeg2k functions work. Could someone help me? Maybe post a veeery simple example of compression/decompression?

By the way, what does the parameter magnBits mean, in the function ippiEncodeLoadCodeBlock_JPEG2K_32s_C1R?

I know theses are noob questions and for that I apologize...

Thanks,

Cesar

[Vladimir_Dudnik]

Hi,

the best way to understand how JPEG2000 functions does work is to download IPP JPEG2000 codec sample and learn how it work (you can download all IPP samples from here http://www.intel.com/software/products/ipp/samples.htm).

BTW, there is some details from our expert about magnBits parameter:

The "magnBits" told to bit-planes coder how many bit-plans should be considered (and may be coded), starting from the most significant 31-st. The least "31 - magnBits" bits will not be passed and will not be coded. [It should not be confused with "number of zero-bit planes" or nubmer of actually significant bits. The nubmer of actually significant bits (sfBits parameter) is counted during encoding and returned by the function and it's independent of "magnBits".]

For example the "magnBits" may be calculated by the subband quantization step exponent (subbandQuantExp) and dynamic range (dynRange).

magnBits = 31-(dynRange - subbandQuantExp);

In this case the "number of zero bit planes" (to write in codestream) can be calculated like "dynRange + guardBits - sfBits - 1".

Regards,
Vladimir

[pawelpus]

Hi,

I want to ask about other thing. IPP JPEG2000 seems to support 8-bit images only. Am I correct?

[Vladimir_Dudnik]

That's correct for IPP v4.1 release. But we are planning to extend our JPEG2000 sample with support of multi-tile, multi-layer all 5 progression and 16-bit images in the next versions. So, please be tuned for updates.

Regards,

Vladimir