Dear I. Pavlov,
I am doing analysis of LZMA for decompression purpose.
For that, refering "LzmaSpec.cpp" file for basic understanding.
Having some doubts regarding to Range Coding and expecting help for the same:
In document "LzmaSpec.txt", it is mentioned that, "CProb variable contains information about estimated probability for symbol 0".
But I found that while calculating probability for symbol "0", the updation of probability is done over "2048 - CProb" value, which logically refers to probability of "1" and also while calculating probability for symbol "1", the updation of probability is done over "CProb" value, which logically refers to probability of "0".
According to my understanding, I seems to be opposite.
Please give me the correct direction.
As we are reading 4B of data at only initially and after that, reading data as per requirement for decoding single symbol in Range Decoding, how we are updating probabilities w.r.t. 32?
Waiting for the response.
Thanking you.
Regards,
Sayali
1) use some example numbers: exact probability value before symbol 0 and after symbol 0.
You can start from 0.25 probability.
2) I don't understand that question.