struggles.
yet another average day. started working on the homework for dsp which is due friday. this time there were a lot of problems and every problem has a lot of calculations. think i should probably use matlab for some of them since the fourier coefficient calculations take forever by hand.
since dr. pfister is in san diego this week for ITA, dr. narayanan took over the lecture for digital communications. last semester i sat in on dr. narayanan's information theory class. though it was a bit of a struggle with all the advanced material, nevertheless, it was such a joy listening to dr. narayanan's lectures. he's always very patient in presenting the material. he almost never rushes in writing things down on the board and carefully explaining each results. since most of his research is in the area of channel coding, he talked about a few algorithms today. the majority of the discussion was focused on huffman code, which is a very simple coding scheme that produces an optimal instantaneous code by utilizing the probability distribution of the source. it is used widely in data compression, such as jpeg and mpeg files. there is a really nice story behind it. definitely worth reading and knowing about. =)
anyway, back to homeworking.
rating 01/30/2008: 7
Subscribe to:
Post Comments (Atom)
2 comments:
When I read "huffman code" I knew that sounded familiar and confirmed my idea that it was related to JPEG after reading the rest of the sentence. Yeah that sounds like some cool stuff. Your blog makes me want to build a robot right now.
hey i didn't see that you made a comment! i thought they were supposed to let me know with an email or something... anyway, yeah huffman code is pretty cool. too bad i didn't get tested on it... if i did, i'd aced that! but then we got tested on this modified limpal-ziv code, which i don't think i did very hot on... =\
Post a Comment