tag:blogger.com,1999:blog-5246987755651065286.post3585325408489207184..comments2024-02-22T16:15:42.388-08:00Comments on cbloom rants: 08-10-10 - Transmission of Huffman Treescbloomhttp://www.blogger.com/profile/10714564834899413045noreply@blogger.comBlogger8125tag:blogger.com,1999:blog-5246987755651065286.post-13721284401408226682010-09-14T16:34:48.589-07:002010-09-14T16:34:48.589-07:00"I'm talking about decompression only and..."I'm talking about decompression only and assuming a fixed probability table."<br /><br />Basically this is just not right.<br /><br />As I demonstrated in great detail, the fastest arithmetic decoder would be one in which the total probabilities was a power of 2, and each individual probability was a power of 2. That's a Huffman code.<br /><br />If you make the total a power of 2 but let each individual once be a sum of two powers of two, that's slightly slower (ala Rissanen-Mohiuden / DCC95).<br /><br />Anything more general is slower still.cbloomhttps://www.blogger.com/profile/10714564834899413045noreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-8671389874866404812010-09-13T22:31:34.105-07:002010-09-13T22:31:34.105-07:00I'm talking about decompression only and assum...I'm talking about decompression only and assuming a fixed probability table.Anonymoushttps://www.blogger.com/profile/08419606671428685946noreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-69271614713275481972010-08-11T12:06:44.821-07:002010-08-11T12:06:44.821-07:00I'm taking your damn trolling and turning into...I'm taking your damn trolling and turning into constructive and interesting gold. Gold, Jerry, gold!cbloomhttps://www.blogger.com/profile/10714564834899413045noreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-47246741461985781962010-08-11T04:20:30.535-07:002010-08-11T04:20:30.535-07:00Round one! Fight!Round one! Fight!Samhttps://www.blogger.com/profile/01843393036648718286noreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-87213399517565020022010-08-10T19:49:32.442-07:002010-08-10T19:49:32.442-07:00Maybe more.Maybe more.cbloomhttps://www.blogger.com/profile/10714564834899413045noreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-84343601324482108892010-08-10T19:49:19.038-07:002010-08-10T19:49:19.038-07:00"Range coder is faster than Huffman. "
..."Range coder is faster than Huffman. "<br /><br />That is 110% wrong.cbloomhttps://www.blogger.com/profile/10714564834899413045noreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-86749061287126580282010-08-10T19:46:04.640-07:002010-08-10T19:46:04.640-07:00Range coder is faster than Huffman. There's re...Range coder is faster than Huffman. There's really no reason to use Huffman since the majority of range coder related IBM patents have now expired.Anonymoushttps://www.blogger.com/profile/08419606671428685946noreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-70552053398112662062010-08-10T19:44:10.751-07:002010-08-10T19:44:10.751-07:00This comment has been removed by the author.Anonymoushttps://www.blogger.com/profile/08419606671428685946noreply@blogger.com