计算机工程
計算機工程
계산궤공정
COMPUTER ENGINEERING
2014年
3期
278-282
,共5页
基于上下文的自适应可变长度编码%零内存存取%码字前缀%一级索引%码字后缀%二级索引
基于上下文的自適應可變長度編碼%零內存存取%碼字前綴%一級索引%碼字後綴%二級索引
기우상하문적자괄응가변장도편마%령내존존취%마자전철%일급색인%마자후철%이급색인
Context Adaptive Variable Length Coding(CAVLC)%zero memory access%code prefix%primary index%code suffix%secondary index
在基于上下文的自适应可变长度编码(CAVLC)解码算法中,对非结构化自适应可变长度编码码表进行解码时需要反复查找码表进行码字匹配,从而导致解码速度慢和需要大量内存存取的问题。为此,提出一种可实现零内存存取的CAVLC解码算法。将CAVLC码字前缀0的个数作为一级索引,同时通过一级索引获得输入码流的可能长度。将码字后缀作为二级索引并获得码字的值,直接通过码字快速获得解码结果。对于确定的输入码字,只需通过无码表查找代码操作即可得到对应的解码输出。测试结果表明,该算法不仅可以实现零内存存取的CAVLC解码,而且其解码速度比标准算法提高了45%。
在基于上下文的自適應可變長度編碼(CAVLC)解碼算法中,對非結構化自適應可變長度編碼碼錶進行解碼時需要反複查找碼錶進行碼字匹配,從而導緻解碼速度慢和需要大量內存存取的問題。為此,提齣一種可實現零內存存取的CAVLC解碼算法。將CAVLC碼字前綴0的箇數作為一級索引,同時通過一級索引穫得輸入碼流的可能長度。將碼字後綴作為二級索引併穫得碼字的值,直接通過碼字快速穫得解碼結果。對于確定的輸入碼字,隻需通過無碼錶查找代碼操作即可得到對應的解碼輸齣。測試結果錶明,該算法不僅可以實現零內存存取的CAVLC解碼,而且其解碼速度比標準算法提高瞭45%。
재기우상하문적자괄응가변장도편마(CAVLC)해마산법중,대비결구화자괄응가변장도편마마표진행해마시수요반복사조마표진행마자필배,종이도치해마속도만화수요대량내존존취적문제。위차,제출일충가실현령내존존취적CAVLC해마산법。장CAVLC마자전철0적개수작위일급색인,동시통과일급색인획득수입마류적가능장도。장마자후철작위이급색인병획득마자적치,직접통과마자쾌속획득해마결과。대우학정적수입마자,지수통과무마표사조대마조작즉가득도대응적해마수출。측시결과표명,해산법불부가이실현령내존존취적CAVLC해마,이차기해마속도비표준산법제고료45%。
This paper proposes a zero memory access algorithm for direct Context-based Adaptive Variable Length Coding(CAVLC) decoding, aiming to solve the problems of slow decoding speed and a large of memory accesses caused by the repeated table look-up for the match codeword during the unstructured variable length coding table decoding. This algorithm takes the numbers of zero in CAVLC code prefix as the primary index, and gets the probably length of the input code stream. It takes the code suffix as the secondary index and gets the codeword value. It can get the decoded result by the codeword quickly. For the specific input code, the decoded output can be directly gotten without any table look-up. Test results show that not only the algorithm can achieve zero memory access for CAVLC decoding, but also the decoding speed of this algorithm can achieve 45%speed-up, compared with the standard algorithm.