Sunday, April 20, 2014
English only      
       
 Advanced Search      
Standards Q&A
Send us your questions concerning ITU-T Recommendations

The objective of the Standards Q&A is to enable people from developed and developing countries to submit their queries concerning their understanding and implementation of ITU-T Recommendations and seek advice from Study Group experts. It is an open forum for questions concerning the standardization work of the International Telecommunication Union (ITU). It offers a unique opportunity to engage with the experts that develop the standards that underpin ICTs. This forum is moderated and information/questions posted to the forum may only be visible to everyone only after moderation.


JPEG (ITU-T T.81 | ISO/IEC 10918-1)
Last Post 19-01-2012 08:44 AM by TSB SG16. 3 Replies.
AddThis - Bookmarking and Sharing Button Printer Friendly
Sort:
PrevPrev NextNext
Author Messages Informative
TSB SG16
New Member
New Member
Posts:11

--
19-01-2012 07:55 AM QuoteQuote ReplyReply
What is JPEG?
TSB SG16
New Member
New Member
Posts:11

--
19-01-2012 08:13 AM QuoteQuote ReplyReply
JPEG is an acronym used to represent both a family of image compression standards and the group of experts that developed and maintains it.

JPEG stands for Joint Picture Experts Group, a joint group between ITU-T Study Group 16 and ISO/IEC JTC1 SC29 that was established in 1986. More details can be found on the JPEG home page.

This group produced three sets of standards, among them what is now commonly referred to as JPEG image compression (the official designation being ITU-T T.81 | ISO/IEC 10918-1), which was first issued in 1992.
Shiyu Ji
Posts:112

--
19-01-2012 08:38 AM QuoteQuote ReplyReply

On Figure F.18 at Page 111 of CCITT Rec. T.81 (1992E), when B2=0, which means it comes to 0xFF00, the end of a block (EOB), it need not return one bit. However, why does it come to the left box and finally returns 1 as the result?

I wish I have clarified my point. Thank you!

TSB SG16
New Member
New Member
Posts:11

--
19-01-2012 08:44 AM QuoteQuote ReplyReply
Posted By on 19-01-2012 09:38 AM

On Figure F.18 at Page 111 of CCITT Rec. T.81 (1992E), when B2=0, which means it comes to 0xFF00, the end of a block (EOB), it need not return one bit. However, why does it come to the left box and finally returns 1 as the result?

I wish I have clarified my point. Thank you!

(NOTE: T.81 should be referred to as ITU-T T.81, not CCITT - this is an old designation that was changed circa 1992!)

The left branch with B2=0 is not taken because an EOB symbol is decoded, but rather because it un-does the "byte-stuffing" of the encoder. That is, JPEG requires that on encoding a zero-byte is inserted after each 0xff byte. The procedure in figure F.18 removes this zero-byte again.
 
Decoding of EOB symbols happens at a higher level of the protocol, after the raw bits have been decoded by the Huffman decoder.

The purpose of the byte-stuffing is to allow a syntactical analysis of the compressed bitstream, especially of the entropy coded data stream, without having to analyse the contents of the entropy coded data itself. The byte-stuffing ensures that no byte sequence that could potentially look like a marker appears in the entropy coded data.
 
JPEG 2000 handled this more efficiently by only stuffing zero-bits, not zero bytes into the stream, but the purpose is quite the same.


 
Active Forums 4.1
Top - Feedback - Contact Us -  Copyright © ITU 2011 All Rights Reserved
Contact for this page : TSB EDH

Share
Copyright 2008 by My Website