XINTRA8 - Revision history

Orlando at 17:21, 17 October 2009

2009-10-17T17:21:32Z

← Older revision		Revision as of 10:21, 17 October 2009
Line 11:		Line 11:

	reference [http://keyj.s2000.ws/files/projects/videocomp.pdf] (in German)		reference [http://keyj.s2000.ws/files/projects/videocomp.pdf] (in German)

			The patent describing the actual encoding and decoding method is [http://www.google.com/patents/about?id=lWaBAAAAEBAJ US 7,263,232]

	== Decoding ==		== Decoding ==

Ce: Xintra8 frames are supported now.

2007-11-10T01:00:54Z

Xintra8 frames are supported now.

← Older revision		Revision as of 18:00, 9 November 2007
Line 1:		Line 1:
	An XIntra8 frame (also known as a J-type frame or J-frame) is a type of frame similar to an [[I-frame]]. They are used in Microsoft's MPEG-4 variation codec [[WMV2]]/[[WMV8]] and the same coding method (with a different header) is used in [[WMV3]] Complex Profile under the name "X8". XIntra8 frames are a type of intra frame ([[spatial prediction]]). This is why they are said to be similar to I-frames, because both B-frames and P-frames use ''inter''-frame prediction ([[temporal prediction]]). XIntra8 frames operate on 8x8 pixel blocks. XIntra8 frame types are missing in WMV9. XIntra8 frames reportedly operate in 10 different modes (more information needed).		An XIntra8 frame (also known as a J-type frame or J-frame) is a type of frame similar to an [[I-frame]]. They are used in Microsoft's MPEG-4 variation codec [[WMV2]]/[[WMV8]] and the same coding method (with a different header) is used in [[WMV3]] Complex Profile under the name "X8". XIntra8 frames are a type of intra frame ([[spatial prediction]]). This is why they are said to be similar to I-frames, because both B-frames and P-frames use ''inter''-frame prediction ([[temporal prediction]]). XIntra8 frames operate on 8x8 pixel blocks. XIntra8 frame types are missing in WMV9. XIntra8 frames reportedly operate in 10 different modes (more information needed).

	~~[[MPlayer]] on x86 can play files containing these frames using a binary codec.~~ FFmpeg can ~~also~~ decode files containing Xintra8 frames, ~~but~~ those frames ~~will be skipped. Xintra8 frames have not been reverse engineered, so there is no documentation available~~. If a stream contains Xintra8 frames the Xintra8 bit will be set in the stream's extradata.		FFmpeg can now decode files containing Xintra8 frames, including those frames. If a stream contains Xintra8 frames the Xintra8 bit will be set in the stream's extradata.

	Note: J-frame is the notation used by FFmpeg developers. The [[XMV]] SDK calls it XIntra8 frame and in WMV3 they are called Intra_X8.		Note: J-frame is the notation used by FFmpeg developers. The [[XMV]] SDK calls it XIntra8 frame and in WMV3 they are called Intra_X8.

DonDiego: spelling/wording/grammar/wikify

2007-06-08T15:03:22Z

spelling/wording/grammar/wikify

← Older revision		Revision as of 08:03, 8 June 2007
Line 1:		Line 1:
	An XIntra8 frame (also known as a J-type frame or J-frame) is a type of frame similar to an [[I-frame]]. They are used in Microsoft's MPEG-4 variation codec [[WMV2]]/[[WMV8]] and the same coding method (with a different header) is used in [[WMV3]] Complex Profile under the name "X8". XIntra8 frames are a type of intra frame ([[spatial prediction]]). This is why they are said to be similar to I-frames, because both B-frames and P-frames use ''inter''-frame prediction ([[temporal prediction]]). XIntra8 frames operate on 8x8 pixel blocks. XIntra8 frame types are missing in WMV9. XIntra8 frames reportedly operate in 10 different modes (~~need~~ more ~~info on this~~).		An XIntra8 frame (also known as a J-type frame or J-frame) is a type of frame similar to an [[I-frame]]. They are used in Microsoft's MPEG-4 variation codec [[WMV2]]/[[WMV8]] and the same coding method (with a different header) is used in [[WMV3]] Complex Profile under the name "X8". XIntra8 frames are a type of intra frame ([[spatial prediction]]). This is why they are said to be similar to I-frames, because both B-frames and P-frames use ''inter''-frame prediction ([[temporal prediction]]). XIntra8 frames operate on 8x8 pixel blocks. XIntra8 frame types are missing in WMV9. XIntra8 frames reportedly operate in 10 different modes (more information needed).

	[[MPlayer]] on x86 can play files containing these frames using a binary codec. ~~ffmpeg~~ can also ~~play~~ files containing Xintra8 frames, but, those frames will be skipped. Xintra8 have not been reverse engineered, so there is no documentation available. If a stream contains Xintra8 frames the Xintra8 bit will be set in the stream's extradata.		[[MPlayer]] on x86 can play files containing these frames using a binary codec. FFmpeg can also decode files containing Xintra8 frames, but those frames will be skipped. Xintra8 frames have not been reverse engineered, so there is no documentation available. If a stream contains Xintra8 frames the Xintra8 bit will be set in the stream's extradata.

	Note: J-frame is the notation used by FFmpeg developers. [[XMV]] SDK calls it XIntra8 frame and in WMV3 ~~it's~~ called Intra_X8.		Note: J-frame is the notation used by FFmpeg developers. The [[XMV]] SDK calls it XIntra8 frame and in WMV3 they are called Intra_X8.

	The main differences between XIntra8-frames and I-frames:		The main differences between XIntra8-frames and I-frames:
	# They are not derived from MPEG-4 standard.		# They are not derived from the MPEG-4 standard.
	# ~~Picture~~ is coded as blocks, ''not'' as macroblocks.		# The picture is coded as blocks, ''not'' as macroblocks.
	# Each frame may use its own Huffman codes.		# Each frame may use its own [[Huffman]] codes.

	~~Reference~~ [http://keyj.s2000.ws/files/projects/videocomp.pdf] (in ~~german~~).		reference [http://keyj.s2000.ws/files/projects/videocomp.pdf] (in German)

	== Decoding ==		== Decoding ==
Line 27:		Line 27:

	=== Huffman Decoding ===		=== Huffman Decoding ===
	When each type code is read first time, reader checks if VLC table is set. If not then it reads ~~codes~~ lengths as nibbles and constructs balanced tree from them.		When each type code is read for the first time, the reader checks if the VLC table is set. If not then it reads code lengths as nibbles and constructs a balanced tree from them.

	=== Block Decoding ===		=== Block Decoding ===

	# Read "joint" code		# Read "joint" code.
	# Divide this code into two parts - DC value and block_coded bit		# Divide this code into two parts - DC value and block_coded bit.
	# If block is coded then decode coefficients		# If block is coded then decode coefficients.
	## Read code		## Read code.
	## Split read code into vector (value, skip, last)		## Split read code into vector (value, skip, last).
	## Skip ''skip'' coefficients and set coefficient value to ''value''		## Skip ''skip'' coefficients and set coefficient value to ''value''.
	## If ''last'' is true then end decoding		## If ''last'' is true then end decoding.

	Decoding AC coefficient:		Decoding AC coefficient:
Line 75:		Line 75:

	=== Tables ===		=== Tables ===
	Arrays with the ~~huffman~~ code lengths that ~~wmv2~~ J-Frames use (not sure all are really used) in a sanitized form (decoder stores 2 per byte and with -1 offset). There are codes for values 0..11, 0..76 (these ~~one~~ are used for AC decoding) and 0..33.		Arrays with the Huffman code lengths that WMV2 J-Frames use (not sure all are really used) in a sanitized form (decoder stores 2 per byte and with -1 offset). There are codes for values 0..11, 0..76 (these are used for AC decoding) and 0..33.
	<pre>		<pre>
	~~huffman~~ header:		Huffman header:
	index for code table: depends on number of code tables for mode (1, 2 or 3 bits)		index for code table: depends on number of code tables for mode (1, 2 or 3 bits)

Merbanan: Added the reconstructionlevels table

2007-06-07T20:02:49Z

Added the reconstructionlevels table

@@ Line 336: / Line 336: @@
 ,  2,  4,  5,  6,  7,  8, 14, 14, 14, 14, 14, 14, 14, 14, 13, 13 }
 };
 </pre>
 [[Category:Multimedia Terminology]]

Kostya: Some more known info

2007-06-07T17:09:15Z

Some more known info

@@ Line 12: / Line 12: @@
 Reference [http://keyj.s2000.ws/files/projects/videocomp.pdf] (in german).
-Arrays with the huffman code lengths that wmv2 J-Frames use (not sure all are really used) in a sanitized form (decoder stores 2 per byte and with -1 offset). There are codes for values 0..11, 0..76 and 0..33.
+== Decoding ==
 <pre>
 huffman header:

Rathann: add J-Frame huffman tables (from ffmpeg-devel post)

2007-06-07T15:16:11Z

add J-Frame huffman tables (from ffmpeg-devel post)

Show changes

DonDiego: spelling/grammar

2007-06-01T16:01:00Z

spelling/grammar

← Older revision		Revision as of 09:01, 1 June 2007
Line 1:		Line 1:
	An XIntra8 frame (also known as a j-type frame or j-frame) is a type of frame similar to an [[i-frame~~\|IFrame~~]]. They are used in Microsoft's ~~mpeg4~~ variation ~~CODEC~~ [[WMV2]]/[[WMV8]] and the same coding method (with a different header) is used in [[WMV3]] Complex Profile under the name "X8". XIntra8 frames are a type of intra frame ([[spatial prediction]]). This is why they are said to be similar to i-frames, because both b-frames and p-frames use ''inter''-frame prediction ([[temporal prediction]]). XIntra8 frames operate on 8x8 pixel blocks. XIntra8 frame types are missing in WMV9. XIntra8 frames reportedly operate in 10 different modes (need more info on this).		An XIntra8 frame (also known as a J-type frame or J-frame) is a type of frame similar to an [[I-frame]]. They are used in Microsoft's MPEG-4 variation codec [[WMV2]]/[[WMV8]] and the same coding method (with a different header) is used in [[WMV3]] Complex Profile under the name "X8". XIntra8 frames are a type of intra frame ([[spatial prediction]]). This is why they are said to be similar to I-frames, because both B-frames and P-frames use ''inter''-frame prediction ([[temporal prediction]]). XIntra8 frames operate on 8x8 pixel blocks. XIntra8 frame types are missing in WMV9. XIntra8 frames reportedly operate in 10 different modes (need more info on this).

	[[MPlayer]] on x86 can play files containing these frames using a binary codec. ffmpeg can also play files containing Xintra8 frames, but, those frames will be skipped. Xintra8 have not been reverse engineered, so there is no documentation available. If a stream contains Xintra8 frames the Xintra8 bit will be set in the ~~streams~~ extradata.		[[MPlayer]] on x86 can play files containing these frames using a binary codec. ffmpeg can also play files containing Xintra8 frames, but, those frames will be skipped. Xintra8 have not been reverse engineered, so there is no documentation available. If a stream contains Xintra8 frames the Xintra8 bit will be set in the stream's extradata.

	Note: j-frame is the ~~notiation~~ used by ~~ffmpeg~~ developers. [[XMV]] SDK calls it XIntra8 frame and in WMV3 it's called Intra_X8.		Note: J-frame is the notation used by FFmpeg developers. [[XMV]] SDK calls it XIntra8 frame and in WMV3 it's called Intra_X8.

	The main differences between XIntra8-frames and i-frames:		The main differences between XIntra8-frames and I-frames:
	# They are not derived from MPEG-4 standard		# They are not derived from MPEG-4 standard.
	# Picture is coded as blocks, ''not'' as macroblocks		# Picture is coded as blocks, ''not'' as macroblocks.
	# Each frame may use its own Huffman codes		# Each frame may use its own Huffman codes.

	Reference [http://keyj.s2000.ws/files/projects/videocomp.pdf] (in german).		Reference [http://keyj.s2000.ws/files/projects/videocomp.pdf] (in german).

			[[Category:Multimedia Terminology]]

Crybaby: convert j-frame page to XIntra8 page.

2007-05-31T06:42:22Z

convert j-frame page to XIntra8 page.

New page

An XIntra8 frame (also known as a j-type frame or j-frame) is a type of frame similar to an [[i-frame|IFrame]]. They are used in Microsoft's mpeg4 variation CODEC [[WMV2]]/[[WMV8]] and the same coding method (with a different header) is used in [[WMV3]] Complex Profile under the name "X8". XIntra8 frames are a type of intra frame ([[spatial prediction]]). This is why they are said to be similar to i-frames, because both b-frames and p-frames use ''inter''-frame prediction ([[temporal prediction]]). XIntra8 frames operate on 8x8 pixel blocks. XIntra8 frame types are missing in WMV9. XIntra8 frames reportedly operate in 10 different modes (need more info on this).

[[MPlayer]] on x86 can play files containing these frames using a binary codec. ffmpeg can also play files containing Xintra8 frames, but, those frames will be skipped. Xintra8 have not been reverse engineered, so there is no documentation available. If a stream contains Xintra8 frames the Xintra8 bit will be set in the streams extradata.

Note: j-frame is the notiation used by ffmpeg developers. [[XMV]] SDK calls it XIntra8 frame and in WMV3 it's called Intra_X8.

The main differences between XIntra8-frames and i-frames:
# They are not derived from MPEG-4 standard
# Picture is coded as blocks, ''not'' as macroblocks
# Each frame may use its own Huffman codes

Reference [http://keyj.s2000.ws/files/projects/videocomp.pdf] (in german).