mediaframe.mpeg1
Class MPEG_video

java.lang.Object
  |
  +--mediaframe.mpeg1.MPEG_video

All Implemented Interfaces:

java.lang.Runnable

public class MPEG_video

extends java.lang.Object

implements java.lang.Runnable

The MPEG_video class implements the video decoder of the MPEG bit stream.

Author:

Joerg Anders

Field Summary

private int	ak_idx The index of the frame coming into being.
private int	Asp_ratio The aspect ratio used by the MPEG standard given in the sequence headerr.
static int	B_TYPE B FRAME frame type constanta.
private int	back_idx The index of the frame for backward prediction.
private motion_data	Backward The object of class "motion_data" to notice and compute the backward motion values.
private int	backward_f The the size of the motion data for backward prediction.
private int	backward_f_code The motion code value for backward prediction.
private int	backward_r_size The size of residual value for backward prediction.
private int	Bit_rate The bit rate specified in the sequence header.
private boolean	Broken_Link The Broken Link value specified in the group of pictures header.
private boolean	Closed_Group The Closed Group value specified in the group of pictures header.
private int	col_y_incr The increment for the luminance line, equals to pixel_per_col_line - 8.
private boolean	const_param True if the constrained parameter flag was set in the sequence header and false otherwise.
private int	dct_dc_cb_past The value of past DC (cb) value.
private int	dct_dc_cr_past The value of past DC (cr) value.
private int	dct_dc_y_past The value of past DC (y) value.
private int[]	dct_recon The values before IDCT transformation.
private boolean	Drop_Flag The Drop Flag value specified in the group of pictures header.
private int	error_count The current amount of printed video errors.
private static int	EXT_START_CODE MPEG VIDEO layer constanta.
private int	forw_f_code The motion code value for forward prediction.
private motion_data	Forward The object of class "motion_data" to notice and compute the forward motion values.
private int	forward_f The the size of the motion data for forward prediction.
private int	forward_r_size The size of residual value for forward prediction.
private int	Frame_nr The frame number in the movie.
private int	Frame_nr_offset The first frame number of last group of pictures.
private boolean	Full_pel_back_vector Equals true if full pixel accuracy is used for backward prediction, equals false if half pixel accuracy is used.
private boolean	Full_pel_forw_vector Equals true if full pixel accuracy is used for forward prediction, equals false if half pixel accuracy is used.
private static int	GOP_START_CODE MPEG video layer constanta.
private int	Height The height in pixels of the movie data.
private int	Hour The Hour time value specified in the group of pictures header.
private Huffmann	Huf The VLC (Hufmann) decoder.
static int	I_TYPE I FRAME frame type constanta.
private IDCT	idct An IDCT object to tranform the DCT coefficients.
private int[]	intramatrix The default intramatrix.
private boolean	lum_block False if in the current macroblock hasn't had any luminance blocks.
private int	lum_y_incr The increment for the luminance line, equals to pixel_per_lum_line - 8.
private int	macro_block_address The actual MB address.
private boolean	macro_block_intra True if a macro block is for intra coded frames.
private boolean	macro_block_motion_backward True if motion vector for backward prediction exists.
private boolean	macro_block_motion_forward True if motion vector for forward prediction exists.
private boolean	macro_block_pattern True if a coded block pattern supplied.
private boolean	macro_block_quant True if a new quantization factor is supplied.
private int	mb_column The actual column position of the macro block.
private int	mb_height The macroblock height.
private int	mb_row The actual row position of the macro block.
private int	mb_width The macroblock width.
private int	Minute The Minute time value specified in the group of pictures header.
private java.lang.Thread	moribund The "Video Decoder" thread that is being stopped.
private int	motion_horiz_back_code The motion code value of horizontal part for backward prediction.
private int	motion_horiz_back_r The motion residual value of horizontal part for backward prediction.
private int	motion_horiz_forw_code The motion code value of horizontal part for forward prediction.
private int	motion_horiz_forw_r The motion residual value of horizontal part for forward prediction.
private int	motion_verti_back_code The motion code value of vertical part for backward prediction.
private int	motion_verti_back_r The motion residual value of vertical part for backward prediction.
private int	motion_verti_forw_code The motion code value of vertical part for forward prediction.
private int	motion_verti_forw_r The motion residual value of vertical part for forward prediction.
private io_tool	mpeg_stream The input movie's data bit stream.
private int[]	non_intramatrix A quantization matrix for non intra frames.
private int[]	nullmatrix A matrix of zeros.
static int	P_TYPE P FRAME frame type constanta.
private int	past_intra_address The past MB address.
private int[][][]	Pel_buffer The "Pel_buffer" is a main feature of the video scanner. 3 frames are stored at: Pel_buffer[0] Pel_buffer[1] Pel_buffer[2] The frame at index "ak_idx" is the frame coming into being.
private int[]	pel1 The array to grab the referred area from predicted frame.
private int[]	pel2 The array to grab the referred area from predicted frame.
private int	Pic_rate The frame rate code used by the MPEG standard given in the sequence header.
private int	Pic_Type The type of the frame (I, B, or P).
private int	Pict_Count The Pict Count value specified in the group of pictures header.
private static int	PICTURE_START_CODE MPEG video layer constanta.
private int	pixel_per_col_line The number of pixels per colour line.
private int	pixel_per_lum_line The number of pixels per luminance line.
private MPEG1	Player The reference to the applet.
private int	pred_idx The index of the frame for forward prediction.
private boolean	quant_matrix True if the quantizer matrix (intra or nonintra) should be loaded from MPEG stream.
private int	Quant_scale The quantization factor.
private int	Second The Second time value specified in the group of pictures header.
private static int	SEQ_END_CODE MPEG video layer constanta.
private static int	SEQ_START_CODE MPEG video layer constanta.
private static int	SLICE_MAX_START_CODE MPEG video layer constanta.
private static int	SLICE_MIN_START_CODE MPEG video layer constanta.
private int	Temp_ref The display order of the picture in the group of pictures.
private static int	USER_START_CODE MPEG VIDEO layer constanta.
private int	VBV_buffer The vbv buffer size specified in the sequence header.
private int	VBV_Delay The VBV delay value of the frame.
private java.lang.Thread	video_thread The "Video Decoder" thread object.
private int	Width The width in pixels of the movie data.
private int[]	zigzag The reverse zigzag scan order.

Constructor Summary

(package private)

MPEG_video(MPEG1 play, io_tool tool) Constructs a MPEG_video object, notices the parameters, inits "non_intra" and "null" matrixes and creates an VLC (Hufmann) decoder object.

Method Summary

private void	correct_col_pixel(int nr) The method "correct_col_pixel" is called in predicted macro blocks.
private void	correct_lum_pixel(int nr) The method "correct_lum_pixel" is called in predicted macro blocks.
void	join() Waits until the "Video Decoder" thread is dead.
private void	Parse_Block(int nr) The method "Parse_Block" parses a block according to ISO 11172-2.
private void	Parse_group_of_pictures() The method "Parse_group_of_pictures" parses the group of pictures according to ISO 11172-2.
private void	Parse_macroblock(int b_nr) The method "Parse_macroblock" parses a macroblock according to ISO 11172-2.
private void	Parse_picture() The method "Parse_picture" parses a picture according to ISO 11172-2.
private void	Parse_sequence_header() The method "Parse_sequence_header" parses the sequence header according to ISO 11172-2.
private void	Parse_slice() The method "Parse_slice" parses a slice according to ISO 11172-2.
void	run() The method "run" parses the MPEG video stream according to ISO 11172-2 and performs some initial steps.
private void	set_col_pixel(int nr) The method "set_col_pixel" takes the re-transformed chrominance values and places them at the appropriate position.
private void	set_dct_diff(int dct_diff, int dct_size) The method "set_dct_diff" computes the DCT difference according to ISO 11172-2.
private void	set_lum_pixel(int nr) The method "set_lum_pixel" takes the re-transformed luminance values and places them at the appropriate position.
void	start() Starts the "Video Decoder" thread.
void	stop() Stops the "Video Decoder" thread.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

mpeg_stream

private io_tool mpeg_stream

The input movie's data bit stream.

Huf

private Huffmann Huf

The VLC (Hufmann) decoder.

nullmatrix

private int[] nullmatrix

A matrix of zeros.

intramatrix

private int[] intramatrix

The default intramatrix.

zigzag

private final int[] zigzag

The reverse zigzag scan order.

idct

private IDCT idct

An IDCT object to tranform the DCT coefficients.

dct_recon

private int[] dct_recon

The values before IDCT transformation.

non_intramatrix

private int[] non_intramatrix

A quantization matrix for non intra frames.

lum_block

private boolean lum_block

False if in the current macroblock hasn't had any luminance blocks.

dct_dc_cr_past

private int dct_dc_cr_past

The value of past DC (cr) value.

dct_dc_cb_past

private int dct_dc_cb_past

The value of past DC (cb) value.

dct_dc_y_past

private int dct_dc_y_past

The value of past DC (y) value.

Pel_buffer

private int[][][] Pel_buffer

The "Pel_buffer" is a main feature of the video scanner. 3 frames are stored at:

				Pel_buffer[0]
				Pel_buffer[1]
				Pel_buffer[2]

The frame at index "ak_idx" is the frame coming into being. The frame at index "pred_idx" is the frame for forward prediction. The frame at index "back_idx" is the frame for backward prediction. The method "Parse_Picture" administers the values of these 3 variables. The index in second dimension determins whether the information is:

		luminance   information 	-	Pel_buffer[?][0]
		chrominance information (cr)	-	Pel_buffer[?][1]
 		chrominance information	(cb)	-	Pel_buffer[?][2]
 

ak_idx

private int ak_idx

The index of the frame coming into being.

pred_idx

private int pred_idx

The index of the frame for forward prediction.

back_idx

private int back_idx

The index of the frame for backward prediction.

Player

private MPEG1 Player

The reference to the applet.

SEQ_END_CODE

private static final int SEQ_END_CODE

MPEG video layer constanta.

See Also:

Constant Field Values

SEQ_START_CODE

private static final int SEQ_START_CODE

MPEG video layer constanta.

See Also:

Constant Field Values

GOP_START_CODE

private static final int GOP_START_CODE

MPEG video layer constanta.

See Also:

Constant Field Values

PICTURE_START_CODE

private static final int PICTURE_START_CODE

MPEG video layer constanta.

See Also:

Constant Field Values

SLICE_MIN_START_CODE

private static final int SLICE_MIN_START_CODE

MPEG video layer constanta.

See Also:

Constant Field Values

SLICE_MAX_START_CODE

private static final int SLICE_MAX_START_CODE

MPEG video layer constanta.

See Also:

Constant Field Values

EXT_START_CODE

private static final int EXT_START_CODE

MPEG VIDEO layer constanta.

See Also:

Constant Field Values

USER_START_CODE

private static final int USER_START_CODE

MPEG VIDEO layer constanta.

See Also:

Constant Field Values

I_TYPE

public static final int I_TYPE

I FRAME frame type constanta.

See Also:

Constant Field Values

P_TYPE

public static final int P_TYPE

P FRAME frame type constanta.

See Also:

Constant Field Values

B_TYPE

public static final int B_TYPE

B FRAME frame type constanta.

See Also:

Constant Field Values

Width

private int Width

The width in pixels of the movie data.

Height

private int Height

The height in pixels of the movie data.

Asp_ratio

private int Asp_ratio

The aspect ratio used by the MPEG standard given in the sequence headerr.

Pic_rate

private int Pic_rate

The frame rate code used by the MPEG standard given in the sequence header.

mb_width

private int mb_width

The macroblock width.

mb_height

private int mb_height

The macroblock height.

Bit_rate

private int Bit_rate

The bit rate specified in the sequence header.

VBV_buffer

private int VBV_buffer

The vbv buffer size specified in the sequence header.

const_param

private boolean const_param

True if the constrained parameter flag was set in the sequence header and false otherwise.

quant_matrix

private boolean quant_matrix

True if the quantizer matrix (intra or nonintra) should be loaded from MPEG stream.

Hour

private int Hour

The Hour time value specified in the group of pictures header.

Minute

private int Minute

The Minute time value specified in the group of pictures header.

Second

private int Second

The Second time value specified in the group of pictures header.

Pict_Count

private int Pict_Count

The Pict Count value specified in the group of pictures header.

Drop_Flag

private boolean Drop_Flag

The Drop Flag value specified in the group of pictures header.

Closed_Group

private boolean Closed_Group

The Closed Group value specified in the group of pictures header.

Broken_Link

private boolean Broken_Link

The Broken Link value specified in the group of pictures header.

Temp_ref

private int Temp_ref

The display order of the picture in the group of pictures.

Pic_Type

private int Pic_Type

The type of the frame (I, B, or P).

Frame_nr_offset

private int Frame_nr_offset

The first frame number of last group of pictures.

Frame_nr

private int Frame_nr

The frame number in the movie.

VBV_Delay

private int VBV_Delay

The VBV delay value of the frame.

Full_pel_forw_vector

private boolean Full_pel_forw_vector

Equals true if full pixel accuracy is used for forward prediction, equals false if half pixel accuracy is used.

forw_f_code

private int forw_f_code

The motion code value for forward prediction.

forward_f

private int forward_f

The the size of the motion data for forward prediction.

forward_r_size

private int forward_r_size

The size of residual value for forward prediction.

motion_horiz_forw_code

private int motion_horiz_forw_code

The motion code value of horizontal part for forward prediction.

motion_horiz_forw_r

private int motion_horiz_forw_r

The motion residual value of horizontal part for forward prediction.

motion_verti_forw_code

private int motion_verti_forw_code

The motion code value of vertical part for forward prediction.

motion_verti_forw_r

private int motion_verti_forw_r

The motion residual value of vertical part for forward prediction.

Full_pel_back_vector

private boolean Full_pel_back_vector

Equals true if full pixel accuracy is used for backward prediction, equals false if half pixel accuracy is used.

backward_f_code

private int backward_f_code

The motion code value for backward prediction.

backward_f

private int backward_f

The the size of the motion data for backward prediction.

backward_r_size

private int backward_r_size

The size of residual value for backward prediction.

motion_horiz_back_code

private int motion_horiz_back_code

The motion code value of horizontal part for backward prediction.

motion_horiz_back_r

private int motion_horiz_back_r

The motion residual value of horizontal part for backward prediction.

motion_verti_back_code

private int motion_verti_back_code

The motion code value of vertical part for backward prediction.

motion_verti_back_r

private int motion_verti_back_r

The motion residual value of vertical part for backward prediction.

Quant_scale

private int Quant_scale

The quantization factor.

macro_block_address

private int macro_block_address

The actual MB address.

past_intra_address

private int past_intra_address

The past MB address.

mb_row

private int mb_row

The actual row position of the macro block.

mb_column

private int mb_column

The actual column position of the macro block.

macro_block_motion_forward

private boolean macro_block_motion_forward

True if motion vector for forward prediction exists.

macro_block_motion_backward

private boolean macro_block_motion_backward

True if motion vector for backward prediction exists.

macro_block_pattern

private boolean macro_block_pattern

True if a coded block pattern supplied.

macro_block_quant

private boolean macro_block_quant

True if a new quantization factor is supplied.

macro_block_intra

private boolean macro_block_intra

True if a macro block is for intra coded frames.

pel1

private int[] pel1

The array to grab the referred area from predicted frame.

pel2

private int[] pel2

The array to grab the referred area from predicted frame.

Forward

private motion_data Forward

The object of class "motion_data" to notice and compute the forward motion values.

Backward

private motion_data Backward

The object of class "motion_data" to notice and compute the backward motion values.

pixel_per_lum_line

private int pixel_per_lum_line

The number of pixels per luminance line.

pixel_per_col_line

private int pixel_per_col_line

The number of pixels per colour line.

lum_y_incr

private int lum_y_incr

The increment for the luminance line, equals to pixel_per_lum_line - 8.

col_y_incr

private int col_y_incr

The increment for the luminance line, equals to pixel_per_col_line - 8.

video_thread

private volatile java.lang.Thread video_thread

The "Video Decoder" thread object.

error_count

private int error_count

The current amount of printed video errors.

moribund

private volatile java.lang.Thread moribund

The "Video Decoder" thread that is being stopped.

Constructor Detail

MPEG_video

MPEG_video(MPEG1 play,
           io_tool tool)

Constructs a MPEG_video object, notices the parameters, inits "non_intra" and "null" matrixes and creates an VLC (Hufmann) decoder object.

Parameters:

play - the reference to the applet.

tool - the input movie's data bit stream.

Method Detail

start

public void start()

Starts the "Video Decoder" thread.

stop

public void stop()

Stops the "Video Decoder" thread.

join

public void join()

Waits until the "Video Decoder" thread is dead.

run

public void run()

The method "run" parses the MPEG video stream according to ISO 11172-2 and performs some initial steps. This method is called by the JRE after the "Video Decoder" thread is started.

Specified by:

run in interface java.lang.Runnable

Parse_sequence_header

private void Parse_sequence_header()
                            throws java.lang.InterruptedException,
                                   java.io.EOFException

The method "Parse_sequence_header" parses the sequence header according to ISO 11172-2.

Throws:

java.lang.InterruptedException - if another process interrupts the current process.

java.io.EOFException - if the end of the file has been reached.

Parse_group_of_pictures

private void Parse_group_of_pictures()
                              throws java.lang.InterruptedException,
                                     java.io.EOFException

The method "Parse_group_of_pictures" parses the group of pictures according to ISO 11172-2.

Throws:

java.lang.InterruptedException - if another process interrupts the current process.

java.io.EOFException - if the end of the file has been reached.

Parse_picture

private void Parse_picture()
                    throws java.lang.InterruptedException,
                           java.io.EOFException

The method "Parse_picture" parses a picture according to ISO 11172-2.

Throws:

java.lang.InterruptedException - if another process interrupts the current process.

java.io.EOFException - if the end of the file has been reached.

Parse_slice

private final void Parse_slice()
                        throws java.lang.InterruptedException,
                               java.io.EOFException

The method "Parse_slice" parses a slice according to ISO 11172-2. It determines quantization scale and the macroblock address of the first macro block of the slice.

Throws:

java.lang.InterruptedException - if another process interrupts the current process.

java.io.EOFException - if the end of the file has been reached.

Parse_macroblock

private final void Parse_macroblock(int b_nr)
                             throws java.lang.InterruptedException,
                                    java.io.EOFException

The method "Parse_macroblock" parses a macroblock according to ISO 11172-2. It is one of the most complex methods because of the great variety of the constitution of a macroblock. The constitution and existence of the the most information fields depends on the constitution and existence of information fields before.

Furthermore the decoding process is controlled by this method. In some situations some variables must be reset to some default values or in case of skipped macroblocks implizit values must be applied.

Bear in mind that some variables used in this method are member (class) variables for later reference!

Parameters:

b_nr - the macroblock number in slice.

Throws:

java.lang.InterruptedException - if another process interrupts the current process.

java.io.EOFException - if the end of the file has been reached.

Parse_Block

private final void Parse_Block(int nr)
                        throws java.lang.InterruptedException,
                               java.io.EOFException

The method "Parse_Block" parses a block according to ISO 11172-2. Thereby the DC and AC coefficients are reconstructed and placed into the "dct_recon" field in de-"zigzag"-ed order. After that the IDCT routine is called. The method counts the coefficients and calls a sparse IDCT method if the coefficient count is 1.

Parameters:

nr - the block number (0..6) in macroblock.

Throws:

java.lang.InterruptedException - if another process interrupts the current process.

java.io.EOFException - if the end of the file has been reached.

set_dct_diff

private void set_dct_diff(int dct_diff,
                          int dct_size)

The method "set_dct_diff" computes the DCT difference according to ISO 11172-2. It sets "dct_recon[0]".

Parameters:

dct_diff - the difference of the DC coefficient

dct_size - the size of the DC coefficient.

set_lum_pixel

private void set_lum_pixel(int nr)

The method "set_lum_pixel" takes the re-transformed luminance values and places them at the appropriate position. Note that the variables:

 			pixel_per_lum_line
			mb_row
			mb_column
 

are computed in "run()" as soon it was possible.

set_col_pixel

private void set_col_pixel(int nr)

The method "set_col_pixel" takes the re-transformed chrominance values and places them at the appropriate position.

Note that the variables:

 			pixel_per_col_line
			mb_row
			mb_column
 

are computed in "run()" as soon it was possible.

Parameters:

nr - the block number (0..6) in macroblock.

correct_lum_pixel

private void correct_lum_pixel(int nr)

The method "correct_lum_pixel" is called in predicted macro blocks. Because the values in "dct_recon" are motion compensation information the task of this method is to correct the already supplied (copied) luminance values.

Note that the variables:

 			pixel_per_lum_line
			lum_y_incr
			mb_row
			mb_column
 

are computed in "run()" as soon it was possible.

Parameters:

nr - the block number (0..6) in macroblock.

correct_col_pixel

private void correct_col_pixel(int nr)

The method "correct_col_pixel" is called in predicted macro blocks. Because the values in "dct_recon" are motion compensation information the task of this method is to correct the already supplied (copied) chrominance values.

Note that the variables:

 			pixel_per_col_line
			col_y_incr
			mb_row
			mb_column
 

are computed in "run()" as soon it was possible.

Parameters:

nr - the block number (0..6) in macroblock.