com.sun.image.codec.jpeg
Interface JPEGEncodeParam

All Superinterfaces:
Cloneable, JPEGDecodeParam

public interface JPEGEncodeParam
extends Cloneable, JPEGDecodeParam

JPEGEncodeParam encapsulates tables and options necessary to control encoding of JPEG data streams. Parameters are either set explicitly by the application for encoding, or read from another JPEG header.

When working with BufferedImages, the codec will attempt to match the encoded JPEG COLOR_ID with the ColorModel in the BufferedImage. This is not always possible (the default mappings are listed below). In cases where unsupported conversions are required (or odd image colorspaces are in use) the user must either convert the image data to a known ColorSpace or encode the data from a raster. When encoding rasters no colorspace adjustments are made, so the user must do any conversions required to get to the encoded COLOR_ID. The COLOR_ID for the encoded images is used to control the JPEG codec's inital values for Huffman and Quantization Tables as well as subsampling factors. It is also used to determine what color conversion should be performed to obtain the best encoding.

Note: The color ids described herein are simply enumerated values that influence data processing by the JPEG codec. JPEG compression is, by definition, color blind. These values are used as hints when compressing JPEG data. Through these values the JPEG codec can perform some default rotation of data into spaces that will aid in getting better compression ratios.

Example behaviour is described below. Since these mappings are likely to change in the future it is strongly recommended that you make use of the @see JPEGImageEncoder.getDefaultParamBlock calls and check the encodedColorID for your particular BufferedImage. In extreme cases is may be necessary for the user to convert the image to the desired colorspace, and encode it from a Raster. In this case the API programmer must specify the colorID of the data in the Raster and no color conversion will take place.

 ENCODING:
 
 BufferedImage Type/Instance        JPEG (Encoded) Color ID
 ========================       =======================
 TYPE_GRAY                      COLOR_ID_GRAYSCALE
 TYPE_RGB                       COLOR_ID_YCbCr
 TYPE_YCbCr                     COLOR_ID_YCbCr
 TYPE_YCbCr/CS_PYCC             COLOR_ID_PYCC
 TYPE_CMYK                      COLOR_ID_CMYK
 TYPE_RGB   (w/ alpha)          COLOR_ID_YCbCrA
 TYPE_YCbCr (w/ alpha)          COLOR_ID_YCbCrA
 TYPE_YCbCr/CS_PYCC (w/ alpha)  COLOR_ID_PYCCA
 Any Other **                COLOR_ID_UNKNOWN
 
When the user wants more control than the BufferedImage conversions provide, the user must encode the data from a Raster. In this case the data undergoes no color conversion at all. It is the user's responsiblity to perform the desired conversions.

If you intend to write a JFIF image (by including the APP0_MARKER) the encoded COLOR_ID must be one of: COLOR_ID_UNKNOWN, COLOR_ID_GRAYSCALE, COLOR_ID_YCbCr, or COLOR_ID_CMYK. In all other instances an ImageformatException will be thrown.

IMPORTANT: an Alpha RGB BufferedImage will not map to a valid JFIF stream, you must strip off the alpha prior to encoding if you want a JFIF file. If the APP0 marker is set and you do not strip off the Alpha, an ImageFormatException will be thrown.

Note that the classes in the com.sun.image.codec.jpeg package are not part of the core Java APIs. They are a part of Sun's JDK and JRE distributions. Although other licensees may choose to distribute these classes, developers cannot depend on their availability in non-Sun implementations. We expect that equivalent functionality will eventually be available in a core API or standard extension.


Field Summary
 
Fields inherited from interface com.sun.image.codec.jpeg.JPEGDecodeParam
APP0_MARKER, APP1_MARKER, APP2_MARKER, APP3_MARKER, APP4_MARKER, APP5_MARKER, APP6_MARKER, APP7_MARKER, APP8_MARKER, APP9_MARKER, APPA_MARKER, APPB_MARKER, APPC_MARKER, APPD_MARKER, APPE_MARKER, APPF_MARKER, COLOR_ID_CMYK, COLOR_ID_GRAY, COLOR_ID_PYCC, COLOR_ID_PYCCA, COLOR_ID_RGB, COLOR_ID_RGBA, COLOR_ID_RGBA_INVERTED, COLOR_ID_UNKNOWN, COLOR_ID_YCbCr, COLOR_ID_YCbCrA, COLOR_ID_YCbCrA_INVERTED, COLOR_ID_YCCK, COMMENT_MARKER, DENSITY_UNIT_ASPECT_RATIO, DENSITY_UNIT_DOTS_CM, DENSITY_UNIT_DOTS_INCH, NUM_COLOR_ID, NUM_DENSITY_UNIT, NUM_TABLES
 
Method Summary
 void addMarkerData(int marker, byte[] data)
          Appends 'data' to the array of byte[] associated with marker.
 Object clone()
           
 void setACHuffmanComponentMapping(int component, int table)
          Sets the mapping between a component and it's AC Huffman Table.
 void setACHuffmanTable(int tableNum, JPEGHuffmanTable huffTable)
          Sets the AC Huffman coding table at index to the table provided.
 void setDCHuffmanComponentMapping(int component, int table)
          Sets the mapping between a component and it's DC Huffman Table.
 void setDCHuffmanTable(int tableNum, JPEGHuffmanTable huffTable)
          Sets the DC Huffman coding table at index to the table provided.
 void setDensityUnit(int unit)
          Set the pixel size units This value is copied into the APP0 marker (if that marker is written).
 void setHorizontalSubsampling(int component, int subsample)
          Set the horizontal subsample factor for the given component.
 void setImageInfoValid(boolean flag)
          Set the flag indicating the validity of the table information in the ParamBlock.
 void setMarkerData(int marker, byte[][] data)
          Sets the marker data to be written to the output data stream.
 void setQTable(int tableNum, JPEGQTable qTable)
          Sets the coefficient quantization tables at index passed.
 void setQTableComponentMapping(int component, int table)
          Sets the mapping between a component and it's Quantization Table.
 void setQuality(float quality, boolean forceBaseline)
          This creates new Quantization tables that replace the currently installed Quantization tables.
 void setRestartInterval(int restartInterval)
          Set the MCUs per restart, or 0 for no restart markers.
 void setTableInfoValid(boolean flag)
          Set the flag indicating the validity of the image information in the ParamBlock.
 void setVerticalSubsampling(int component, int subsample)
          Set the vertical subsample factor for the given component.
 void setXDensity(int density)
          Set the horizontal pixel density This value is written into the APP0 marker.
 void setYDensity(int density)
          Set the vertical pixel density.
 
Methods inherited from interface com.sun.image.codec.jpeg.JPEGDecodeParam
getACHuffmanComponentMapping, getACHuffmanTable, getACHuffmanTableForComponent, getDCHuffmanComponentMapping, getDCHuffmanTable, getDCHuffmanTableForComponent, getDensityUnit, getEncodedColorID, getHeight, getHorizontalSubsampling, getMarker, getMarkerData, getNumComponents, getQTable, getQTableComponentMapping, getQTableForComponent, getRestartInterval, getVerticalSubsampling, getWidth, getXDensity, getYDensity, isImageInfoValid, isTableInfoValid
 

Method Detail

clone

public Object clone()
Specified by:
clone in interface JPEGDecodeParam

setHorizontalSubsampling

public void setHorizontalSubsampling(int component,
                                     int subsample)
Set the horizontal subsample factor for the given component. Note that the subsample factor is the number of input pixels that contribute to each output pixel (ussually 2 for YCC).

Parameters:
component - The component being specified.
subsample - The subsampling factor being specified.

setVerticalSubsampling

public void setVerticalSubsampling(int component,
                                   int subsample)
Set the vertical subsample factor for the given component. Note that the subsample factor is the number of input pixels that contribute to each output pixel (ussually 2 for YCC).

Parameters:
component - The component being specified.
subsample - The subsampling factor being specified.

setQTable

public void setQTable(int tableNum,
                      JPEGQTable qTable)
Sets the coefficient quantization tables at index passed. tableNum must range in value from 0 - 3.

Parameters:
tableNum - the index of the table to be set.

setDCHuffmanTable

public void setDCHuffmanTable(int tableNum,
                              JPEGHuffmanTable huffTable)
Sets the DC Huffman coding table at index to the table provided.

Parameters:
huffTable - JPEGHuffmanTable that will be assigned to index tableNum.
tableNum - - the index of the table to be set.
Throws:
IllegalArgumentException - - thrown if the tableNum is out of range. Index must range in value from 0 - 3.

setACHuffmanTable

public void setACHuffmanTable(int tableNum,
                              JPEGHuffmanTable huffTable)
Sets the AC Huffman coding table at index to the table provided.

Parameters:
huffTable - JPEGHuffmanTable that will be assigned to index tableNum.
tableNum - - the index of the table to be set.
Throws:
IllegalArgumentException - - thrown if the tableNum is out of range. Index must range in value from 0 - 3.

setDCHuffmanComponentMapping

public void setDCHuffmanComponentMapping(int component,
                                         int table)
Sets the mapping between a component and it's DC Huffman Table.

Parameters:
component - The component to set the mapping for
table - The DC Huffman table to use for component

setACHuffmanComponentMapping

public void setACHuffmanComponentMapping(int component,
                                         int table)
Sets the mapping between a component and it's AC Huffman Table.

Parameters:
component - The component to set the mapping for
table - The AC Huffman table to use for component

setQTableComponentMapping

public void setQTableComponentMapping(int component,
                                      int table)
Sets the mapping between a component and it's Quantization Table.

Parameters:
component - The component to set the mapping for
table - The Quantization Table to use for component

setImageInfoValid

public void setImageInfoValid(boolean flag)
Set the flag indicating the validity of the table information in the ParamBlock. This is used to indicate if tables should be included when encoding.


setTableInfoValid

public void setTableInfoValid(boolean flag)
Set the flag indicating the validity of the image information in the ParamBlock. This is used to indicates if image data should be written when encoding.


setMarkerData

public void setMarkerData(int marker,
                          byte[][] data)
Sets the marker data to be written to the output data stream. This removes any existing marker data in the JPEParm object. This can be used to remove the default APP0 marker by calling it with data set to null.

Parameters:
marker - The marker to set the data for.
data - the new set of data to be written.

addMarkerData

public void addMarkerData(int marker,
                          byte[] data)
Appends 'data' to the array of byte[] associated with marker. This will result in additional instance of the marker being written (one for each byte[] in the array.).

Parameters:
marker - The marker to add and instance of.
data - the data to be written.

setRestartInterval

public void setRestartInterval(int restartInterval)
Set the MCUs per restart, or 0 for no restart markers.

Parameters:
restartInterval - number MCUs per restart marker.

setDensityUnit

public void setDensityUnit(int unit)
Set the pixel size units This value is copied into the APP0 marker (if that marker is written). This value isn't used by the JPEG code.


setXDensity

public void setXDensity(int density)
Set the horizontal pixel density This value is written into the APP0 marker. It isn't used by the JPEG code.


setYDensity

public void setYDensity(int density)
Set the vertical pixel density. This value is copied into the JFIF APP0 marker. It isn't used by the JPEG code.

Parameters:
density - The verticle pixel density, in units described by @see JPEGParam.getDensityUnit.

setQuality

public void setQuality(float quality,
                       boolean forceBaseline)
This creates new Quantization tables that replace the currently installed Quantization tables. It also updates the Component QTable mapping to the default for the current encoded COLOR_ID. The Created Quantization table varies from very high compression, very low quality, (0.0) to low compression, very high quality (1.0) based on the quality parameter.

At a quality level of 1.0 the table will be all 1's which will lead to no loss of data due to quantization (however chrominace subsampling, if used, and roundoff error in the DCT will still degrade the image some what).

This is a linear manipulation of the standard Chrominance Q-Table.

Some guidelines: 0.75 high quality
                 0.5  medium quality
                 0.25 low quality
 

Parameters:
quality - 0.0-1.0 setting of desired quality level.
forceBaseline - force baseline quantization table


Submit a bug or feature
Java is a trademark or registered trademark of Sun Microsystems, Inc. in the US and other countries.
Copyright 1993-2002 Sun Microsystems, Inc. 901 San Antonio Road
Palo Alto, California, 94303, U.S.A. All Rights Reserved.