TurboJPEG 3.0.1
Loading...
Searching...
No Matches
Data Structures | Macros | Typedefs | Enumerations | Functions | Variables
TurboJPEG

TurboJPEG API. More...

Data Structures

struct  tjscalingfactor
 Scaling factor. More...
 
struct  tjregion
 Cropping region. More...
 
struct  tjtransform
 Lossless transform. More...
 

Macros

#define TJ_NUMINIT
 The number of initialization options.
 
#define TJ_NUMSAMP
 The number of chrominance subsampling options.
 
#define TJ_NUMPF
 The number of pixel formats.
 
#define TJ_NUMCS
 The number of JPEG colorspaces.
 
#define TJ_NUMERR
 The number of error codes.
 
#define TJ_NUMXOP
 The number of transform operations.
 
#define TJXOPT_PERFECT
 This option will cause tj3Transform() to return an error if the transform is not perfect.
 
#define TJXOPT_TRIM
 This option will cause tj3Transform() to discard any partial MCU blocks that cannot be transformed.
 
#define TJXOPT_CROP
 This option will enable lossless cropping.
 
#define TJXOPT_GRAY
 This option will discard the color data in the source image and produce a grayscale destination image.
 
#define TJXOPT_NOOUTPUT
 This option will prevent tj3Transform() from outputting a JPEG image for this particular transform.
 
#define TJXOPT_PROGRESSIVE
 This option will enable progressive entropy coding in the JPEG image generated by this particular transform.
 
#define TJXOPT_COPYNONE
 This option will prevent tj3Transform() from copying any extra markers (including EXIF and ICC profile data) from the source image to the destination image.
 
#define TJXOPT_ARITHMETIC
 This option will enable arithmetic entropy coding in the JPEG image generated by this particular transform.
 
#define TJXOPT_OPTIMIZE
 This option will enable optimized baseline entropy coding in the JPEG image generated by this particular transform.
 
#define TJSCALED(dimension, scalingFactor)
 Compute the scaled value of dimension using the given scaling factor.
 

Typedefs

typedef struct tjtransform tjtransform
 Lossless transform.
 
typedef void * tjhandle
 TurboJPEG instance handle.
 

Enumerations

enum  TJINIT { TJINIT_COMPRESS , TJINIT_DECOMPRESS , TJINIT_TRANSFORM }
 Initialization options. More...
 
enum  TJSAMP {
  TJSAMP_444 , TJSAMP_422 , TJSAMP_420 , TJSAMP_GRAY ,
  TJSAMP_440 , TJSAMP_411 , TJSAMP_441 , TJSAMP_UNKNOWN
}
 Chrominance subsampling options. More...
 
enum  TJPF {
  TJPF_RGB , TJPF_BGR , TJPF_RGBX , TJPF_BGRX ,
  TJPF_XBGR , TJPF_XRGB , TJPF_GRAY , TJPF_RGBA ,
  TJPF_BGRA , TJPF_ABGR , TJPF_ARGB , TJPF_CMYK ,
  TJPF_UNKNOWN
}
 Pixel formats. More...
 
enum  TJCS {
  TJCS_RGB , TJCS_YCbCr , TJCS_GRAY , TJCS_CMYK ,
  TJCS_YCCK
}
 JPEG colorspaces. More...
 
enum  TJPARAM {
  TJPARAM_STOPONWARNING , TJPARAM_BOTTOMUP , TJPARAM_NOREALLOC , TJPARAM_QUALITY ,
  TJPARAM_SUBSAMP , TJPARAM_JPEGWIDTH , TJPARAM_JPEGHEIGHT , TJPARAM_PRECISION ,
  TJPARAM_COLORSPACE , TJPARAM_FASTUPSAMPLE , TJPARAM_FASTDCT , TJPARAM_OPTIMIZE ,
  TJPARAM_PROGRESSIVE , TJPARAM_SCANLIMIT , TJPARAM_ARITHMETIC , TJPARAM_LOSSLESS ,
  TJPARAM_LOSSLESSPSV , TJPARAM_LOSSLESSPT , TJPARAM_RESTARTBLOCKS , TJPARAM_RESTARTROWS ,
  TJPARAM_XDENSITY , TJPARAM_YDENSITY , TJPARAM_DENSITYUNITS , TJPARAM_MAXMEMORY ,
  TJPARAM_MAXPIXELS
}
 Parameters. More...
 
enum  TJERR { TJERR_WARNING , TJERR_FATAL }
 Error codes. More...
 
enum  TJXOP {
  TJXOP_NONE , TJXOP_HFLIP , TJXOP_VFLIP , TJXOP_TRANSPOSE ,
  TJXOP_TRANSVERSE , TJXOP_ROT90 , TJXOP_ROT180 , TJXOP_ROT270
}
 Transform operations for tj3Transform() More...
 

Functions

DLLEXPORT tjhandle tj3Init (int initType)
 Create a new TurboJPEG instance.
 
DLLEXPORT int tj3Set (tjhandle handle, int param, int value)
 Set the value of a parameter.
 
DLLEXPORT int tj3Get (tjhandle handle, int param)
 Get the value of a parameter.
 
DLLEXPORT int tj3Compress8 (tjhandle handle, const unsigned char *srcBuf, int width, int pitch, int height, int pixelFormat, unsigned char **jpegBuf, size_t *jpegSize)
 Compress an 8-bit-per-sample packed-pixel RGB, grayscale, or CMYK image into an 8-bit-per-sample JPEG image.
 
DLLEXPORT int tj3Compress12 (tjhandle handle, const short *srcBuf, int width, int pitch, int height, int pixelFormat, unsigned char **jpegBuf, size_t *jpegSize)
 Compress a 12-bit-per-sample packed-pixel RGB, grayscale, or CMYK image into a 12-bit-per-sample JPEG image.
 
DLLEXPORT int tj3Compress16 (tjhandle handle, const unsigned short *srcBuf, int width, int pitch, int height, int pixelFormat, unsigned char **jpegBuf, size_t *jpegSize)
 Compress a 16-bit-per-sample packed-pixel RGB, grayscale, or CMYK image into a 16-bit-per-sample lossless JPEG image.
 
DLLEXPORT int tj3CompressFromYUV8 (tjhandle handle, const unsigned char *srcBuf, int width, int align, int height, unsigned char **jpegBuf, size_t *jpegSize)
 Compress an 8-bit-per-sample unified planar YUV image into an 8-bit-per-sample JPEG image.
 
DLLEXPORT int tj3CompressFromYUVPlanes8 (tjhandle handle, const unsigned char *const *srcPlanes, int width, const int *strides, int height, unsigned char **jpegBuf, size_t *jpegSize)
 Compress a set of 8-bit-per-sample Y, U (Cb), and V (Cr) image planes into an 8-bit-per-sample JPEG image.
 
DLLEXPORT size_t tj3JPEGBufSize (int width, int height, int jpegSubsamp)
 The maximum size of the buffer (in bytes) required to hold a JPEG image with the given parameters.
 
DLLEXPORT size_t tj3YUVBufSize (int width, int align, int height, int subsamp)
 The size of the buffer (in bytes) required to hold a unified planar YUV image with the given parameters.
 
DLLEXPORT size_t tj3YUVPlaneSize (int componentID, int width, int stride, int height, int subsamp)
 The size of the buffer (in bytes) required to hold a YUV image plane with the given parameters.
 
DLLEXPORT int tj3YUVPlaneWidth (int componentID, int width, int subsamp)
 The plane width of a YUV image plane with the given parameters.
 
DLLEXPORT int tj3YUVPlaneHeight (int componentID, int height, int subsamp)
 The plane height of a YUV image plane with the given parameters.
 
DLLEXPORT int tj3EncodeYUV8 (tjhandle handle, const unsigned char *srcBuf, int width, int pitch, int height, int pixelFormat, unsigned char *dstBuf, int align)
 Encode an 8-bit-per-sample packed-pixel RGB or grayscale image into an 8-bit-per-sample unified planar YUV image.
 
DLLEXPORT int tj3EncodeYUVPlanes8 (tjhandle handle, const unsigned char *srcBuf, int width, int pitch, int height, int pixelFormat, unsigned char **dstPlanes, int *strides)
 Encode an 8-bit-per-sample packed-pixel RGB or grayscale image into separate 8-bit-per-sample Y, U (Cb), and V (Cr) image planes.
 
DLLEXPORT int tj3DecompressHeader (tjhandle handle, const unsigned char *jpegBuf, size_t jpegSize)
 Retrieve information about a JPEG image without decompressing it, or prime the decompressor with quantization and Huffman tables.
 
DLLEXPORT tjscalingfactortj3GetScalingFactors (int *numScalingFactors)
 Returns a list of fractional scaling factors that the JPEG decompressor supports.
 
DLLEXPORT int tj3SetScalingFactor (tjhandle handle, tjscalingfactor scalingFactor)
 Set the scaling factor for subsequent lossy decompression operations.
 
DLLEXPORT int tj3SetCroppingRegion (tjhandle handle, tjregion croppingRegion)
 Set the cropping region for partially decompressing a lossy JPEG image into a packed-pixel image.
 
DLLEXPORT int tj3Decompress8 (tjhandle handle, const unsigned char *jpegBuf, size_t jpegSize, unsigned char *dstBuf, int pitch, int pixelFormat)
 Decompress an 8-bit-per-sample JPEG image into an 8-bit-per-sample packed-pixel RGB, grayscale, or CMYK image.
 
DLLEXPORT int tj3Decompress12 (tjhandle handle, const unsigned char *jpegBuf, size_t jpegSize, short *dstBuf, int pitch, int pixelFormat)
 Decompress a 12-bit-per-sample JPEG image into a 12-bit-per-sample packed-pixel RGB, grayscale, or CMYK image.
 
DLLEXPORT int tj3Decompress16 (tjhandle handle, const unsigned char *jpegBuf, size_t jpegSize, unsigned short *dstBuf, int pitch, int pixelFormat)
 Decompress a 16-bit-per-sample lossless JPEG image into a 16-bit-per-sample packed-pixel RGB, grayscale, or CMYK image.
 
DLLEXPORT int tj3DecompressToYUV8 (tjhandle handle, const unsigned char *jpegBuf, size_t jpegSize, unsigned char *dstBuf, int align)
 Decompress an 8-bit-per-sample JPEG image into an 8-bit-per-sample unified planar YUV image.
 
DLLEXPORT int tj3DecompressToYUVPlanes8 (tjhandle handle, const unsigned char *jpegBuf, size_t jpegSize, unsigned char **dstPlanes, int *strides)
 Decompress an 8-bit-per-sample JPEG image into separate 8-bit-per-sample Y, U (Cb), and V (Cr) image planes.
 
DLLEXPORT int tj3DecodeYUV8 (tjhandle handle, const unsigned char *srcBuf, int align, unsigned char *dstBuf, int width, int pitch, int height, int pixelFormat)
 Decode an 8-bit-per-sample unified planar YUV image into an 8-bit-per-sample packed-pixel RGB or grayscale image.
 
DLLEXPORT int tj3DecodeYUVPlanes8 (tjhandle handle, const unsigned char *const *srcPlanes, const int *strides, unsigned char *dstBuf, int width, int pitch, int height, int pixelFormat)
 Decode a set of 8-bit-per-sample Y, U (Cb), and V (Cr) image planes into an 8-bit-per-sample packed-pixel RGB or grayscale image.
 
DLLEXPORT int tj3Transform (tjhandle handle, const unsigned char *jpegBuf, size_t jpegSize, int n, unsigned char **dstBufs, size_t *dstSizes, const tjtransform *transforms)
 Losslessly transform a JPEG image into another JPEG image.
 
DLLEXPORT void tj3Destroy (tjhandle handle)
 Destroy a TurboJPEG instance.
 
DLLEXPORT void * tj3Alloc (size_t bytes)
 Allocate a byte buffer for use with TurboJPEG.
 
DLLEXPORT unsigned char * tj3LoadImage8 (tjhandle handle, const char *filename, int *width, int align, int *height, int *pixelFormat)
 Load an 8-bit-per-sample packed-pixel image from disk into memory.
 
DLLEXPORT short * tj3LoadImage12 (tjhandle handle, const char *filename, int *width, int align, int *height, int *pixelFormat)
 Load a 12-bit-per-sample packed-pixel image from disk into memory.
 
DLLEXPORT unsigned short * tj3LoadImage16 (tjhandle handle, const char *filename, int *width, int align, int *height, int *pixelFormat)
 Load a 16-bit-per-sample packed-pixel image from disk into memory.
 
DLLEXPORT int tj3SaveImage8 (tjhandle handle, const char *filename, const unsigned char *buffer, int width, int pitch, int height, int pixelFormat)
 Save an 8-bit-per-sample packed-pixel image from memory to disk.
 
DLLEXPORT int tj3SaveImage12 (tjhandle handle, const char *filename, const short *buffer, int width, int pitch, int height, int pixelFormat)
 Save a 12-bit-per-sample packed-pixel image from memory to disk.
 
DLLEXPORT int tj3SaveImage16 (tjhandle handle, const char *filename, const unsigned short *buffer, int width, int pitch, int height, int pixelFormat)
 Save a 16-bit-per-sample packed-pixel image from memory to disk.
 
DLLEXPORT void tj3Free (void *buffer)
 Free a byte buffer previously allocated by TurboJPEG.
 
DLLEXPORT char * tj3GetErrorStr (tjhandle handle)
 Returns a descriptive error message explaining why the last command failed.
 
DLLEXPORT int tj3GetErrorCode (tjhandle handle)
 Returns a code indicating the severity of the last error.
 

Variables

static const int tjMCUWidth [TJ_NUMSAMP]
 MCU block width (in pixels) for a given level of chrominance subsampling.
 
static const int tjMCUHeight [TJ_NUMSAMP]
 MCU block height (in pixels) for a given level of chrominance subsampling.
 
static const int tjRedOffset [TJ_NUMPF]
 Red offset (in samples) for a given pixel format.
 
static const int tjGreenOffset [TJ_NUMPF]
 Green offset (in samples) for a given pixel format.
 
static const int tjBlueOffset [TJ_NUMPF]
 Blue offset (in samples) for a given pixel format.
 
static const int tjAlphaOffset [TJ_NUMPF]
 Alpha offset (in samples) for a given pixel format.
 
static const int tjPixelSize [TJ_NUMPF]
 Pixel size (in samples) for a given pixel format.
 
static const tjregion TJUNCROPPED
 A tjregion structure that specifies no cropping.
 
static const tjscalingfactor TJUNSCALED
 A tjscalingfactor structure that specifies a scaling factor of 1/1 (no scaling)
 

Detailed Description

TurboJPEG API.

This API provides an interface for generating, decoding, and transforming planar YUV and JPEG images in memory.

YUV Image Format Notes

Technically, the JPEG format uses the YCbCr colorspace (which is technically not a colorspace but a color transform), but per the convention of the digital video community, the TurboJPEG API uses "YUV" to refer to an image format consisting of Y, Cb, and Cr image planes.

Each plane is simply a 2D array of bytes, each byte representing the value of one of the components (Y, Cb, or Cr) at a particular location in the image. The width and height of each plane are determined by the image width, height, and level of chrominance subsampling. The luminance plane width is the image width padded to the nearest multiple of the horizontal subsampling factor (1 in the case of 4:4:4, grayscale, 4:4:0, or 4:4:1; 2 in the case of 4:2:2 or 4:2:0; 4 in the case of 4:1:1.) Similarly, the luminance plane height is the image height padded to the nearest multiple of the vertical subsampling factor (1 in the case of 4:4:4, 4:2:2, grayscale, or 4:1:1; 2 in the case of 4:2:0 or 4:4:0; 4 in the case of 4:4:1.) This is irrespective of any additional padding that may be specified as an argument to the various YUV functions. The chrominance plane width is equal to the luminance plane width divided by the horizontal subsampling factor, and the chrominance plane height is equal to the luminance plane height divided by the vertical subsampling factor.

For example, if the source image is 35 x 35 pixels and 4:2:2 subsampling is used, then the luminance plane would be 36 x 35 bytes, and each of the chrominance planes would be 18 x 35 bytes. If you specify a row alignment of 4 bytes on top of this, then the luminance plane would be 36 x 35 bytes, and each of the chrominance planes would be 20 x 35 bytes.

Macro Definition Documentation

◆ TJ_NUMCS

#define TJ_NUMCS

The number of JPEG colorspaces.

◆ TJ_NUMERR

#define TJ_NUMERR

The number of error codes.

◆ TJ_NUMINIT

#define TJ_NUMINIT

The number of initialization options.

◆ TJ_NUMPF

#define TJ_NUMPF

The number of pixel formats.

◆ TJ_NUMSAMP

#define TJ_NUMSAMP

The number of chrominance subsampling options.

◆ TJ_NUMXOP

#define TJ_NUMXOP

The number of transform operations.

◆ TJSCALED

#define TJSCALED (   dimension,
  scalingFactor 
)

Compute the scaled value of dimension using the given scaling factor.

This macro performs the integer equivalent of ceil(dimension * scalingFactor).

◆ TJXOPT_ARITHMETIC

#define TJXOPT_ARITHMETIC

This option will enable arithmetic entropy coding in the JPEG image generated by this particular transform.

Arithmetic entropy coding will generally improve compression relative to Huffman entropy coding (the default), but it will reduce decompression performance considerably. Can be combined with TJXOPT_PROGRESSIVE.

◆ TJXOPT_COPYNONE

#define TJXOPT_COPYNONE

This option will prevent tj3Transform() from copying any extra markers (including EXIF and ICC profile data) from the source image to the destination image.

◆ TJXOPT_CROP

#define TJXOPT_CROP

This option will enable lossless cropping.

See tj3Transform() for more information.

◆ TJXOPT_GRAY

#define TJXOPT_GRAY

This option will discard the color data in the source image and produce a grayscale destination image.

◆ TJXOPT_NOOUTPUT

#define TJXOPT_NOOUTPUT

This option will prevent tj3Transform() from outputting a JPEG image for this particular transform.

(This can be used in conjunction with a custom filter to capture the transformed DCT coefficients without transcoding them.)

◆ TJXOPT_OPTIMIZE

#define TJXOPT_OPTIMIZE

This option will enable optimized baseline entropy coding in the JPEG image generated by this particular transform.

Optimized baseline entropy coding will improve compression slightly (generally 5% or less.)

◆ TJXOPT_PERFECT

#define TJXOPT_PERFECT

This option will cause tj3Transform() to return an error if the transform is not perfect.

Lossless transforms operate on MCU blocks, whose size depends on the level of chrominance subsampling used (see tjMCUWidth and tjMCUHeight.) If the image's width or height is not evenly divisible by the MCU block size, then there will be partial MCU blocks on the right and/or bottom edges. It is not possible to move these partial MCU blocks to the top or left of the image, so any transform that would require that is "imperfect." If this option is not specified, then any partial MCU blocks that cannot be transformed will be left in place, which will create odd-looking strips on the right or bottom edge of the image.

◆ TJXOPT_PROGRESSIVE

#define TJXOPT_PROGRESSIVE

This option will enable progressive entropy coding in the JPEG image generated by this particular transform.

Progressive entropy coding will generally improve compression relative to baseline entropy coding (the default), but it will reduce decompression performance considerably. Can be combined with TJXOPT_ARITHMETIC. Implies TJXOPT_OPTIMIZE unless TJXOPT_ARITHMETIC is also specified.

◆ TJXOPT_TRIM

#define TJXOPT_TRIM

This option will cause tj3Transform() to discard any partial MCU blocks that cannot be transformed.

Typedef Documentation

◆ tjhandle

typedef void* tjhandle

TurboJPEG instance handle.

◆ tjtransform

typedef struct tjtransform tjtransform

Lossless transform.

Enumeration Type Documentation

◆ TJCS

enum TJCS

JPEG colorspaces.

Enumerator
TJCS_RGB 

RGB colorspace.

When compressing the JPEG image, the R, G, and B components in the source image are reordered into image planes, but no colorspace conversion or subsampling is performed. RGB JPEG images can be compressed from and decompressed to packed-pixel images with any of the extended RGB or grayscale pixel formats, but they cannot be compressed from or decompressed to planar YUV images.

TJCS_YCbCr 

YCbCr colorspace.

YCbCr is not an absolute colorspace but rather a mathematical transformation of RGB designed solely for storage and transmission. YCbCr images must be converted to RGB before they can actually be displayed. In the YCbCr colorspace, the Y (luminance) component represents the black & white portion of the original image, and the Cb and Cr (chrominance) components represent the color portion of the original image. Originally, the analog equivalent of this transformation allowed the same signal to drive both black & white and color televisions, but JPEG images use YCbCr primarily because it allows the color data to be optionally subsampled for the purposes of reducing network or disk usage. YCbCr is the most common JPEG colorspace, and YCbCr JPEG images can be compressed from and decompressed to packed-pixel images with any of the extended RGB or grayscale pixel formats. YCbCr JPEG images can also be compressed from and decompressed to planar YUV images.

TJCS_GRAY 

Grayscale colorspace.

The JPEG image retains only the luminance data (Y component), and any color data from the source image is discarded. Grayscale JPEG images can be compressed from and decompressed to packed-pixel images with any of the extended RGB or grayscale pixel formats, or they can be compressed from and decompressed to planar YUV images.

TJCS_CMYK 

CMYK colorspace.

When compressing the JPEG image, the C, M, Y, and K components in the source image are reordered into image planes, but no colorspace conversion or subsampling is performed. CMYK JPEG images can only be compressed from and decompressed to packed-pixel images with the CMYK pixel format.

TJCS_YCCK 

YCCK colorspace.

YCCK (AKA "YCbCrK") is not an absolute colorspace but rather a mathematical transformation of CMYK designed solely for storage and transmission. It is to CMYK as YCbCr is to RGB. CMYK pixels can be reversibly transformed into YCCK, and as with YCbCr, the chrominance components in the YCCK pixels can be subsampled without incurring major perceptual loss. YCCK JPEG images can only be compressed from and decompressed to packed-pixel images with the CMYK pixel format.

◆ TJERR

enum TJERR

Error codes.

Enumerator
TJERR_WARNING 

The error was non-fatal and recoverable, but the destination image may still be corrupt.

TJERR_FATAL 

The error was fatal and non-recoverable.

◆ TJINIT

enum TJINIT

Initialization options.

Enumerator
TJINIT_COMPRESS 

Initialize the TurboJPEG instance for compression.

TJINIT_DECOMPRESS 

Initialize the TurboJPEG instance for decompression.

TJINIT_TRANSFORM 

Initialize the TurboJPEG instance for lossless transformation (both compression and decompression.)

◆ TJPARAM

enum TJPARAM

Parameters.

Enumerator
TJPARAM_STOPONWARNING 

Error handling behavior.

Value

  • 0 [default] Allow the current compression/decompression/transform operation to complete unless a fatal error is encountered.
  • 1 Immediately discontinue the current compression/decompression/transform operation if a warning (non-fatal error) occurs.
TJPARAM_BOTTOMUP 

Row order in packed-pixel source/destination images.

Value

  • 0 [default] top-down (X11) order
  • 1 bottom-up (Windows, OpenGL) order
TJPARAM_NOREALLOC 

JPEG destination buffer (re)allocation [compression, lossless transformation].

Value

  • 0 [default] Attempt to allocate or reallocate the JPEG destination buffer as needed.
  • 1 Generate an error if the JPEG destination buffer is invalid or too small.
TJPARAM_QUALITY 

Perceptual quality of lossy JPEG images [compression only].

Value

  • 1-100 (1 = worst quality but best compression, 100 = best quality but worst compression) [no default; must be explicitly specified]
TJPARAM_SUBSAMP 

Chrominance subsampling level.

The JPEG or YUV image uses (decompression, decoding) or will use (lossy compression, encoding) the specified level of chrominance subsampling.

Value

TJPARAM_JPEGWIDTH 

JPEG width (in pixels) [decompression only, read-only].

TJPARAM_JPEGHEIGHT 

JPEG height (in pixels) [decompression only, read-only].

TJPARAM_PRECISION 

JPEG data precision (bits per sample) [decompression only, read-only].

The JPEG image uses the specified number of bits per sample.

Value

  • 8, 12, or 16

12-bit data precision implies TJPARAM_OPTIMIZE unless TJPARAM_ARITHMETIC is set.

TJPARAM_COLORSPACE 

JPEG colorspace.

The JPEG image uses (decompression) or will use (lossy compression) the specified colorspace.

Value

  • One of the JPEG colorspaces [default for lossy compression: automatically selected based on the subsampling level and pixel format]
TJPARAM_FASTUPSAMPLE 

Chrominance upsampling algorithm [lossy decompression only].

Value

  • 0 [default] Use smooth upsampling when decompressing a JPEG image that was compressed using chrominance subsampling. This creates a smooth transition between neighboring chrominance components in order to reduce upsampling artifacts in the decompressed image.
  • 1 Use the fastest chrominance upsampling algorithm available, which may combine upsampling with color conversion.
TJPARAM_FASTDCT 

DCT/IDCT algorithm [lossy compression and decompression].

Value

  • 0 [default] Use the most accurate DCT/IDCT algorithm available.
  • 1 Use the fastest DCT/IDCT algorithm available.

This parameter is provided mainly for backward compatibility with libjpeg, which historically implemented several different DCT/IDCT algorithms because of performance limitations with 1990s CPUs. In the libjpeg-turbo implementation of the TurboJPEG API:

  • The "fast" and "accurate" DCT/IDCT algorithms perform similarly on modern x86/x86-64 CPUs that support AVX2 instructions.
  • The "fast" algorithm is generally only about 5-15% faster than the "accurate" algorithm on other types of CPUs.
  • The difference in accuracy between the "fast" and "accurate" algorithms is the most pronounced at JPEG quality levels above 90 and tends to be more pronounced with decompression than with compression.
  • The "fast" algorithm degrades and is not fully accelerated for JPEG quality levels above 97, so it will be slower than the "accurate" algorithm.
TJPARAM_OPTIMIZE 

Optimized baseline entropy coding [lossy compression only].

Value

  • 0 [default] The JPEG image will use the default Huffman tables.
  • 1 Optimal Huffman tables will be computed for the JPEG image. For lossless transformation, this can also be specified using TJXOPT_OPTIMIZE.

Optimized baseline entropy coding will improve compression slightly (generally 5% or less), but it will reduce compression performance considerably.

TJPARAM_PROGRESSIVE 

Progressive entropy coding.

Value

  • 0 [default for compression, lossless transformation] The lossy JPEG image uses (decompression) or will use (compression, lossless transformation) baseline entropy coding.
  • 1 The lossy JPEG image uses (decompression) or will use (compression, lossless transformation) progressive entropy coding. For lossless transformation, this can also be specified using TJXOPT_PROGRESSIVE.

Progressive entropy coding will generally improve compression relative to baseline entropy coding, but it will reduce compression and decompression performance considerably. Can be combined with TJPARAM_ARITHMETIC. Implies TJPARAM_OPTIMIZE unless TJPARAM_ARITHMETIC is also set.

TJPARAM_SCANLIMIT 

Progressive JPEG scan limit for lossy JPEG images [decompression, lossless transformation].

Setting this parameter will cause the decompression and transform functions to return an error if the number of scans in a progressive JPEG image exceeds the specified limit. The primary purpose of this is to allow security-critical applications to guard against an exploit of the progressive JPEG format described in this report.

Value

  • maximum number of progressive JPEG scans that the decompression and transform functions will process [default: 0 (no limit)]
See also
TJPARAM_PROGRESSIVE
TJPARAM_ARITHMETIC 

Arithmetic entropy coding.

Value

  • 0 [default for compression, lossless transformation] The lossy JPEG image uses (decompression) or will use (compression, lossless transformation) Huffman entropy coding.
  • 1 The lossy JPEG image uses (decompression) or will use (compression, lossless transformation) arithmetic entropy coding. For lossless transformation, this can also be specified using TJXOPT_ARITHMETIC.

Arithmetic entropy coding will generally improve compression relative to Huffman entropy coding, but it will reduce compression and decompression performance considerably. Can be combined with TJPARAM_PROGRESSIVE.

TJPARAM_LOSSLESS 

Lossless JPEG.

Value

  • 0 [default for compression] The JPEG image is (decompression) or will be (compression) lossy/DCT-based.
  • 1 The JPEG image is (decompression) or will be (compression) lossless/predictive.

In most cases, compressing and decompressing lossless JPEG images is considerably slower than compressing and decompressing lossy JPEG images, and lossless JPEG images are much larger than lossy JPEG images. Thus, lossless JPEG images are typically used only for applications that require mathematically lossless compression. Also note that the following features are not available with lossless JPEG images:

  • Colorspace conversion (lossless JPEG images always use TJCS_RGB, TJCS_GRAY, or TJCS_CMYK, depending on the pixel format of the source image)
  • Chrominance subsampling (lossless JPEG images always use TJSAMP_444)
  • JPEG quality selection
  • DCT/IDCT algorithm selection
  • Progressive entropy coding
  • Arithmetic entropy coding
  • Compression from/decompression to planar YUV images
  • Decompression scaling
  • Lossless transformation
See also
TJPARAM_LOSSLESSPSV, TJPARAM_LOSSLESSPT
TJPARAM_LOSSLESSPSV 

Lossless JPEG predictor selection value (PSV)

Value

  • 1-7 [default for compression: 1]

Lossless JPEG compression shares no algorithms with lossy JPEG compression. Instead, it uses differential pulse-code modulation (DPCM), an algorithm whereby each sample is encoded as the difference between the sample's value and a "predictor", which is based on the values of neighboring samples. If Ra is the sample immediately to the left of the current sample, Rb is the sample immediately above the current sample, and Rc is the sample diagonally to the left and above the current sample, then the relationship between the predictor selection value and the predictor is as follows:

PSV Predictor
1 Ra
2 Rb
3 Rc
4 Ra + Rb – Rc
5 Ra + (Rb – Rc) / 2
6 Rb + (Ra – Rc) / 2
7 (Ra + Rb) / 2

Predictors 1-3 are 1-dimensional predictors, whereas Predictors 4-7 are 2-dimensional predictors. The best predictor for a particular image depends on the image.

See also
TJPARAM_LOSSLESS
TJPARAM_LOSSLESSPT 

Lossless JPEG point transform (Pt)

Value

  • 0 through precision - 1, where precision is the JPEG data precision in bits [default for compression: 0]

A point transform value of 0 is necessary in order to generate a fully lossless JPEG image. (A non-zero point transform value right-shifts the input samples by the specified number of bits, which is effectively a form of lossy color quantization.)

See also
TJPARAM_LOSSLESS, TJPARAM_PRECISION
TJPARAM_RESTARTBLOCKS 

JPEG restart marker interval in MCU blocks (lossy) or samples (lossless) [compression only].

The nature of entropy coding is such that a corrupt JPEG image cannot be decompressed beyond the point of corruption unless it contains restart markers. A restart marker stops and restarts the entropy coding algorithm so that, if a JPEG image is corrupted, decompression can resume at the next marker. Thus, adding more restart markers improves the fault tolerance of the JPEG image, but adding too many restart markers can adversely affect the compression ratio and performance.

Value

  • the number of MCU blocks or samples between each restart marker [default: 0 (no restart markers)]

Setting this parameter to a non-zero value sets TJPARAM_RESTARTROWS to 0.

TJPARAM_RESTARTROWS 

JPEG restart marker interval in MCU rows (lossy) or sample rows (lossless) [compression only].

See TJPARAM_RESTARTBLOCKS for a description of restart markers.

Value

  • the number of MCU rows or sample rows between each restart marker [default: 0 (no restart markers)]

Setting this parameter to a non-zero value sets TJPARAM_RESTARTBLOCKS to 0.

TJPARAM_XDENSITY 

JPEG horizontal pixel density.

Value

  • The JPEG image has (decompression) or will have (compression) the specified horizontal pixel density [default for compression: 1].

This value is stored in or read from the JPEG header. It does not affect the contents of the JPEG image. Note that this parameter is set by tj3LoadImage8() when loading a Windows BMP file that contains pixel density information, and the value of this parameter is stored to a Windows BMP file by tj3SaveImage8() if the value of TJPARAM_DENSITYUNITS is 2.

See also
TJPARAM_DENSITYUNITS
TJPARAM_YDENSITY 

JPEG vertical pixel density.

Value

  • The JPEG image has (decompression) or will have (compression) the specified vertical pixel density [default for compression: 1].

This value is stored in or read from the JPEG header. It does not affect the contents of the JPEG image. Note that this parameter is set by tj3LoadImage8() when loading a Windows BMP file that contains pixel density information, and the value of this parameter is stored to a Windows BMP file by tj3SaveImage8() if the value of TJPARAM_DENSITYUNITS is 2.

See also
TJPARAM_DENSITYUNITS
TJPARAM_DENSITYUNITS 

JPEG pixel density units.

Value

  • 0 [default for compression] The pixel density of the JPEG image is expressed (decompression) or will be expressed (compression) in unknown units.
  • 1 The pixel density of the JPEG image is expressed (decompression) or will be expressed (compression) in units of pixels/inch.
  • 2 The pixel density of the JPEG image is expressed (decompression) or will be expressed (compression) in units of pixels/cm.

This value is stored in or read from the JPEG header. It does not affect the contents of the JPEG image. Note that this parameter is set by tj3LoadImage8() when loading a Windows BMP file that contains pixel density information, and the value of this parameter is stored to a Windows BMP file by tj3SaveImage8() if the value is 2.

See also
TJPARAM_XDENSITY, TJPARAM_YDENSITY
TJPARAM_MAXMEMORY 

Memory limit for intermediate buffers.

Value

  • the maximum amount of memory (in megabytes) that will be allocated for intermediate buffers, which are used with progressive JPEG compression and decompression, optimized baseline entropy coding, lossless JPEG compression, and lossless transformation [default: 0 (no limit)]
TJPARAM_MAXPIXELS 

Image size limit [decompression, lossless transformation, packed-pixel image loading].

Setting this parameter will cause the decompression, transform, and image loading functions to return an error if the number of pixels in the source image exceeds the specified limit. This allows security-critical applications to guard against excessive memory consumption.

Value

  • maximum number of pixels that the decompression, transform, and image loading functions will process [default: 0 (no limit)]

◆ TJPF

enum TJPF

Pixel formats.

Enumerator
TJPF_RGB 

RGB pixel format.

The red, green, and blue components in the image are stored in 3-sample pixels in the order R, G, B from lowest to highest memory address within each pixel.

TJPF_BGR 

BGR pixel format.

The red, green, and blue components in the image are stored in 3-sample pixels in the order B, G, R from lowest to highest memory address within each pixel.

TJPF_RGBX 

RGBX pixel format.

The red, green, and blue components in the image are stored in 4-sample pixels in the order R, G, B from lowest to highest memory address within each pixel. The X component is ignored when compressing and undefined when decompressing.

TJPF_BGRX 

BGRX pixel format.

The red, green, and blue components in the image are stored in 4-sample pixels in the order B, G, R from lowest to highest memory address within each pixel. The X component is ignored when compressing and undefined when decompressing.

TJPF_XBGR 

XBGR pixel format.

The red, green, and blue components in the image are stored in 4-sample pixels in the order R, G, B from highest to lowest memory address within each pixel. The X component is ignored when compressing and undefined when decompressing.

TJPF_XRGB 

XRGB pixel format.

The red, green, and blue components in the image are stored in 4-sample pixels in the order B, G, R from highest to lowest memory address within each pixel. The X component is ignored when compressing and undefined when decompressing.

TJPF_GRAY 

Grayscale pixel format.

Each 1-sample pixel represents a luminance (brightness) level from 0 to the maximum sample value (255 for 8-bit samples, 4095 for 12-bit samples, and 65535 for 16-bit samples.)

TJPF_RGBA 

RGBA pixel format.

This is the same as TJPF_RGBX, except that when decompressing, the X component is guaranteed to be equal to the maximum sample value, which can be interpreted as an opaque alpha channel.

TJPF_BGRA 

BGRA pixel format.

This is the same as TJPF_BGRX, except that when decompressing, the X component is guaranteed to be equal to the maximum sample value, which can be interpreted as an opaque alpha channel.

TJPF_ABGR 

ABGR pixel format.

This is the same as TJPF_XBGR, except that when decompressing, the X component is guaranteed to be equal to the maximum sample value, which can be interpreted as an opaque alpha channel.

TJPF_ARGB 

ARGB pixel format.

This is the same as TJPF_XRGB, except that when decompressing, the X component is guaranteed to be equal to the maximum sample value, which can be interpreted as an opaque alpha channel.

TJPF_CMYK 

CMYK pixel format.

Unlike RGB, which is an additive color model used primarily for display, CMYK (Cyan/Magenta/Yellow/Key) is a subtractive color model used primarily for printing. In the CMYK color model, the value of each color component typically corresponds to an amount of cyan, magenta, yellow, or black ink that is applied to a white background. In order to convert between CMYK and RGB, it is necessary to use a color management system (CMS.) A CMS will attempt to map colors within the printer's gamut to perceptually similar colors in the display's gamut and vice versa, but the mapping is typically not 1:1 or reversible, nor can it be defined with a simple formula. Thus, such a conversion is out of scope for a codec library. However, the TurboJPEG API allows for compressing packed-pixel CMYK images into YCCK JPEG images (see TJCS_YCCK) and decompressing YCCK JPEG images into packed-pixel CMYK images.

TJPF_UNKNOWN 

Unknown pixel format.

Currently this is only used by tj3LoadImage8(), tj3LoadImage12(), and tj3LoadImage16().

◆ TJSAMP

enum TJSAMP

Chrominance subsampling options.

When pixels are converted from RGB to YCbCr (see TJCS_YCbCr) or from CMYK to YCCK (see TJCS_YCCK) as part of the JPEG compression process, some of the Cb and Cr (chrominance) components can be discarded or averaged together to produce a smaller image with little perceptible loss of image clarity. (The human eye is more sensitive to small changes in brightness than to small changes in color.) This is called "chrominance subsampling".

Enumerator
TJSAMP_444 

4:4:4 chrominance subsampling (no chrominance subsampling).

The JPEG or YUV image will contain one chrominance component for every pixel in the source image.

TJSAMP_422 

4:2:2 chrominance subsampling.

The JPEG or YUV image will contain one chrominance component for every 2x1 block of pixels in the source image.

TJSAMP_420 

4:2:0 chrominance subsampling.

The JPEG or YUV image will contain one chrominance component for every 2x2 block of pixels in the source image.

TJSAMP_GRAY 

Grayscale.

The JPEG or YUV image will contain no chrominance components.

TJSAMP_440 

4:4:0 chrominance subsampling.

The JPEG or YUV image will contain one chrominance component for every 1x2 block of pixels in the source image.

Note
4:4:0 subsampling is not fully accelerated in libjpeg-turbo.
TJSAMP_411 

4:1:1 chrominance subsampling.

The JPEG or YUV image will contain one chrominance component for every 4x1 block of pixels in the source image. JPEG images compressed with 4:1:1 subsampling will be almost exactly the same size as those compressed with 4:2:0 subsampling, and in the aggregate, both subsampling methods produce approximately the same perceptual quality. However, 4:1:1 is better able to reproduce sharp horizontal features.

Note
4:1:1 subsampling is not fully accelerated in libjpeg-turbo.
TJSAMP_441 

4:4:1 chrominance subsampling.

The JPEG or YUV image will contain one chrominance component for every 1x4 block of pixels in the source image. JPEG images compressed with 4:4:1 subsampling will be almost exactly the same size as those compressed with 4:2:0 subsampling, and in the aggregate, both subsampling methods produce approximately the same perceptual quality. However, 4:4:1 is better able to reproduce sharp vertical features.

Note
4:4:1 subsampling is not fully accelerated in libjpeg-turbo.
TJSAMP_UNKNOWN 

Unknown subsampling.

The JPEG image uses an unusual type of chrominance subsampling. Such images can be decompressed into packed-pixel images, but they cannot be

  • decompressed into planar YUV images,
  • losslessly transformed if TJXOPT_CROP is specified, or
  • partially decompressed using a cropping region.

◆ TJXOP

enum TJXOP

Transform operations for tj3Transform()

Enumerator
TJXOP_NONE 

Do not transform the position of the image pixels.

TJXOP_HFLIP 

Flip (mirror) image horizontally.

This transform is imperfect if there are any partial MCU blocks on the right edge (see TJXOPT_PERFECT.)

TJXOP_VFLIP 

Flip (mirror) image vertically.

This transform is imperfect if there are any partial MCU blocks on the bottom edge (see TJXOPT_PERFECT.)

TJXOP_TRANSPOSE 

Transpose image (flip/mirror along upper left to lower right axis.) This transform is always perfect.

TJXOP_TRANSVERSE 

Transverse transpose image (flip/mirror along upper right to lower left axis.) This transform is imperfect if there are any partial MCU blocks in the image (see TJXOPT_PERFECT.)

TJXOP_ROT90 

Rotate image clockwise by 90 degrees.

This transform is imperfect if there are any partial MCU blocks on the bottom edge (see TJXOPT_PERFECT.)

TJXOP_ROT180 

Rotate image 180 degrees.

This transform is imperfect if there are any partial MCU blocks in the image (see TJXOPT_PERFECT.)

TJXOP_ROT270 

Rotate image counter-clockwise by 90 degrees.

This transform is imperfect if there are any partial MCU blocks on the right edge (see TJXOPT_PERFECT.)

Function Documentation

◆ tj3Alloc()

DLLEXPORT void * tj3Alloc ( size_t  bytes)

Allocate a byte buffer for use with TurboJPEG.

You should always use this function to allocate the JPEG destination buffer(s) for the compression and transform functions unless you are disabling automatic buffer (re)allocation (by setting TJPARAM_NOREALLOC.)

Parameters
bytesthe number of bytes to allocate
Returns
a pointer to a newly-allocated buffer with the specified number of bytes.
See also
tj3Free()

◆ tj3Compress12()

DLLEXPORT int tj3Compress12 ( tjhandle  handle,
const short *  srcBuf,
int  width,
int  pitch,
int  height,
int  pixelFormat,
unsigned char **  jpegBuf,
size_t *  jpegSize 
)

Compress a 12-bit-per-sample packed-pixel RGB, grayscale, or CMYK image into a 12-bit-per-sample JPEG image.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for compression
srcBufpointer to a buffer containing a packed-pixel RGB, grayscale, or CMYK source image to be compressed. This buffer should normally be pitch * height samples in size. However, you can also use this parameter to compress from a specific region of a larger buffer.
widthwidth (in pixels) of the source image
pitchsamples per row in the source image. Normally this should be width * tjPixelSize[pixelFormat], if the image is unpadded. (Setting this parameter to 0 is the equivalent of setting it to width * tjPixelSize[pixelFormat].) However, you can also use this parameter to specify the row alignment/padding of the source image, to skip rows, or to compress from a specific region of a larger buffer.
heightheight (in pixels) of the source image
pixelFormatpixel format of the source image (see Pixel formats.)
jpegBufaddress of a pointer to a byte buffer that will receive the JPEG image. TurboJPEG has the ability to reallocate the JPEG buffer to accommodate the size of the JPEG image. Thus, you can choose to:
  1. pre-allocate the JPEG buffer with an arbitrary size using tj3Alloc() and let TurboJPEG grow the buffer as needed,
  2. set *jpegBuf to NULL to tell TurboJPEG to allocate the buffer for you, or
  3. pre-allocate the buffer to a "worst case" size determined by calling tj3JPEGBufSize(). This should ensure that the buffer never has to be re-allocated. (Setting TJPARAM_NOREALLOC guarantees that it won't be.)
If you choose option 1, then *jpegSize should be set to the size of your pre-allocated buffer. In any case, unless you have set TJPARAM_NOREALLOC, you should always check *jpegBuf upon return from this function, as it may have changed.
jpegSizepointer to a size_t variable that holds the size of the JPEG buffer. If *jpegBuf points to a pre-allocated buffer, then *jpegSize should be set to the size of the buffer. Upon return, *jpegSize will contain the size of the JPEG image (in bytes.) If *jpegBuf points to a JPEG buffer that is being reused from a previous call to one of the JPEG compression functions, then *jpegSize is ignored.
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3Compress16()

DLLEXPORT int tj3Compress16 ( tjhandle  handle,
const unsigned short *  srcBuf,
int  width,
int  pitch,
int  height,
int  pixelFormat,
unsigned char **  jpegBuf,
size_t *  jpegSize 
)

Compress a 16-bit-per-sample packed-pixel RGB, grayscale, or CMYK image into a 16-bit-per-sample lossless JPEG image.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for compression
srcBufpointer to a buffer containing a packed-pixel RGB, grayscale, or CMYK source image to be compressed. This buffer should normally be pitch * height samples in size. However, you can also use this parameter to compress from a specific region of a larger buffer.
widthwidth (in pixels) of the source image
pitchsamples per row in the source image. Normally this should be width * tjPixelSize[pixelFormat], if the image is unpadded. (Setting this parameter to 0 is the equivalent of setting it to width * tjPixelSize[pixelFormat].) However, you can also use this parameter to specify the row alignment/padding of the source image, to skip rows, or to compress from a specific region of a larger buffer.
heightheight (in pixels) of the source image
pixelFormatpixel format of the source image (see Pixel formats.)
jpegBufaddress of a pointer to a byte buffer that will receive the JPEG image. TurboJPEG has the ability to reallocate the JPEG buffer to accommodate the size of the JPEG image. Thus, you can choose to:
  1. pre-allocate the JPEG buffer with an arbitrary size using tj3Alloc() and let TurboJPEG grow the buffer as needed,
  2. set *jpegBuf to NULL to tell TurboJPEG to allocate the buffer for you, or
  3. pre-allocate the buffer to a "worst case" size determined by calling tj3JPEGBufSize(). This should ensure that the buffer never has to be re-allocated. (Setting TJPARAM_NOREALLOC guarantees that it won't be.)
If you choose option 1, then *jpegSize should be set to the size of your pre-allocated buffer. In any case, unless you have set TJPARAM_NOREALLOC, you should always check *jpegBuf upon return from this function, as it may have changed.
jpegSizepointer to a size_t variable that holds the size of the JPEG buffer. If *jpegBuf points to a pre-allocated buffer, then *jpegSize should be set to the size of the buffer. Upon return, *jpegSize will contain the size of the JPEG image (in bytes.) If *jpegBuf points to a JPEG buffer that is being reused from a previous call to one of the JPEG compression functions, then *jpegSize is ignored.
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3Compress8()

DLLEXPORT int tj3Compress8 ( tjhandle  handle,
const unsigned char *  srcBuf,
int  width,
int  pitch,
int  height,
int  pixelFormat,
unsigned char **  jpegBuf,
size_t *  jpegSize 
)

Compress an 8-bit-per-sample packed-pixel RGB, grayscale, or CMYK image into an 8-bit-per-sample JPEG image.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for compression
srcBufpointer to a buffer containing a packed-pixel RGB, grayscale, or CMYK source image to be compressed. This buffer should normally be pitch * height samples in size. However, you can also use this parameter to compress from a specific region of a larger buffer.
widthwidth (in pixels) of the source image
pitchsamples per row in the source image. Normally this should be width * tjPixelSize[pixelFormat], if the image is unpadded. (Setting this parameter to 0 is the equivalent of setting it to width * tjPixelSize[pixelFormat].) However, you can also use this parameter to specify the row alignment/padding of the source image, to skip rows, or to compress from a specific region of a larger buffer.
heightheight (in pixels) of the source image
pixelFormatpixel format of the source image (see Pixel formats.)
jpegBufaddress of a pointer to a byte buffer that will receive the JPEG image. TurboJPEG has the ability to reallocate the JPEG buffer to accommodate the size of the JPEG image. Thus, you can choose to:
  1. pre-allocate the JPEG buffer with an arbitrary size using tj3Alloc() and let TurboJPEG grow the buffer as needed,
  2. set *jpegBuf to NULL to tell TurboJPEG to allocate the buffer for you, or
  3. pre-allocate the buffer to a "worst case" size determined by calling tj3JPEGBufSize(). This should ensure that the buffer never has to be re-allocated. (Setting TJPARAM_NOREALLOC guarantees that it won't be.)
If you choose option 1, then *jpegSize should be set to the size of your pre-allocated buffer. In any case, unless you have set TJPARAM_NOREALLOC, you should always check *jpegBuf upon return from this function, as it may have changed.
jpegSizepointer to a size_t variable that holds the size of the JPEG buffer. If *jpegBuf points to a pre-allocated buffer, then *jpegSize should be set to the size of the buffer. Upon return, *jpegSize will contain the size of the JPEG image (in bytes.) If *jpegBuf points to a JPEG buffer that is being reused from a previous call to one of the JPEG compression functions, then *jpegSize is ignored.
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3CompressFromYUV8()

DLLEXPORT int tj3CompressFromYUV8 ( tjhandle  handle,
const unsigned char *  srcBuf,
int  width,
int  align,
int  height,
unsigned char **  jpegBuf,
size_t *  jpegSize 
)

Compress an 8-bit-per-sample unified planar YUV image into an 8-bit-per-sample JPEG image.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for compression
srcBufpointer to a buffer containing a unified planar YUV source image to be compressed. The size of this buffer should match the value returned by tj3YUVBufSize() for the given image width, height, row alignment, and level of chrominance subsampling (see TJPARAM_SUBSAMP.) The Y, U (Cb), and V (Cr) image planes should be stored sequentially in the buffer. (Refer to YUV Image Format Notes.)
widthwidth (in pixels) of the source image. If the width is not an even multiple of the MCU block width (see tjMCUWidth), then an intermediate buffer copy will be performed.
alignrow alignment (in bytes) of the source image (must be a power of 2.) Setting this parameter to n indicates that each row in each plane of the source image is padded to the nearest multiple of n bytes (1 = unpadded.)
heightheight (in pixels) of the source image. If the height is not an even multiple of the MCU block height (see tjMCUHeight), then an intermediate buffer copy will be performed.
jpegBufaddress of a pointer to a byte buffer that will receive the JPEG image. TurboJPEG has the ability to reallocate the JPEG buffer to accommodate the size of the JPEG image. Thus, you can choose to:
  1. pre-allocate the JPEG buffer with an arbitrary size using tj3Alloc() and let TurboJPEG grow the buffer as needed,
  2. set *jpegBuf to NULL to tell TurboJPEG to allocate the buffer for you, or
  3. pre-allocate the buffer to a "worst case" size determined by calling tj3JPEGBufSize(). This should ensure that the buffer never has to be re-allocated. (Setting TJPARAM_NOREALLOC guarantees that it won't be.)
If you choose option 1, then *jpegSize should be set to the size of your pre-allocated buffer. In any case, unless you have set TJPARAM_NOREALLOC, you should always check *jpegBuf upon return from this function, as it may have changed.
jpegSizepointer to a size_t variable that holds the size of the JPEG buffer. If *jpegBuf points to a pre-allocated buffer, then *jpegSize should be set to the size of the buffer. Upon return, *jpegSize will contain the size of the JPEG image (in bytes.) If *jpegBuf points to a JPEG buffer that is being reused from a previous call to one of the JPEG compression functions, then *jpegSize is ignored.
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3CompressFromYUVPlanes8()

DLLEXPORT int tj3CompressFromYUVPlanes8 ( tjhandle  handle,
const unsigned char *const *  srcPlanes,
int  width,
const int *  strides,
int  height,
unsigned char **  jpegBuf,
size_t *  jpegSize 
)

Compress a set of 8-bit-per-sample Y, U (Cb), and V (Cr) image planes into an 8-bit-per-sample JPEG image.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for compression
srcPlanesan array of pointers to Y, U (Cb), and V (Cr) image planes (or just a Y plane, if compressing a grayscale image) that contain a YUV source image to be compressed. These planes can be contiguous or non-contiguous in memory. The size of each plane should match the value returned by tj3YUVPlaneSize() for the given image width, height, strides, and level of chrominance subsampling (see TJPARAM_SUBSAMP.) Refer to YUV Image Format Notes for more details.
widthwidth (in pixels) of the source image. If the width is not an even multiple of the MCU block width (see tjMCUWidth), then an intermediate buffer copy will be performed.
stridesan array of integers, each specifying the number of bytes per row in the corresponding plane of the YUV source image. Setting the stride for any plane to 0 is the same as setting it to the plane width (see YUV Image Format Notes.) If strides is NULL, then the strides for all planes will be set to their respective plane widths. You can adjust the strides in order to specify an arbitrary amount of row padding in each plane or to create a JPEG image from a subregion of a larger planar YUV image.
heightheight (in pixels) of the source image. If the height is not an even multiple of the MCU block height (see tjMCUHeight), then an intermediate buffer copy will be performed.
jpegBufaddress of a pointer to a byte buffer that will receive the JPEG image. TurboJPEG has the ability to reallocate the JPEG buffer to accommodate the size of the JPEG image. Thus, you can choose to:
  1. pre-allocate the JPEG buffer with an arbitrary size using tj3Alloc() and let TurboJPEG grow the buffer as needed,
  2. set *jpegBuf to NULL to tell TurboJPEG to allocate the buffer for you, or
  3. pre-allocate the buffer to a "worst case" size determined by calling tj3JPEGBufSize(). This should ensure that the buffer never has to be re-allocated. (Setting TJPARAM_NOREALLOC guarantees that it won't be.)
If you choose option 1, then *jpegSize should be set to the size of your pre-allocated buffer. In any case, unless you have set TJPARAM_NOREALLOC, you should always check *jpegBuf upon return from this function, as it may have changed.
jpegSizepointer to a size_t variable that holds the size of the JPEG buffer. If *jpegBuf points to a pre-allocated buffer, then *jpegSize should be set to the size of the buffer. Upon return, *jpegSize will contain the size of the JPEG image (in bytes.) If *jpegBuf points to a JPEG buffer that is being reused from a previous call to one of the JPEG compression functions, then *jpegSize is ignored.
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3DecodeYUV8()

DLLEXPORT int tj3DecodeYUV8 ( tjhandle  handle,
const unsigned char *  srcBuf,
int  align,
unsigned char *  dstBuf,
int  width,
int  pitch,
int  height,
int  pixelFormat 
)

Decode an 8-bit-per-sample unified planar YUV image into an 8-bit-per-sample packed-pixel RGB or grayscale image.

This function performs color conversion (which is accelerated in the libjpeg-turbo implementation) but does not execute any of the other steps in the JPEG decompression process.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for decompression
srcBufpointer to a buffer containing a unified planar YUV source image to be decoded. The size of this buffer should match the value returned by tj3YUVBufSize() for the given image width, height, row alignment, and level of chrominance subsampling (see TJPARAM_SUBSAMP.) The Y, U (Cb), and V (Cr) image planes should be stored sequentially in the source buffer. (Refer to YUV Image Format Notes.)
alignrow alignment (in bytes) of the YUV source image (must be a power of 2.) Setting this parameter to n indicates that each row in each plane of the YUV source image is padded to the nearest multiple of n bytes (1 = unpadded.)
dstBufpointer to a buffer that will receive the packed-pixel decoded image. This buffer should normally be pitch * height bytes in size. However, you can also use this parameter to decode into a specific region of a larger buffer.
widthwidth (in pixels) of the source and destination images
pitchbytes per row in the destination image. Normally this should be set to width * tjPixelSize[pixelFormat], if the destination image should be unpadded. (Setting this parameter to 0 is the equivalent of setting it to width * tjPixelSize[pixelFormat].) However, you can also use this parameter to specify the row alignment/padding of the destination image, to skip rows, or to decode into a specific region of a larger buffer.
heightheight (in pixels) of the source and destination images
pixelFormatpixel format of the destination image (see Pixel formats.)
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3DecodeYUVPlanes8()

DLLEXPORT int tj3DecodeYUVPlanes8 ( tjhandle  handle,
const unsigned char *const *  srcPlanes,
const int *  strides,
unsigned char *  dstBuf,
int  width,
int  pitch,
int  height,
int  pixelFormat 
)

Decode a set of 8-bit-per-sample Y, U (Cb), and V (Cr) image planes into an 8-bit-per-sample packed-pixel RGB or grayscale image.

This function performs color conversion (which is accelerated in the libjpeg-turbo implementation) but does not execute any of the other steps in the JPEG decompression process.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for decompression
srcPlanesan array of pointers to Y, U (Cb), and V (Cr) image planes (or just a Y plane, if decoding a grayscale image) that contain a YUV image to be decoded. These planes can be contiguous or non-contiguous in memory. The size of each plane should match the value returned by tj3YUVPlaneSize() for the given image width, height, strides, and level of chrominance subsampling (see TJPARAM_SUBSAMP.) Refer to YUV Image Format Notes for more details.
stridesan array of integers, each specifying the number of bytes per row in the corresponding plane of the YUV source image. Setting the stride for any plane to 0 is the same as setting it to the plane width (see YUV Image Format Notes.) If strides is NULL, then the strides for all planes will be set to their respective plane widths. You can adjust the strides in order to specify an arbitrary amount of row padding in each plane or to decode a subregion of a larger planar YUV image.
dstBufpointer to a buffer that will receive the packed-pixel decoded image. This buffer should normally be pitch * height bytes in size. However, you can also use this parameter to decode into a specific region of a larger buffer.
widthwidth (in pixels) of the source and destination images
pitchbytes per row in the destination image. Normally this should be set to width * tjPixelSize[pixelFormat], if the destination image should be unpadded. (Setting this parameter to 0 is the equivalent of setting it to width * tjPixelSize[pixelFormat].) However, you can also use this parameter to specify the row alignment/padding of the destination image, to skip rows, or to decode into a specific region of a larger buffer.
heightheight (in pixels) of the source and destination images
pixelFormatpixel format of the destination image (see Pixel formats.)
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3Decompress12()

DLLEXPORT int tj3Decompress12 ( tjhandle  handle,
const unsigned char *  jpegBuf,
size_t  jpegSize,
short *  dstBuf,
int  pitch,
int  pixelFormat 
)

Decompress a 12-bit-per-sample JPEG image into a 12-bit-per-sample packed-pixel RGB, grayscale, or CMYK image.

The parameters that describe the JPEG image will be set when this function returns.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for decompression
jpegBufpointer to a byte buffer containing the JPEG image to decompress
jpegSizesize of the JPEG image (in bytes)
dstBufpointer to a buffer that will receive the packed-pixel decompressed image. This buffer should normally be pitch * destinationHeight samples in size. However, you can also use this parameter to decompress into a specific region of a larger buffer. NOTE: If the JPEG image is lossy, then destinationHeight is either the scaled JPEG height (see TJSCALED(), TJPARAM_JPEGHEIGHT, and tj3SetScalingFactor()) or the height of the cropping region (see tj3SetCroppingRegion().) If the JPEG image is lossless, then destinationHeight is the JPEG height.
pitchsamples per row in the destination image. Normally this should be set to destinationWidth * tjPixelSize[pixelFormat], if the destination image should be unpadded. (Setting this parameter to 0 is the equivalent of setting it to destinationWidth * tjPixelSize[pixelFormat].) However, you can also use this parameter to specify the row alignment/padding of the destination image, to skip rows, or to decompress into a specific region of a larger buffer. NOTE: If the JPEG image is lossy, then destinationWidth is either the scaled JPEG width (see TJSCALED(), TJPARAM_JPEGWIDTH, and tj3SetScalingFactor()) or the width of the cropping region (see tj3SetCroppingRegion().) If the JPEG image is lossless, then destinationWidth is the JPEG width.
pixelFormatpixel format of the destination image (see Pixel formats.)
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3Decompress16()

DLLEXPORT int tj3Decompress16 ( tjhandle  handle,
const unsigned char *  jpegBuf,
size_t  jpegSize,
unsigned short *  dstBuf,
int  pitch,
int  pixelFormat 
)

Decompress a 16-bit-per-sample lossless JPEG image into a 16-bit-per-sample packed-pixel RGB, grayscale, or CMYK image.

The parameters that describe the JPEG image will be set when this function returns.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for decompression
jpegBufpointer to a byte buffer containing the JPEG image to decompress
jpegSizesize of the JPEG image (in bytes)
dstBufpointer to a buffer that will receive the packed-pixel decompressed image. This buffer should normally be pitch * destinationHeight samples in size. However, you can also use this parameter to decompress into a specific region of a larger buffer. NOTE: If the JPEG image is lossy, then destinationHeight is either the scaled JPEG height (see TJSCALED(), TJPARAM_JPEGHEIGHT, and tj3SetScalingFactor()) or the height of the cropping region (see tj3SetCroppingRegion().) If the JPEG image is lossless, then destinationHeight is the JPEG height.
pitchsamples per row in the destination image. Normally this should be set to destinationWidth * tjPixelSize[pixelFormat], if the destination image should be unpadded. (Setting this parameter to 0 is the equivalent of setting it to destinationWidth * tjPixelSize[pixelFormat].) However, you can also use this parameter to specify the row alignment/padding of the destination image, to skip rows, or to decompress into a specific region of a larger buffer. NOTE: If the JPEG image is lossy, then destinationWidth is either the scaled JPEG width (see TJSCALED(), TJPARAM_JPEGWIDTH, and tj3SetScalingFactor()) or the width of the cropping region (see tj3SetCroppingRegion().) If the JPEG image is lossless, then destinationWidth is the JPEG width.
pixelFormatpixel format of the destination image (see Pixel formats.)
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3Decompress8()

DLLEXPORT int tj3Decompress8 ( tjhandle  handle,
const unsigned char *  jpegBuf,
size_t  jpegSize,
unsigned char *  dstBuf,
int  pitch,
int  pixelFormat 
)

Decompress an 8-bit-per-sample JPEG image into an 8-bit-per-sample packed-pixel RGB, grayscale, or CMYK image.

The parameters that describe the JPEG image will be set when this function returns.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for decompression
jpegBufpointer to a byte buffer containing the JPEG image to decompress
jpegSizesize of the JPEG image (in bytes)
dstBufpointer to a buffer that will receive the packed-pixel decompressed image. This buffer should normally be pitch * destinationHeight samples in size. However, you can also use this parameter to decompress into a specific region of a larger buffer. NOTE: If the JPEG image is lossy, then destinationHeight is either the scaled JPEG height (see TJSCALED(), TJPARAM_JPEGHEIGHT, and tj3SetScalingFactor()) or the height of the cropping region (see tj3SetCroppingRegion().) If the JPEG image is lossless, then destinationHeight is the JPEG height.
pitchsamples per row in the destination image. Normally this should be set to destinationWidth * tjPixelSize[pixelFormat], if the destination image should be unpadded. (Setting this parameter to 0 is the equivalent of setting it to destinationWidth * tjPixelSize[pixelFormat].) However, you can also use this parameter to specify the row alignment/padding of the destination image, to skip rows, or to decompress into a specific region of a larger buffer. NOTE: If the JPEG image is lossy, then destinationWidth is either the scaled JPEG width (see TJSCALED(), TJPARAM_JPEGWIDTH, and tj3SetScalingFactor()) or the width of the cropping region (see tj3SetCroppingRegion().) If the JPEG image is lossless, then destinationWidth is the JPEG width.
pixelFormatpixel format of the destination image (see Pixel formats.)
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3DecompressHeader()

DLLEXPORT int tj3DecompressHeader ( tjhandle  handle,
const unsigned char *  jpegBuf,
size_t  jpegSize 
)

Retrieve information about a JPEG image without decompressing it, or prime the decompressor with quantization and Huffman tables.

If a JPEG image is passed to this function, then the parameters that describe the JPEG image will be set when the function returns.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for decompression
jpegBufpointer to a byte buffer containing a JPEG image or an "abbreviated table specification" (AKA "tables-only") datastream. Passing a tables-only datastream to this function primes the decompressor with quantization and Huffman tables that can be used when decompressing subsequent "abbreviated image" datastreams. This is useful, for instance, when decompressing video streams in which all frames share the same quantization and Huffman tables.
jpegSizesize of the JPEG image or tables-only datastream (in bytes)
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3DecompressToYUV8()

DLLEXPORT int tj3DecompressToYUV8 ( tjhandle  handle,
const unsigned char *  jpegBuf,
size_t  jpegSize,
unsigned char *  dstBuf,
int  align 
)

Decompress an 8-bit-per-sample JPEG image into an 8-bit-per-sample unified planar YUV image.

This function performs JPEG decompression but leaves out the color conversion step, so a planar YUV image is generated instead of a packed-pixel image. The parameters that describe the JPEG image will be set when this function returns.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for decompression
jpegBufpointer to a byte buffer containing the JPEG image to decompress
jpegSizesize of the JPEG image (in bytes)
dstBufpointer to a buffer that will receive the unified planar YUV decompressed image. Use tj3YUVBufSize() to determine the appropriate size for this buffer based on the scaled JPEG width and height (see TJSCALED(), TJPARAM_JPEGWIDTH, TJPARAM_JPEGHEIGHT, and tj3SetScalingFactor()), row alignment, and level of chrominance subsampling (see TJPARAM_SUBSAMP.) The Y, U (Cb), and V (Cr) image planes will be stored sequentially in the buffer. (Refer to YUV Image Format Notes.)
alignrow alignment (in bytes) of the YUV image (must be a power of 2.) Setting this parameter to n will cause each row in each plane of the YUV image to be padded to the nearest multiple of n bytes (1 = unpadded.) To generate images suitable for X Video, align should be set to 4.
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3DecompressToYUVPlanes8()

DLLEXPORT int tj3DecompressToYUVPlanes8 ( tjhandle  handle,
const unsigned char *  jpegBuf,
size_t  jpegSize,
unsigned char **  dstPlanes,
int *  strides 
)

Decompress an 8-bit-per-sample JPEG image into separate 8-bit-per-sample Y, U (Cb), and V (Cr) image planes.

This function performs JPEG decompression but leaves out the color conversion step, so a planar YUV image is generated instead of a packed-pixel image. The parameters that describe the JPEG image will be set when this function returns.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for decompression
jpegBufpointer to a byte buffer containing the JPEG image to decompress
jpegSizesize of the JPEG image (in bytes)
dstPlanesan array of pointers to Y, U (Cb), and V (Cr) image planes (or just a Y plane, if decompressing a grayscale image) that will receive the decompressed image. These planes can be contiguous or non-contiguous in memory. Use tj3YUVPlaneSize() to determine the appropriate size for each plane based on the scaled JPEG width and height (see TJSCALED(), TJPARAM_JPEGWIDTH, TJPARAM_JPEGHEIGHT, and tj3SetScalingFactor()), strides, and level of chrominance subsampling (see TJPARAM_SUBSAMP.) Refer to YUV Image Format Notes for more details.
stridesan array of integers, each specifying the number of bytes per row in the corresponding plane of the YUV image. Setting the stride for any plane to 0 is the same as setting it to the scaled plane width (see YUV Image Format Notes.) If strides is NULL, then the strides for all planes will be set to their respective scaled plane widths. You can adjust the strides in order to add an arbitrary amount of row padding to each plane or to decompress the JPEG image into a subregion of a larger planar YUV image.
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3Destroy()

DLLEXPORT void tj3Destroy ( tjhandle  handle)

Destroy a TurboJPEG instance.

Parameters
handlehandle to a TurboJPEG instance. If the handle is NULL, then this function has no effect.

◆ tj3EncodeYUV8()

DLLEXPORT int tj3EncodeYUV8 ( tjhandle  handle,
const unsigned char *  srcBuf,
int  width,
int  pitch,
int  height,
int  pixelFormat,
unsigned char *  dstBuf,
int  align 
)

Encode an 8-bit-per-sample packed-pixel RGB or grayscale image into an 8-bit-per-sample unified planar YUV image.

This function performs color conversion (which is accelerated in the libjpeg-turbo implementation) but does not execute any of the other steps in the JPEG compression process.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for compression
srcBufpointer to a buffer containing a packed-pixel RGB or grayscale source image to be encoded. This buffer should normally be pitch * height bytes in size. However, you can also use this parameter to encode from a specific region of a larger buffer.
widthwidth (in pixels) of the source image
pitchbytes per row in the source image. Normally this should be width * tjPixelSize[pixelFormat], if the image is unpadded. (Setting this parameter to 0 is the equivalent of setting it to width * tjPixelSize[pixelFormat].) However, you can also use this parameter to specify the row alignment/padding of the source image, to skip rows, or to encode from a specific region of a larger packed-pixel image.
heightheight (in pixels) of the source image
pixelFormatpixel format of the source image (see Pixel formats.)
dstBufpointer to a buffer that will receive the unified planar YUV image. Use tj3YUVBufSize() to determine the appropriate size for this buffer based on the image width, height, row alignment, and level of chrominance subsampling (see TJPARAM_SUBSAMP.) The Y, U (Cb), and V (Cr) image planes will be stored sequentially in the buffer. (Refer to YUV Image Format Notes.)
alignrow alignment (in bytes) of the YUV image (must be a power of 2.) Setting this parameter to n will cause each row in each plane of the YUV image to be padded to the nearest multiple of n bytes (1 = unpadded.) To generate images suitable for X Video, align should be set to 4.
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3EncodeYUVPlanes8()

DLLEXPORT int tj3EncodeYUVPlanes8 ( tjhandle  handle,
const unsigned char *  srcBuf,
int  width,
int  pitch,
int  height,
int  pixelFormat,
unsigned char **  dstPlanes,
int *  strides 
)

Encode an 8-bit-per-sample packed-pixel RGB or grayscale image into separate 8-bit-per-sample Y, U (Cb), and V (Cr) image planes.

This function performs color conversion (which is accelerated in the libjpeg-turbo implementation) but does not execute any of the other steps in the JPEG compression process.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for compression
srcBufpointer to a buffer containing a packed-pixel RGB or grayscale source image to be encoded. This buffer should normally be pitch * height bytes in size. However, you can also use this parameter to encode from a specific region of a larger buffer.
widthwidth (in pixels) of the source image
pitchbytes per row in the source image. Normally this should be width * tjPixelSize[pixelFormat], if the image is unpadded. (Setting this parameter to 0 is the equivalent of setting it to width * tjPixelSize[pixelFormat].) However, you can also use this parameter to specify the row alignment/padding of the source image, to skip rows, or to encode from a specific region of a larger packed-pixel image.
heightheight (in pixels) of the source image
pixelFormatpixel format of the source image (see Pixel formats.)
dstPlanesan array of pointers to Y, U (Cb), and V (Cr) image planes (or just a Y plane, if generating a grayscale image) that will receive the encoded image. These planes can be contiguous or non-contiguous in memory. Use tj3YUVPlaneSize() to determine the appropriate size for each plane based on the image width, height, strides, and level of chrominance subsampling (see TJPARAM_SUBSAMP.) Refer to YUV Image Format Notes for more details.
stridesan array of integers, each specifying the number of bytes per row in the corresponding plane of the YUV image. Setting the stride for any plane to 0 is the same as setting it to the plane width (see YUV Image Format Notes.) If strides is NULL, then the strides for all planes will be set to their respective plane widths. You can adjust the strides in order to add an arbitrary amount of row padding to each plane or to encode an RGB or grayscale image into a subregion of a larger planar YUV image.
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3Free()

DLLEXPORT void tj3Free ( void *  buffer)

Free a byte buffer previously allocated by TurboJPEG.

You should always use this function to free JPEG destination buffer(s) that were automatically (re)allocated by the compression and transform functions or that were manually allocated using tj3Alloc().

Parameters
bufferaddress of the buffer to free. If the address is NULL, then this function has no effect.
See also
tj3Alloc()

◆ tj3Get()

DLLEXPORT int tj3Get ( tjhandle  handle,
int  param 
)

Get the value of a parameter.

Parameters
handlehandle to a TurboJPEG instance
paramone of the parameters
Returns
the value of the specified parameter, or -1 if the value is unknown.

◆ tj3GetErrorCode()

DLLEXPORT int tj3GetErrorCode ( tjhandle  handle)

Returns a code indicating the severity of the last error.

See Error codes.

Parameters
handlehandle to a TurboJPEG instance
Returns
a code indicating the severity of the last error. See Error codes.

◆ tj3GetErrorStr()

DLLEXPORT char * tj3GetErrorStr ( tjhandle  handle)

Returns a descriptive error message explaining why the last command failed.

Parameters
handlehandle to a TurboJPEG instance, or NULL if the error was generated by a global function (but note that retrieving the error message for a global function is thread-safe only on platforms that support thread-local storage.)
Returns
a descriptive error message explaining why the last command failed.

◆ tj3GetScalingFactors()

DLLEXPORT tjscalingfactor * tj3GetScalingFactors ( int *  numScalingFactors)

Returns a list of fractional scaling factors that the JPEG decompressor supports.

Parameters
numScalingFactorspointer to an integer variable that will receive the number of elements in the list
Returns
a pointer to a list of fractional scaling factors, or NULL if an error is encountered (see tj3GetErrorStr().)

◆ tj3Init()

DLLEXPORT tjhandle tj3Init ( int  initType)

Create a new TurboJPEG instance.

Parameters
initTypeone of the initialization options
Returns
a handle to the newly-created instance, or NULL if an error occurred (see tj3GetErrorStr().)

◆ tj3JPEGBufSize()

DLLEXPORT size_t tj3JPEGBufSize ( int  width,
int  height,
int  jpegSubsamp 
)

The maximum size of the buffer (in bytes) required to hold a JPEG image with the given parameters.

The number of bytes returned by this function is larger than the size of the uncompressed source image. The reason for this is that the JPEG format uses 16-bit coefficients, so it is possible for a very high-quality source image with very high-frequency content to expand rather than compress when converted to the JPEG format. Such images represent very rare corner cases, but since there is no way to predict the size of a JPEG image prior to compression, the corner cases have to be handled.

Parameters
widthwidth (in pixels) of the image
heightheight (in pixels) of the image
jpegSubsampthe level of chrominance subsampling to be used when generating the JPEG image (see Chrominance subsampling options.) TJSAMP_UNKNOWN is treated like TJSAMP_444, since a buffer large enough to hold a JPEG image with no subsampling should also be large enough to hold a JPEG image with an arbitrary level of subsampling. Note that lossless JPEG images always use TJSAMP_444.
Returns
the maximum size of the buffer (in bytes) required to hold the image, or 0 if the arguments are out of bounds.

◆ tj3LoadImage12()

DLLEXPORT short * tj3LoadImage12 ( tjhandle  handle,
const char *  filename,
int *  width,
int  align,
int *  height,
int *  pixelFormat 
)

Load a 12-bit-per-sample packed-pixel image from disk into memory.

Parameters
handlehandle to a TurboJPEG instance
filenamename of a file containing a packed-pixel image in Windows BMP or PBMPLUS (PPM/PGM) format. Windows BMP files require 8-bit-per-sample data precision. If the data precision of the PBMPLUS file does not match the target data precision, then upconverting or downconverting will be performed.
widthpointer to an integer variable that will receive the width (in pixels) of the packed-pixel image
alignrow alignment (in samples) of the packed-pixel buffer to be returned (must be a power of 2.) Setting this parameter to n will cause all rows in the buffer to be padded to the nearest multiple of n samples (1 = unpadded.)
heightpointer to an integer variable that will receive the height (in pixels) of the packed-pixel image
pixelFormatpointer to an integer variable that specifies or will receive the pixel format of the packed-pixel buffer. The behavior of this function will vary depending on the value of *pixelFormat passed to the function:
  • TJPF_UNKNOWN : The packed-pixel buffer returned by this function will use the most optimal pixel format for the file type, and *pixelFormat will contain the ID of that pixel format upon successful return from this function.
  • TJPF_GRAY : Only PGM files and 8-bit-per-pixel BMP files with a grayscale colormap can be loaded.
  • TJPF_CMYK : The RGB or grayscale pixels stored in the file will be converted using a quick & dirty algorithm that is suitable only for testing purposes. (Proper conversion between CMYK and other formats requires a color management system.)
  • Other pixel formats : The packed-pixel buffer will use the specified pixel format, and pixel format conversion will be performed if necessary.
Returns
a pointer to a newly-allocated buffer containing the packed-pixel image, converted to the chosen pixel format and with the chosen row alignment, or NULL if an error occurred (see tj3GetErrorStr().) This buffer should be freed using tj3Free().

◆ tj3LoadImage16()

DLLEXPORT unsigned short * tj3LoadImage16 ( tjhandle  handle,
const char *  filename,
int *  width,
int  align,
int *  height,
int *  pixelFormat 
)

Load a 16-bit-per-sample packed-pixel image from disk into memory.

Parameters
handlehandle to a TurboJPEG instance
filenamename of a file containing a packed-pixel image in Windows BMP or PBMPLUS (PPM/PGM) format. Windows BMP files require 8-bit-per-sample data precision. If the data precision of the PBMPLUS file does not match the target data precision, then upconverting or downconverting will be performed.
widthpointer to an integer variable that will receive the width (in pixels) of the packed-pixel image
alignrow alignment (in samples) of the packed-pixel buffer to be returned (must be a power of 2.) Setting this parameter to n will cause all rows in the buffer to be padded to the nearest multiple of n samples (1 = unpadded.)
heightpointer to an integer variable that will receive the height (in pixels) of the packed-pixel image
pixelFormatpointer to an integer variable that specifies or will receive the pixel format of the packed-pixel buffer. The behavior of this function will vary depending on the value of *pixelFormat passed to the function:
  • TJPF_UNKNOWN : The packed-pixel buffer returned by this function will use the most optimal pixel format for the file type, and *pixelFormat will contain the ID of that pixel format upon successful return from this function.
  • TJPF_GRAY : Only PGM files and 8-bit-per-pixel BMP files with a grayscale colormap can be loaded.
  • TJPF_CMYK : The RGB or grayscale pixels stored in the file will be converted using a quick & dirty algorithm that is suitable only for testing purposes. (Proper conversion between CMYK and other formats requires a color management system.)
  • Other pixel formats : The packed-pixel buffer will use the specified pixel format, and pixel format conversion will be performed if necessary.
Returns
a pointer to a newly-allocated buffer containing the packed-pixel image, converted to the chosen pixel format and with the chosen row alignment, or NULL if an error occurred (see tj3GetErrorStr().) This buffer should be freed using tj3Free().

◆ tj3LoadImage8()

DLLEXPORT unsigned char * tj3LoadImage8 ( tjhandle  handle,
const char *  filename,
int *  width,
int  align,
int *  height,
int *  pixelFormat 
)

Load an 8-bit-per-sample packed-pixel image from disk into memory.

Parameters
handlehandle to a TurboJPEG instance
filenamename of a file containing a packed-pixel image in Windows BMP or PBMPLUS (PPM/PGM) format. Windows BMP files require 8-bit-per-sample data precision. If the data precision of the PBMPLUS file does not match the target data precision, then upconverting or downconverting will be performed.
widthpointer to an integer variable that will receive the width (in pixels) of the packed-pixel image
alignrow alignment (in samples) of the packed-pixel buffer to be returned (must be a power of 2.) Setting this parameter to n will cause all rows in the buffer to be padded to the nearest multiple of n samples (1 = unpadded.)
heightpointer to an integer variable that will receive the height (in pixels) of the packed-pixel image
pixelFormatpointer to an integer variable that specifies or will receive the pixel format of the packed-pixel buffer. The behavior of this function will vary depending on the value of *pixelFormat passed to the function:
  • TJPF_UNKNOWN : The packed-pixel buffer returned by this function will use the most optimal pixel format for the file type, and *pixelFormat will contain the ID of that pixel format upon successful return from this function.
  • TJPF_GRAY : Only PGM files and 8-bit-per-pixel BMP files with a grayscale colormap can be loaded.
  • TJPF_CMYK : The RGB or grayscale pixels stored in the file will be converted using a quick & dirty algorithm that is suitable only for testing purposes. (Proper conversion between CMYK and other formats requires a color management system.)
  • Other pixel formats : The packed-pixel buffer will use the specified pixel format, and pixel format conversion will be performed if necessary.
Returns
a pointer to a newly-allocated buffer containing the packed-pixel image, converted to the chosen pixel format and with the chosen row alignment, or NULL if an error occurred (see tj3GetErrorStr().) This buffer should be freed using tj3Free().

◆ tj3SaveImage12()

DLLEXPORT int tj3SaveImage12 ( tjhandle  handle,
const char *  filename,
const short *  buffer,
int  width,
int  pitch,
int  height,
int  pixelFormat 
)

Save a 12-bit-per-sample packed-pixel image from memory to disk.

Parameters
handlehandle to a TurboJPEG instance
filenamename of a file to which to save the packed-pixel image. The image will be stored in Windows BMP or PBMPLUS (PPM/PGM) format, depending on the file extension. Windows BMP files require 8-bit-per-sample data precision.
bufferpointer to a buffer containing a packed-pixel RGB, grayscale, or CMYK image to be saved
widthwidth (in pixels) of the packed-pixel image
pitchsamples per row in the packed-pixel image. Setting this parameter to 0 is the equivalent of setting it to width * tjPixelSize[pixelFormat].
heightheight (in pixels) of the packed-pixel image
pixelFormatpixel format of the packed-pixel image (see Pixel formats.) If this parameter is set to TJPF_GRAY, then the image will be stored in PGM or 8-bit-per-pixel (indexed color) BMP format. Otherwise, the image will be stored in PPM or 24-bit-per-pixel BMP format. If this parameter is set to TJPF_CMYK, then the CMYK pixels will be converted to RGB using a quick & dirty algorithm that is suitable only for testing purposes. (Proper conversion between CMYK and other formats requires a color management system.)
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr().)

◆ tj3SaveImage16()

DLLEXPORT int tj3SaveImage16 ( tjhandle  handle,
const char *  filename,
const unsigned short *  buffer,
int  width,
int  pitch,
int  height,
int  pixelFormat 
)

Save a 16-bit-per-sample packed-pixel image from memory to disk.

Parameters
handlehandle to a TurboJPEG instance
filenamename of a file to which to save the packed-pixel image. The image will be stored in Windows BMP or PBMPLUS (PPM/PGM) format, depending on the file extension. Windows BMP files require 8-bit-per-sample data precision.
bufferpointer to a buffer containing a packed-pixel RGB, grayscale, or CMYK image to be saved
widthwidth (in pixels) of the packed-pixel image
pitchsamples per row in the packed-pixel image. Setting this parameter to 0 is the equivalent of setting it to width * tjPixelSize[pixelFormat].
heightheight (in pixels) of the packed-pixel image
pixelFormatpixel format of the packed-pixel image (see Pixel formats.) If this parameter is set to TJPF_GRAY, then the image will be stored in PGM or 8-bit-per-pixel (indexed color) BMP format. Otherwise, the image will be stored in PPM or 24-bit-per-pixel BMP format. If this parameter is set to TJPF_CMYK, then the CMYK pixels will be converted to RGB using a quick & dirty algorithm that is suitable only for testing purposes. (Proper conversion between CMYK and other formats requires a color management system.)
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr().)

◆ tj3SaveImage8()

DLLEXPORT int tj3SaveImage8 ( tjhandle  handle,
const char *  filename,
const unsigned char *  buffer,
int  width,
int  pitch,
int  height,
int  pixelFormat 
)

Save an 8-bit-per-sample packed-pixel image from memory to disk.

Parameters
handlehandle to a TurboJPEG instance
filenamename of a file to which to save the packed-pixel image. The image will be stored in Windows BMP or PBMPLUS (PPM/PGM) format, depending on the file extension. Windows BMP files require 8-bit-per-sample data precision.
bufferpointer to a buffer containing a packed-pixel RGB, grayscale, or CMYK image to be saved
widthwidth (in pixels) of the packed-pixel image
pitchsamples per row in the packed-pixel image. Setting this parameter to 0 is the equivalent of setting it to width * tjPixelSize[pixelFormat].
heightheight (in pixels) of the packed-pixel image
pixelFormatpixel format of the packed-pixel image (see Pixel formats.) If this parameter is set to TJPF_GRAY, then the image will be stored in PGM or 8-bit-per-pixel (indexed color) BMP format. Otherwise, the image will be stored in PPM or 24-bit-per-pixel BMP format. If this parameter is set to TJPF_CMYK, then the CMYK pixels will be converted to RGB using a quick & dirty algorithm that is suitable only for testing purposes. (Proper conversion between CMYK and other formats requires a color management system.)
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr().)

◆ tj3Set()

DLLEXPORT int tj3Set ( tjhandle  handle,
int  param,
int  value 
)

Set the value of a parameter.

Parameters
handlehandle to a TurboJPEG instance
paramone of the parameters
valuevalue of the parameter (refer to parameter documentation)
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr().)

◆ tj3SetCroppingRegion()

DLLEXPORT int tj3SetCroppingRegion ( tjhandle  handle,
tjregion  croppingRegion 
)

Set the cropping region for partially decompressing a lossy JPEG image into a packed-pixel image.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for decompression
croppingRegiontjregion structure that specifies a subregion of the JPEG image to decompress, or TJUNCROPPED for no cropping. The left boundary of the cropping region must be evenly divisible by the scaled MCU block width (TJSCALED(tjMCUWidth[subsamp], scalingFactor), where subsamp is the level of chrominance subsampling in the JPEG image (see TJPARAM_SUBSAMP) and scalingFactor is the decompression scaling factor (see tj3SetScalingFactor().) The cropping region should be specified relative to the scaled image dimensions. Unless croppingRegion is TJUNCROPPED, the JPEG header must be read (see tj3DecompressHeader()) prior to calling this function.
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr().)

◆ tj3SetScalingFactor()

DLLEXPORT int tj3SetScalingFactor ( tjhandle  handle,
tjscalingfactor  scalingFactor 
)

Set the scaling factor for subsequent lossy decompression operations.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for decompression
scalingFactortjscalingfactor structure that specifies a fractional scaling factor that the decompressor supports (see tj3GetScalingFactors()), or TJUNSCALED for no scaling. Decompression scaling is a function of the IDCT algorithm, so scaling factors are generally limited to multiples of 1/8. If the entire JPEG image will be decompressed, then the width and height of the scaled destination image can be determined by calling TJSCALED() with the JPEG width and height (see TJPARAM_JPEGWIDTH and TJPARAM_JPEGHEIGHT) and the specified scaling factor. When decompressing into a planar YUV image, an intermediate buffer copy will be performed if the width or height of the scaled destination image is not an even multiple of the MCU block size (see tjMCUWidth and tjMCUHeight.) Note that decompression scaling is not available (and the specified scaling factor is ignored) when decompressing lossless JPEG images (see TJPARAM_LOSSLESS), since the IDCT algorithm is not used with those images. Note also that TJPARAM_FASTDCT is ignored when decompression scaling is enabled.
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr().)

◆ tj3Transform()

DLLEXPORT int tj3Transform ( tjhandle  handle,
const unsigned char *  jpegBuf,
size_t  jpegSize,
int  n,
unsigned char **  dstBufs,
size_t *  dstSizes,
const tjtransform transforms 
)

Losslessly transform a JPEG image into another JPEG image.

Lossless transforms work by moving the raw DCT coefficients from one JPEG image structure to another without altering the values of the coefficients. While this is typically faster than decompressing the image, transforming it, and re-compressing it, lossless transforms are not free. Each lossless transform requires reading and performing entropy decoding on all of the coefficients in the source image, regardless of the size of the destination image. Thus, this function provides a means of generating multiple transformed images from the same source or applying multiple transformations simultaneously, in order to eliminate the need to read the source coefficients multiple times.

Parameters
handlehandle to a TurboJPEG instance that has been initialized for lossless transformation
jpegBufpointer to a byte buffer containing the JPEG source image to transform
jpegSizesize of the JPEG source image (in bytes)
nthe number of transformed JPEG images to generate
dstBufspointer to an array of n byte buffers. dstBufs[i] will receive a JPEG image that has been transformed using the parameters in transforms[i]. TurboJPEG has the ability to reallocate the JPEG destination buffer to accommodate the size of the transformed JPEG image. Thus, you can choose to:
  1. pre-allocate the JPEG destination buffer with an arbitrary size using tj3Alloc() and let TurboJPEG grow the buffer as needed,
  2. set dstBufs[i] to NULL to tell TurboJPEG to allocate the buffer for you, or
  3. pre-allocate the buffer to a "worst case" size determined by calling tj3JPEGBufSize() with the transformed or cropped width and height and the level of subsampling used in the source image. Under normal circumstances, this should ensure that the buffer never has to be re-allocated. (Setting TJPARAM_NOREALLOC guarantees that it won't be.) Note, however, that there are some rare cases (such as transforming images with a large amount of embedded EXIF or ICC profile data) in which the transformed JPEG image will be larger than the worst-case size, and TJPARAM_NOREALLOC cannot be used in those cases.
If you choose option 1, then dstSizes[i] should be set to the size of your pre-allocated buffer. In any case, unless you have set TJPARAM_NOREALLOC, you should always check dstBufs[i] upon return from this function, as it may have changed.
dstSizespointer to an array of n size_t variables that will receive the actual sizes (in bytes) of each transformed JPEG image. If dstBufs[i] points to a pre-allocated buffer, then dstSizes[i] should be set to the size of the buffer. Upon return, dstSizes[i] will contain the size of the transformed JPEG image (in bytes.)
transformspointer to an array of n tjtransform structures, each of which specifies the transform parameters and/or cropping region for the corresponding transformed JPEG image.
Returns
0 if successful, or -1 if an error occurred (see tj3GetErrorStr() and tj3GetErrorCode().)

◆ tj3YUVBufSize()

DLLEXPORT size_t tj3YUVBufSize ( int  width,
int  align,
int  height,
int  subsamp 
)

The size of the buffer (in bytes) required to hold a unified planar YUV image with the given parameters.

Parameters
widthwidth (in pixels) of the image
alignrow alignment (in bytes) of the image (must be a power of 2.) Setting this parameter to n specifies that each row in each plane of the image will be padded to the nearest multiple of n bytes (1 = unpadded.)
heightheight (in pixels) of the image
subsamplevel of chrominance subsampling in the image (see Chrominance subsampling options.)
Returns
the size of the buffer (in bytes) required to hold the image, or 0 if the arguments are out of bounds.

◆ tj3YUVPlaneHeight()

DLLEXPORT int tj3YUVPlaneHeight ( int  componentID,
int  height,
int  subsamp 
)

The plane height of a YUV image plane with the given parameters.

Refer to YUV Image Format Notes for a description of plane height.

Parameters
componentIDID number of the image plane (0 = Y, 1 = U/Cb, 2 = V/Cr)
heightheight (in pixels) of the YUV image
subsamplevel of chrominance subsampling in the image (see Chrominance subsampling options.)
Returns
the plane height of a YUV image plane with the given parameters, or 0 if the arguments are out of bounds.

◆ tj3YUVPlaneSize()

DLLEXPORT size_t tj3YUVPlaneSize ( int  componentID,
int  width,
int  stride,
int  height,
int  subsamp 
)

The size of the buffer (in bytes) required to hold a YUV image plane with the given parameters.

Parameters
componentIDID number of the image plane (0 = Y, 1 = U/Cb, 2 = V/Cr)
widthwidth (in pixels) of the YUV image. NOTE: this is the width of the whole image, not the plane width.
stridebytes per row in the image plane. Setting this to 0 is the equivalent of setting it to the plane width.
heightheight (in pixels) of the YUV image. NOTE: this is the height of the whole image, not the plane height.
subsamplevel of chrominance subsampling in the image (see Chrominance subsampling options.)
Returns
the size of the buffer (in bytes) required to hold the YUV image plane, or 0 if the arguments are out of bounds.

◆ tj3YUVPlaneWidth()

DLLEXPORT int tj3YUVPlaneWidth ( int  componentID,
int  width,
int  subsamp 
)

The plane width of a YUV image plane with the given parameters.

Refer to YUV Image Format Notes for a description of plane width.

Parameters
componentIDID number of the image plane (0 = Y, 1 = U/Cb, 2 = V/Cr)
widthwidth (in pixels) of the YUV image
subsamplevel of chrominance subsampling in the image (see Chrominance subsampling options.)
Returns
the plane width of a YUV image plane with the given parameters, or 0 if the arguments are out of bounds.

Variable Documentation

◆ tjAlphaOffset

const int tjAlphaOffset[TJ_NUMPF]
static

Alpha offset (in samples) for a given pixel format.

This specifies the number of samples that the alpha component is offset from the start of the pixel. For instance, if an 8-bit-per-component pixel of format TJPF_BGRA is stored in unsigned char pixel[], then the alpha component will be pixel[tjAlphaOffset[TJPF_BGRA]]. This will be -1 if the pixel format does not have an alpha component.

◆ tjBlueOffset

const int tjBlueOffset[TJ_NUMPF]
static

Blue offset (in samples) for a given pixel format.

This specifies the number of samples that the blue component is offset from the start of the pixel. For instance, if an 8-bit-per-component pixel of format TJPF_BGRX is stored in unsigned char pixel[], then the blue component will be pixel[tjBlueOffset[TJPF_BGRX]]. This will be -1 if the pixel format does not have a blue component.

◆ tjGreenOffset

const int tjGreenOffset[TJ_NUMPF]
static

Green offset (in samples) for a given pixel format.

This specifies the number of samples that the green component is offset from the start of the pixel. For instance, if an 8-bit-per-component pixel of format TJPF_BGRX is stored in unsigned char pixel[], then the green component will be pixel[tjGreenOffset[TJPF_BGRX]]. This will be -1 if the pixel format does not have a green component.

◆ tjMCUHeight

const int tjMCUHeight[TJ_NUMSAMP]
static

MCU block height (in pixels) for a given level of chrominance subsampling.

MCU block sizes:

  • 8x8 for no subsampling or grayscale
  • 16x8 for 4:2:2
  • 8x16 for 4:4:0
  • 16x16 for 4:2:0
  • 32x8 for 4:1:1
  • 8x32 for 4:4:1

◆ tjMCUWidth

const int tjMCUWidth[TJ_NUMSAMP]
static

MCU block width (in pixels) for a given level of chrominance subsampling.

MCU block sizes:

  • 8x8 for no subsampling or grayscale
  • 16x8 for 4:2:2
  • 8x16 for 4:4:0
  • 16x16 for 4:2:0
  • 32x8 for 4:1:1
  • 8x32 for 4:4:1

◆ tjPixelSize

const int tjPixelSize[TJ_NUMPF]
static

Pixel size (in samples) for a given pixel format.

◆ tjRedOffset

const int tjRedOffset[TJ_NUMPF]
static

Red offset (in samples) for a given pixel format.

This specifies the number of samples that the red component is offset from the start of the pixel. For instance, if an 8-bit-per-component pixel of format TJPF_BGRX is stored in unsigned char pixel[], then the red component will be pixel[tjRedOffset[TJPF_BGRX]]. This will be -1 if the pixel format does not have a red component.

◆ TJUNCROPPED

const tjregion TJUNCROPPED
static

A tjregion structure that specifies no cropping.

◆ TJUNSCALED

const tjscalingfactor TJUNSCALED
static

A tjscalingfactor structure that specifies a scaling factor of 1/1 (no scaling)