IRIX Base Documentation 2002 November/SGI IRIX Base Documentation 2002 November.iso/usr/share/catman/p_man/cat3dm/old-compression/cl_mpeg1.z/cl_mpeg1

Text File | 2002-10-03 | 18.5 KB | 397 lines

ccccllll____mmmmppppeeeegggg1111((((3333ddddmmmm))))                                                    ccccllll____mmmmppppeeeegggg1111((((3333ddddmmmm))))



NNNNAAAAMMMMEEEE
     cl_mpeg1 - MPEG-1 schemes in the Compression Library

SSSSYYYYNNNNOOOOPPPPSSSSIIIISSSS
     ####iiiinnnncccclllluuuuddddeeee <<<<ddddmmmmeeeeddddiiiiaaaa////ccccllll....hhhh>>>>

DDDDEEEESSSSCCCCRRRRIIIIPPPPTTTTIIIIOOOONNNN
     The MPEG compression standard, mpeg(4), is supported in the Compression
     Library, CLintro(3dm), with two schemes

     1. CCCCLLLL____MMMMPPPPEEEEGGGG1111____VVVVIIIIDDDDEEEEOOOO____SSSSOOOOFFFFTTTTWWWWAAAARRRREEEE     video compressor/decompressor
     2. CCCCLLLL____MMMMPPPPEEEEGGGG1111____AAAAUUUUDDDDIIIIOOOO____SSSSOOOOFFFFTTTTWWWWAAAARRRREEEE     audio compressor/decompressor

     Each scheme is accessed using the basic CL API and parameter set.  This
     man page describes the features supported by these schemes, as well as
     the general CL and MPEG-specific parameters that are relevant for their
     operation.

     The Compression Library also provides MPEG audio support through a third
     scheme, CL_MPEG1_AUDIO_AWARE, which uses somewhat different parameters.
     For a discussion of this and other compression schemes from Aware, Inc.,
     see cl_aware(3dm).


FFFFEEEEAAAATTTTUUUURRRREEEESSSS
     VVVViiiiddddeeeeoooo CCCCoooommmmpppprrrreeeessssssssiiiioooonnnn
          o Cropping and smoothing of input frames
          o High-quality rate control
          o Automatic I-frame insertion at scene changes


     AAAAuuuuddddiiiioooo CCCCoooommmmpppprrrreeeessssssssiiiioooonnnn
          o Layer I and II support
          o Stereo, joint, dual, and single-channel modes



PPPPRRRROOOOGGGGRRRRAAAAMMMMMMMMIIIINNNNGGGG MMMMPPPPEEEEGGGG
     The two MPEG schemes, CL_MPEG1_VIDEO_SOFTWARE and
     CL_MPEG1_AUDIO_SOFTWARE, are accessed through the standard CL API.  See
     CLintro(3dm) for an overview of the CL calls and their usage.


     Only _I_m_p_l_i_c_i_t _B_u_f_f_e_r_i_n_g is currently supported for input buffers of video
     and audio decompressors, and for both buffers of video compressors.  This
     model requires that the implicit buffer be created with ccccllllCCCCrrrreeeeaaaatttteeeeBBBBuuuuffff((((3333ddddmmmm))))
     prior to calling ccccllllDDDDeeeeccccoooommmmpppprrrreeeessssssss((((3333ddddmmmm)))) or ccccllllCCCCoooommmmpppprrrreeeessssssss((((3333ddddmmmm)))).  Once created, the
     same implicit buffer should be used for the lifetime of the processing
     object.






                                                                        PPPPaaaaggggeeee 1111






ccccllll____mmmmppppeeeegggg1111((((3333ddddmmmm))))                                                    ccccllll____mmmmppppeeeegggg1111((((3333ddddmmmm))))



   PPPPAAAARRRRAAAAMMMMEEEETTTTEEEERRRRSSSS
     Various parameters are useful for guiding and monitoring MPEG processing.
     These are discussed in several groups : General CL, MPEG-1, MPEG-1 Video,
     and MPEG-1 Audio.


   PPPPAAAARRRRAAAAMMMMEEEETTTTEEEERRRRSSSS :::: GGGGEEEENNNNEEEERRRRAAAALLLL CCCCLLLL
     All of the general CL parameters are described in CLintro(3dm).  Below we
     mention only those for which additional information is useful (such as
     parameter ranges and other constraints).

     CCCCLLLL____IIIIMMMMAAAAGGGGEEEE____WWWWIIIIDDDDTTTTHHHH,,,, CCCCLLLL____IIIIMMMMAAAAGGGGEEEE____HHHHEEEEIIIIGGGGHHHHTTTT
          For video compressors, these should be set to match the input
          frame size.  Each dimension has a minimum value of 16 and a
          maximum value of 4080.  For video decompressors, these are
          read-only and match the encoded ("internal") dimensions.

     CCCCLLLL____IIIINNNNTTTTEEEERRRRNNNNAAAALLLL____IIIIMMMMAAAAGGGGEEEE____WWWWIIIIDDDDTTTTHHHH,,,, CCCCLLLL____IIIINNNNTTTTEEEERRRRNNNNAAAALLLL____IIIIMMMMAAAAGGGGEEEE____HHHHEEEEIIIIGGGGHHHHTTTT
          For video compressors, these default to the "source input format"
          (SIF), which is 352x240 (CCIR601-525 pixel aspect ratio), 352x288
          (CCIR601-625 pixel aspect ratio), or 320x240 (square pixels).  For
          inputs with other pixel aspect ratios, the closest of these three
          applies.

          If cropping is used (see below), the cropped image is scaled to
          the internal dimensions prior to encoding.  The internal width and
          height must each be between 16 and 4080, inclusive, and be a
          multiple of 16.

          For video decompressors, these parameters are read-only.

     CCCCLLLL____IIIIMMMMAAAAGGGGEEEE____CCCCRRRROOOOPPPP____LLLLEEEEFFFFTTTT,,,, CCCCLLLL____IIIIMMMMAAAAGGGGEEEE____CCCCRRRROOOOPPPP____RRRRIIIIGGGGHHHHTTTT,,,,
     CCCCLLLL____IIIIMMMMAAAAGGGGEEEE____CCCCRRRROOOOPPPP____TTTTOOOOPPPP,,,, CCCCLLLL____IIIIMMMMAAAAGGGGEEEE____CCCCRRRROOOOPPPP____BBBBOOOOTTTTTTTTOOOOMMMM
          For video compressors, these indicate how many pixels to crop
          from each of the four sides of the input image.  These may be
          set to any value from 0 to 4064.  On the first call to clCompress,
          the cropped region is checked to make sure it is at least 16x16
          pixels in size.

     CCCCLLLL____IIIINNNNTTTTEEEERRRRNNNNAAAALLLL____FFFFOOOORRRRMMMMAAAATTTT
          For video, this is always CL_FORMAT_YCbCr422DC.  To be MPEG-
          compliant, the bitstream is actually encoded in 4:2:0 format,
          where the chroma is subsampled by 2 vertically in addition to
          horizontally, but the current implementation exposes only the
          duplicate chroma format.

     CCCCLLLL____BBBBIIIITTTTSSSS____PPPPEEEERRRR____CCCCOOOOMMMMPPPPOOOONNNNEEEENNNNTTTT
          For audio, only 16-bit samples are supported.

     CCCCLLLL____FFFFRRRRAAAAMMMMEEEE____RRRRAAAATTTTEEEE
          For compressors, this should be set to indicate the frame rate of
          the input.  The following rates are allowed.



                                                                        PPPPaaaaggggeeee 2222






ccccllll____mmmmppppeeeegggg1111((((3333ddddmmmm))))                                                    ccccllll____mmmmppppeeeegggg1111((((3333ddddmmmm))))



              Video : 23.976, 24, 25, 29.97, 30, 50, 59.94, 60
              Audio : 32000, 44100, 48000

          For decompressors, this is read-only.

     CCCCLLLL____PPPPIIIIXXXXEEEELLLL____AAAASSSSPPPPEEEECCCCTTTT
          Pixel aspect ratio, defined as the vertical extent of a pixel
          divided by its horizontal extent.  For video compressors, this
          should be set to match the aspect ratio of the input pixels.
          The closest value from the following list is used.

                1.0000   Square
                0.6735
                0.7031   16:9, 625line
                0.7615
                0.8055
                0.8437   16:9, 525line
                0.8935
                0.9157   CCIR601, 625line
                0.9815
                1.0255
                1.0695
                1.0950   CCIR601, 525line
                1.1575
                1.2015

          For video decompressors, this is read-only.

     CCCCLLLL____CCCCOOOOMMMMPPPPRRRREEEESSSSSSSSIIIIOOOONNNN____RRRRAAAATTTTIIIIOOOO
          For compressors, setting this parameter will result in an
          appropriate CL_BITRATE to be calculated (and vice versa).

     CCCCLLLL____EEEEXXXXAAAACCCCTTTT____CCCCOOOOMMMMPPPPRRRREEEESSSSSSSSIIIIOOOONNNN____RRRRAAAATTTTIIIIOOOO
          TRUE for video and audio compressors, since bitrate targets are
          generally met.  However, some deviation from the requested
          compression ratio may occur due to the bitrates actually allowed
          (see below).

     CCCCLLLL____BBBBIIIITTTTRRRRAAAATTTTEEEE
          May be set for compressors to indicate a target bitrate.

          For compressors, setting this parameter will result in an
          appropriate CL_COMPRESSION_RATIO to be calculated (and vice
          versa).

          The allowed bitrates are

          Video : 100 Kbps - 4.5 Mbps
          Audio : (in Kbps)
           layer I  - 32, 64, 96,128,160,192,224,256,288,320,352,384,416,448
           layer II - 32, 48, 56, 64, 80, 96,112,128,160,192,224,256,320,384




                                                                        PPPPaaaaggggeeee 3333






ccccllll____mmmmppppeeeegggg1111((((3333ddddmmmm))))                                                    ccccllll____mmmmppppeeeegggg1111((((3333ddddmmmm))))



          The default bitrates are 1.158 Mbps for video and 96/192 Kbps for
          mono/stereo audio.  These correspond to single-speed CD-ROM, and
          work best for SIF-size images (352x240, 352x288, 320x240).  Values
          of CL_INTERNAL_WIDTH and CL_INTERNAL_HEIGHT that are larger should
          use higher bitrates.

          For decompressors, this is read-only.

     CCCCLLLL____NNNNUUUUMMMMBBBBEEEERRRR____OOOOFFFF____FFFFRRRRAAAAMMMMEEEESSSS
          For video and audio decompressors, this parameter keeps track of
          the known number of frames in the sequence.  As the sequence is
          decoded, the number of frames may increase.



   PPPPAAAARRRRAAAAMMMMEEEETTTTEEEERRRRSSSS :::: MMMMPPPPEEEEGGGG----1111
     This parameter is shared by the MPEG schemes.

     CCCCLLLL____MMMMPPPPEEEEGGGG1111____EEEENNNNDDDD____OOOOFFFF____SSSSTTTTRRRREEEEAAAAMMMM
          Type : CL_ENUM_VALUE
          A read-only parameter which indicates that the end of stream has
          been reached.  Used by compressors to indicate there is no more
          data to write.  Used by decompressors to indicate there is no data
          left to read.



   PPPPAAAARRRRAAAAMMMMEEEETTTTEEEERRRRSSSS :::: MMMMPPPPEEEEGGGG----1111 VVVVIIIIDDDDEEEEOOOO
     The following parameters apply only to the MPEG-1 Video scheme.

     CCCCLLLL____MMMMPPPPEEEEGGGG1111____VVVVIIIIDDDDEEEEOOOO____MMMM
     CCCCLLLL____MMMMPPPPEEEEGGGG1111____VVVVIIIIDDDDEEEEOOOO____NNNN
          Type : CL_RANGE_VALUE
          Used to select IPB pattern for video compressors.  M is the
          distance between reference frames (I or P), and N is the distance
          between I frames, measured in numbers of frames.  M ranges from
          1 to 12, and N ranges from 1 to 60.  The first call to clCompress
          checks that N/M is an integer.  Default is M=4, N=16.  The actual
          encoded pattern may be different due to automatic I frame insertion
          at scene changes.

               M      N        Pattern
               1      1        IIIIII IIIIII
               1      3        IPPIPP IPPIPP
               3      3        BBIBBI BBIBBI
               3      6        BBIBBP BBIBBP

     CCCCLLLL____MMMMPPPPEEEEGGGG1111____VVVVIIIIDDDDEEEEOOOO____MMMMAAAAXXXX____MMMMOOOOTTTTIIIIOOOONNNN____VVVVEEEECCCCTTTTOOOORRRR____LLLLEEEENNNNGGGGTTTTHHHH____PPPP____XXXX,,,,
     CCCCLLLL____MMMMPPPPEEEEGGGG1111____VVVVIIIIDDDDEEEEOOOO____MMMMAAAAXXXX____MMMMOOOOTTTTIIIIOOOONNNN____VVVVEEEECCCCTTTTOOOORRRR____LLLLEEEENNNNGGGGTTTTHHHH____PPPP____YYYY,,,,
     CCCCLLLL____MMMMPPPPEEEEGGGG1111____VVVVIIIIDDDDEEEEOOOO____MMMMAAAAXXXX____MMMMOOOOTTTTIIIIOOOONNNN____VVVVEEEECCCCTTTTOOOORRRR____LLLLEEEENNNNGGGGTTTTHHHH____BBBB____XXXX,,,,
     CCCCLLLL____MMMMPPPPEEEEGGGG1111____VVVVIIIIDDDDEEEEOOOO____MMMMAAAAXXXX____MMMMOOOOTTTTIIIIOOOONNNN____VVVVEEEECCCCTTTTOOOORRRR____LLLLEEEENNNNGGGGTTTTHHHH____BBBB____YYYY
          Type : CL_RANGE_VALUE



                                                                        PPPPaaaaggggeeee 4444






ccccllll____mmmmppppeeeegggg1111((((3333ddddmmmm))))                                                    ccccllll____mmmmppppeeeegggg1111((((3333ddddmmmm))))



          For video compressors, the maximum search range (in pixels) for
          motion vectors.  Vectors in P frames have maximum lengths of P_X
          and P_Y, and in B frames have maximum lengths of B_X and B_Y.
          Higher values may give better image quality at the expense of more
          computation.  Allowable parameter values range from 1 to 60.
          Defaults are P_X=48, P_Y=32, B_X=32, and B_Y=24.

     CCCCLLLL____MMMMPPPPEEEEGGGG1111____VVVVIIIIDDDDEEEEOOOO____SSSSOOOOFFFFTTTTNNNNEEEESSSSSSSS
          Type : CL_ENUM_VALUE
          The smoothing filter parameter for video compressors.  May be set
          to one of

                CL_MPEG1_VIDEO_SOFTNESS_NONE (default)
                CL_MPEG1_VIDEO_SOFTNESS_MEDIUM
                CL_MPEG1_VIDEO_SOFTNESS_MAXIMUM

          Setting this parameter modifies the filter that is applied when
          scaling from the cropped window to the internal dimensions.



   PPPPAAAARRRRAAAAMMMMEEEETTTTEEEERRRRSSSS :::: MMMMPPPPEEEEGGGG----1111 AAAAUUUUDDDDIIIIOOOO
     The following parameters apply only to the MPEG-1 Audio scheme.

     CCCCLLLL____MMMMPPPPEEEEGGGG1111____AAAAUUUUDDDDIIIIOOOO____LLLLAAAAYYYYEEEERRRR
          Type : CL_ENUM_VALUE
          The layer used for audio encoding.  Allowed values are

                 CL_MPEG1_AUDIO_LAYER_I
                 CL_MPEG1_AUDIO_LAYER_II

          For compressors this may be set to select the layer.  For
          decompressors, this is a read-only parameter.

     CCCCLLLL____MMMMPPPPEEEEGGGG1111____AAAAUUUUDDDDIIIIOOOO____MMMMOOOODDDDEEEE
          Type : CL_ENUM_VALUE
          The mode in which input channel(s) are encoded.  Allowed values are

                 CL_MPEG1_AUDIO_MODE_STEREO
                 CL_MPEG1_AUDIO_MODE_JOINT
                 CL_MPEG1_AUDIO_MODE_DUAL
                 CL_MPEG1_AUDIO_MODE_SINGLE

          For compressors, this may be set to select the encoding mode.
          Stereo channels should be encoded with CL_MPEG1_AUDIO_MODE_STEREO
          or CL_MPEG1_AUDIO_MODE_JOINT.  The latter exploits correlation
          between channels for more compression.  Two independent channels
          should be encoded with CL_MPEG1_AUDIO_MODE_DUAL.  A mono signal
          should be encoded with CL_MPEG1_AUDIO_MODE_SINGLE.

          For decompressors, this is a read-only parameter.




                                                                        PPPPaaaaggggeeee 5555






ccccllll____mmmmppppeeeegggg1111((((3333ddddmmmm))))                                                    ccccllll____mmmmppppeeeegggg1111((((3333ddddmmmm))))



NNNNOOOOTTTTEEEESSSS
     This section provides miscellaneous information and tips on how to use
     the MPEG schemes.


   EEEErrrrrrrroooorrrr CCCCooooddddeeeessss
     CL_BAD_LICENSE may be returned by clOpenCompressor if the appropriate
     license is not installed.

     CL_MPEG1_ERROR may be returned by a CL call if an MPEG-specific error
     occurs, e.g. a non-compliant data stream.


   VVVViiiiddddeeeeoooo CCCCoooommmmpppprrrreeeessssssssiiiioooonnnn
     Since the MPEG video compressor reorders input frames, compressed bits
     that are generated by a call to clCompress do not necessarily correspond
     to the frame that is consumed from the input buffer.

     The compressor maps the cropped window to the internal width and height.
     If the cropped window and the internal dimensions are not related by the
     same scale factor in each dimension, the final image is stretched more
     along one axis than the other.



FFFFIIIILLLLEEEESSSS
     /usr/include/dmedia/cl.h


SSSSEEEEEEEE AAAALLLLSSSSOOOO
     CLintro(3dm), mpeg(4)
























                                                                        PPPPaaaaggggeeee 6666