← Back to Patent Explorer
EP0778709A1 Expired / Lapsed EP ⏱️ 0.0 years remaining

MPEG-2 decoding with a reduced RAM requisite by ADPCM recompression before storing MPEG decompressed data

Assignee: STMicroelectronics SRL
•
Filing: 1996-03-11
•
Granted: —
•
Expires: 2016-03-11
📄 Original PDF 🌐 Google Patents 🔄 Sync Now

Abstract

Abstract The video RAM requisite of an MPEG-2 decoder is reduced by recompressing according to an adaptive pulse code modulation scheme (ADPCM) at least the I and P pictures, after MPEG-2 decompression and before storing the relative data in the video RAM. The ADPCM recompressed and coded data written in the video RAM are decoded and decompressed during the reconstruction of a B-picture to be displayed.

Inventors

  • 👤 Danilo Pau
  • 👤 Roberto Sannino

Bibliographic Details

Application Number —
Jurisdiction EP
Data Source google_patents
Last Synchronized 2026-11-07 21:08
CPC Classifications
H; H04; H04N; H04N19/00; H04N19/42; H04N19/423; H04N19/426; H04N19/428; H04N19/50; H04N19/593; H04N19/60; H04N19/61

Claims Analysis

Total Claims: 8
Claim 1
Independent Claim 249 words
A method for reducing the video memory requisite of an MPEG-2 decoder comprising a decompression stage of the respective I, P and optionally B-pictures of the MPEG compression algorithm and where the data relative these pictures are stored in respective buffers organized in said video memory, characterized in that it comprises recompressing at least the I and P pictures after said MPEG-2 decompression and before storing the relative data in said video memory buffers, estimating the variance of each block of n∗m pels relative to the discrete cosine transform of the decompressed picture data, multiplying the value of the variance of each block by a set of coefficients while configuring a multilevel adaptive quantizer, and coding a first pel of each block with a p number of bits, the estimated value of the variance with a n-h number of bits, where h is an integer number greater than zero, and the differences between each other pel following the first pel and the average value of all the pels of the block with p-k number of bits, where k is an integer number greater than zero, according to an adaptive pulse code modulation scheme (ADPCM); storing the so coded data relative to the recompressed I and P pictures in said respective buffers of the video memory; decoding the stored data relative to the pels of said recompressed I and P images, decompressing said images according to an adaptive pulse code modulation scheme (ADPCM) and directing them toward a display unit.
Claim 2
Dependent Claim 37 words
The method according to claim 1, characterized in that it comprises preventively subdividing each block of n∗m pels into two sub-blocks of (n/2)∗m pels and carrying out said ADPCM recompression, coding-decoding and ADPCM decompression on said sub-blocks.
Claim 3
Dependent Claim 254 words
MPEG-2 video decoder, interfacing with a control bus and a data bus through which video data are written and read in respective RAM buffers extemal to the "core" of the video decoder which comprises a first "first-in-first-out" (FIFO) buffer for data acquisition and writing in a first buffer (BIT BUFFER) of an extemal DRAM memory, a start code detecting circuit (START CODE DETECTOR) synchronized by a controller (CONTROLLER), a bi-directional buffer (I/O UNIT) for on screen display data (OSD), a variable length decoder (VDL) of the compressed data input stream (BITSTREAM), an MPEG decompression block (PIPELINE-RLD, I_QUANT, I_DCT, PREDICTOR CONSTRUCTION) of the data decoded by said variable length decoder, comprising a "run length" decoder, an inverse quantizer circuit, an inverse discrete cosine transform processor, a "predictor generating circuit, a "Macroblock scan to raster scan" conversion circuit for a current B-picture upstream of a display unit, characterized in that further comprises a coding and recompression circuit (ADPCM ENCODER), according to an adaptive differential pulse code modulation scheme of the differential type, capable of recompressing decompressed I and P pictures and of coding the respective data to be stored in the respective buffers of said external memory; a decoding and decompressing and decoding circuit (ADPCM DECODER) of the stored data relative to said recompressed I and P pictures read from the respective buffers of said external memory capable of generating a stream of decoded and decompressed data relative to I and P pictures; means of motion compensation of B-pictures and of conversion "MACROBLOCK TO RASTER SCAN".
Claim 4
Dependent Claim 306 words
The decoder according to claim 3, characterized in that said recompressing and coding circuit (ADPCM ENCODER) comprises a buffer (BLOCK BUFFER) for the acquisition of blocks of decompressed I-DCT data produced by said pipeline of MPEG-2 decompression; a circuit (VARIANCE ESTIMATOR) capable of estimating the energy content of said buffer (BLOCK BUFFER) and of generating a digital value of the variance of the pel values of a block of I-DCT data to be stored in a respective buffer of said memory; programmable means for storing a plurality of pre-established digital values (ROM, PLA) selected by said digital value of variance read from said memory; a multilevel quantizer circuit (QUANTIZER) coherently conditioned by the digital value selected by the current value of the variance; a differentiating circuit (-) capable of receiving through a first input the stream of said I-DCT decompressed data block and through a second input a predictor value and to output a stream of data that are input to said multilevel quantizer circuit (QUANTIZER); a write coding circuit (CODER) of ADPCM recompressed data in the respective memory buffers, capable of receiving as input the output stream of said quantizer circuit; a generating circuit of said predictor value comprising a multiplexer (MUX) receiving through a first input the decompressed I-DCT data stream and through a second input the predictor value generated by the circuit, an adder (+) receiving through a first input the output of said quantizer circuit and through a second input the output of said multiplexer (MUX) and outputting a stream of summed data, a limiting circuit (LIM 0-255) receiving through an input said stream of summed data produced by said adder (+), followed in cascade by a circuit that generates said predictor value (COEFF) whose output is fed to a second input of said differentiating circuit (-) and of said multiplexer (MUX).
Claim 5
Dependent Claim 22 words
The decoder according to claim 4, characterized in that said programmable means (ROM, PLA) are constituted by a read only memory (ROM).
Claim 6
Dependent Claim 22 words
The decoder according to claim 4, characterized in that said programmable means of storage are constituted by a programmable logic array (PLA).
Claim 7
Dependent Claim 24 words
The decoder according to claim 4, characterized in that it comprises means (ADPCM MEMORY) for storing part of ADPCM recompressed data within the decoder.
Claim 8
Dependent Claim 24 words
The decoder according to claim 5, characterized in that said storing means (ADPCM MEMORY) are constituted by a RAM memory of 194.400 bit capacity.

References Cited (Backward Citations)

US5473377A Cited
EP0696874A2 Cited

USPTO Annuity & Maintenance Windows

Maintenance schedule is calculated for granted US utility patents.

Assigned Categories & Tags

No categories assigned to this patent yet.

Assign Category Tag