US5202847A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Digital signal processing
Assignee: STMicroelectronics Ltd Great Britain
•
Filing: 1990-07-31
•
Granted: 1993-04-13
•
Expires: 2010-07-31
Abstract
Abstract Digital signal processing circuitry for calculating separable two dimensional linear transforms on blocks of data elements includes two processors coupled as a linear pipeline. Each processor carries out a one dimensional linear transform effecting multiplication of transform coefficients by repeated addition in a carry save adder network to form a plurality of inner products. The two processors may effect simultaneously different transforms on respective blocks of data.
Inventors
- 👤 John P. Bolton
- 👤 Kenneth N. Burgin
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-27 10:06
CPC Classifications
G; G06; G06F; G06F17/00; G06F17/10; G06F17/14; G06F17/147
Claims Analysis
Total Claims: 15
Claim 1
Independent Claim
164 words
Digital signal processing circuitry comprising first and second processors each for effecting a linear transform on respective blocks of digital data, each of said processors including transform variation circuitry for varying the linear transform effected by the respective processor, said first and second processors being coupled as a linear pipeline by means of pipeline circuitry, said pipeline circuitry including timing controlled circuitry arranged to control the transmission of a succession of blocks of digital data through each of said first and second processors in succession to effect a combined transform, and selection circuitry coupled to the transform variation circuitry of each of said first and second processors and arranged to select the linear transform effected by the respective processor independently of the selection of linear transform of the other processor, whereby the combined transform to be performed on each block of data by the digital signal processing circuitry is selectable independently of the combined transform to be performed on a successive block of data.
Claim 2
Dependent Claim
164 words
Digital signal processing circuitry comprising first and second processors each for effecting a linear transform on respective blocks of digital data, each of said processors including transform variation circuitry for varying the linear transform effected by the respective processor, said first and second processors being coupled as a linear pipeline by means of pipeline circuitry, said pipeline circuitry including timing controlled circuitry arranged to control the transmission of a succession of blocks of digital data through each of said first and second processors in succession to effect a combined transform, and selection circuitry coupled to the transform variation circuitry of each of said first and second processors and arranged to select the linear transform effected by the respective processor independently of the selection of linear transform of the other processor, whereby the combined transform to be performed on each block of data by the digital signal processing circuitry is selectable independently of the combined transform to be performed on a successive block of data.
Claim 3
Dependent Claim
51 words
Digital signal processing circuitry according to claim 1 in which each processor includes a bank of carry save adders to effect multiplication of transform coefficients by repeated addition to form a plurality of inner products each addition corresponding to one bit position of a data word in said blocks of data.
Claim 4
Dependent Claim
44 words
Digital signal processing circuitry according to claim 2 in which a carry propagate adder is provided in said pipeline after said bank of carry save adders of said first processor whereby carry signals are resolved before transform data is supplied to said second processor.
Claim 5
Dependent Claim
30 words
Digital signal processing circuitry according to claim 3 in which a carry propagate adder is provided in said pipeline after said bank of carry save adders in said second processor.
Claim 6
Dependent Claim
22 words
Digital signal processing circuitry according to claim 1 in which each of said first and second processors comprise substantially the same circuitry.
Claim 7
Dependent Claim
57 words
Digital signal processing circuitry according to claim 1 in which each of said processors includes a memory having pre-computed combinations of transform coefficients, memory addressing circuitry to address said memory in accordance with data in said blocks of digital data and accumulator circuitry to accumulate data from said memory in response to successive addressing of the memory.
Claim 8
Dependent Claim
45 words
Digital signal processing circuitry according to claim 6 in which each processor includes storage circuitry for storing data words forming a row or column of data from each of said blocks and supplying data bitwise from each of said words to said address in circuitry.
Claim 9
Dependent Claim
20 words
Digital signal processing circuitry according to claim 1 in which said first and second processors are implemented in CMOS circuitry.
Claim 10
Dependent Claim
43 words
Digital signal processing circuitry according to claim 1 in which each bank of adders operates in two's complement format and inverter circuitry is provided to invert data from said first processor corresponding to the most significant bits of data supplied from each word.
Claim 11
Dependent Claim
58 words
Digital signal processing circuitry according to claim 9 including an adder array comprising a plurality of adders performing a succession of adder stages wherein each adder is arranged to invert its output relative to its input and said inverter circuitry comprises a cross-over switch arranged to interchange two signals fed to respective successive stages of the adder array.
Claim 12
Independent Claim
125 words
A method of computing linear transforms on blocks of digital data, said method comprising transmitting blocks of digital data successively through first and second processors each operable to perform a plurality of different selectable linear transforms, controlling flow of data through said first and second processors in a time controlled manner to effect a linear pipeline, selecting a first set of coefficients for use in said first processor, and selecting a second set of coefficients for use in said second processor, the selection of the first and second set of coefficients being independently variable whereby a combined linear transform is computed on each block of data by the first and second processors independently of the combined linear transform computed on a successive block of data.
Claim 13
Independent Claim
125 words
A method of computing linear transforms on blocks of digital data, said method comprising transmitting blocks of digital data successively through first and second processors each operable to perform a plurality of different selectable linear transforms, controlling flow of data through said first and second processors in a time controlled manner to effect a linear pipeline, selecting a first set of coefficients for use in said first processor, and selecting a second set of coefficients for use in said second processor, the selection of the first and second set of coefficients being independently variable whereby a combined linear transform is computed on each block of data by the first and second processors independently of the combined linear transform computed on a successive block of data.
Claim 14
Dependent Claim
36 words
A method according to claim 11 comprising accumulation of a plurality of inner products representing multiplication of input data with a plurality of transform coefficients, said accumulation being effected by operation of carry save adder network.
Claim 15
Dependent Claim
27 words
A method according to claim 12 wherein said inner products are obtained by successive addressing of a memory having a plurality of precalculated combinations of transform coefficients.
References Cited (Backward Citations)
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1996-04-12 | 1996-10-12 | 1997-04-13 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2000-04-12 | 2000-10-12 | 2001-04-13 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2004-04-12 | 2004-10-12 | 2005-04-13 | $7700.00 | Window Closed / Lapsed |