`
`
`
`
`(12) United States Patent
`
`
`
`US 7,012,960 B2
`(10) Patent N0.:
`
`
`
`
`
`
`
`(45) Date of Patent:
`*Mar. 14, 2006
`Bourge et al.
`
`
`
`U3007012960B2
`
`
`
`(54) METHOD OF TRANSCODING AND
`
`
`
`
`TRANSCODING DEVICE WITH EMBEDDED
`
`
`
`FILTERS
`
`
`
`(75)
`
`
`
`( * ) Notice:
`
`
`
`
`
`
`
`
`
`
`Inventors: Arnaud Bourge, Paris (FR); Anthony
`More], Xi’An (FR)
`
`
`
`
`
`
`
`(73) Assignee: Koninklijke Philips Electronics N.V.,
`
`
`Eindhoven (NL)
`
`
`Subject to any disclaimer, the term of this
`
`
`
`
`
`
`
`patent is extended or adjusted under 35
`
`
`
`
`
`
`
`
`U.S.C. 154(b) by 712 days.
`This patent is subject to a terminal dis-
`
`
`
`
`elaimer.
`
`
`1,1
`5,493 513 A x
`
`5,537,440 A X
`
`
`2/1996 Keith et al.
`................. 709/247
`
`
`
`
`7/1996 Eyuboglu et al.
`...... 375/24003
`
`
`
`
`
`
`
`
`
`
`EP
`
`EP
`
`WO
`
`
`(Continued)
`
`FOREIGN PATENT DOCUMENTS
`
`
`0690392 A1
`6/1995
`
`
`1032217 A2
`12/1999
`
`
`W00051357
`/2000
`
`
`OTHER PUBLICATIONS
`
`
`A. Morel et a1, “Spatial and Temporal Filtering in a Low-
`
`
`
`
`
`
`
`
`Cost MPEG Bit-Rate Transcoder”, 2001 IEEE International
`
`
`
`
`
`
`
`Conference on Acoustics, Speech, and Signal Processing;
`
`
`
`
`
`
`
`Salt Lake, UT, vol. 3, 2001, pp. 1885-1888, XP002189333.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`(21) Appl. N0.: 10/082,860
`(22) Filed:
`Oct. 19, 2001
`
`
`
`
`
`Prior Publication Data
`
`
`
`
`US 2002/0136311 A1
`Sep. 26, 2002
`
`
`
`
`
`(65)
`
`
`
`
`Primary Examiner—Shawn S. An
`
`
`
`(Continued)
`
`(57)
`
`
`
`ABSTRACT
`
`
`
`(30)
`
`(51)
`
`(56)
`
`
`
`
`
`
`
`
`Foreign Application Priority Data
`Oct. 24, 2000
`(EP)
`.................................. 00402939
`
`
`
`
`
`
`Mar. 6, 2001
`.................................. 01400588
`(EP)
`
`
`
`
`Int. Cl.
`
`
`
`(2006.01)
`H04B 1/66
`
`
`
`........................... 375/240.12; 375/240.03;
`(52) US. Cl.
`
`
`
`
`
`375/24016; 375/24025; 375/240.26; 375/240.24;
`
`
`
`348/699; 382/251; 382/233; 382/236; 382/235
`
`
`
`
`(58) Field of Classification Search ........... 375/240.12,
`
`
`
`
`
`375/240.25, 240.24, 240.29, 240.16, 240.03,
`
`
`
`
`375/240.26, 348/699; 382/251, 233, 236,
`
`
`382/235
`See application file for complete search history.
`
`
`
`
`
`
`
`References Cited
`
`
`U.S. PATENT DOCUMENTS
`
`
`
`.......... 348/443
`5,121,191 A *
`6/1992 Cassereau et a1.
`
`
`
`
`
`5,451,954 A *
`9/1995 Davis et a1.
`................ 341/200
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`The present invention relates to a method of transcoding
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`(200) a primary encoded signal (81) into a secondary
`
`
`
`
`
`
`
`encoded signal (52). Said transcoding method comprising at
`least a step of decoding a current picture of the primary
`
`
`
`
`
`
`
`
`
`encoded signal, said decoding step comprising a dequantiz-
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`ing sub-step (12) for producing a first transformed signal
`
`
`
`
`
`
`
`
`
`(R1), an encoding step, following the decoding step, for
`obtaining the secondary encoded signal, said encoding step
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`comprising a quantizing sub-step (13), and a step of pre-
`
`
`
`
`
`dicting a transformed motion-compensated signal (Rmc)
`from a transformed encoding error (Re) derived from the
`
`
`
`
`
`
`
`
`encoding step, said prediction step being situated between
`
`
`
`
`
`
`
`
`the encoding and decoding steps. Said method of transcod-
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`ing further comprises a filtering step (21), between the
`
`
`
`
`
`
`
`dequantizing sub-step and the quantizing sub-step,
`for
`
`
`
`
`
`
`
`
`obtaining a better picture quality for low bitrate applications.
`
`
`
`
`8 Claims, 3 Drawing Sheets
`
`
`
`
`
`
`
`
`
`
`Page 1 ofll
`
`GOOGLE EXHIBIT 1001
`
`Page 1 of 11
`
`GOOGLE EXHIBIT 1001
`
`

`

`
`
`US 7,012,960 B2
`Page 2
`
`
`U.S. PATENT DOCUMENTS
`
`
`
`Kim ...................... 375/24U.16
`4/1997
`5,621,468 A *
`
`
`
`
`
`Zhu ........................... 375/240
`2/1999
`5,870,146 A *
`
`
`
`
`
`6,041,068 A *
`3/2000
`Rosengren et al.
`......... 370/538
`
`
`
`
`
`
`6,178,205 B1 *
`1/2001
`Chcung ct a1.
`........ 375/240.29
`
`
`
`
`
`
`
`
`6,434,197 B1 *
`8/2002
`Wang et a1.
`..
`.. 375/240.29
`
`
`
`
`
`
`
`Vetro et a1.
`............ 375/240.16
`6,671,322 B1 * 12/2003
`
`
`
`
`
`
`
`
`
`
`OTHER PUBLICATIONS
`
`
`Shih-Fu Chang et a1; “Manipulationi and Compositing of
`
`
`
`
`
`
`MC—DCT Compressed Video” IEEE Journal on Selected
`
`
`
`
`
`
`
`Areas in Cmmunicatins, IEEE INC. NY, V01. 13, No. 1,
`
`
`
`
`
`
`
`
`1995, pp. 1-11, XP000492740,
`
`
`
`
`* Cited by cxamincr
`
`
`
`
`
`Page 2 ofll
`
`Page 2 of 11
`
`

`

`
`US. Patent
`
`
`
`
`h43r.14,2006
`
`
`
`Sheet],0f3
`
`
`
`
`
`US 7,012,960 132
`
`f'""-""'""'"""'""'""'"""""""""""“'l
`
`
`FIG. 1
`
`
`
`
`
`
`
`
`FIG. 2
`
`Page 3 ofll
`
`Page 3 of 11
`
`
`

`

`
`US. Patent
`
`
`
`
`h43r.14,2006
`
`
`
`SheetZ 0f3
`
`
`
`
`US 7,012,960 132
`
`
`
`
`
`
`
`
`FIG. 3
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`—-———————.——_———————_—.———————__.—.
`
`
`
`FIG. 4
`
`Page 4 ofll
`
`Page 4 of 11
`
`
`

`

`
`US. Patent
`
`
`
`
`Mar. 14, 2006
`
`
`
`
`Sheet 3 0f3
`
`
`
`US 7,012,960 B2
`
`
`
`
`
` u—_—.——-————-a_.—.—_——-——
`
`-————nn-——————-—fiu——n—‘n—
`
`I l I I l I I I l I I I l I I I l I I I I I I I I I I I I
`
`.l‘
`
`l
`
`I—
`
`ml8!$1
`
`Page 5 ofll
`
`Page 5 of 11
`
`

`

`
`
`US 7,012,960 B2
`
`1
`
`METHOD OF TRANSCODING AND
`
`
`
`TRANSCODING DEVICE WITH EMBEDDED
`
`
`
`FILTERS
`
`
`FIELD OF THE INVENTION
`
`
`
`
`
`
`
`The present invention relates to a method of transcoding
`
`
`
`
`
`
`
`
`
`
`
`a primary encoded signal comprising a sequence of pictures,
`
`
`
`
`
`
`into a secondary encoded signal, said transcoding method
`comprising at least the steps of:
`
`
`
`
`
`decoding a current picture of the primary encoded signal,
`
`
`
`
`
`
`
`
`
`
`
`
`
`said decoding step comprising a dequantizing sub—step for
`producing a first transformed signal,
`
`
`
`
`
`
`
`
`
`
`
`encoding, following the decoding step, for obtaining the
`
`
`
`
`
`
`secondary encoded signal, said encoding step comprising
`
`
`
`a quantizing sub-step.
`The invention also relates to a corresponding device for
`
`
`
`
`
`
`
`
`
`
`
`carrying out such a method of transcoding.
`This invention is particularly relevant to the transcoding
`
`
`
`
`
`of MPEG encoded video signals.
`
`
`
`
`
`BACKGROUND OF THE INVENTION
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Bitratc transcoding is a technique which allows a primary
`
`
`
`
`
`
`video stream encoded at a bitrate BR1 to be converted into
`
`
`
`
`
`
`
`a secondary video stream encoded at a bitrate BR2, lower
`
`
`
`
`
`
`
`
`than BR1, the bitrate reduction being performed in order to
`
`
`
`
`
`
`
`meet requirements imposed by the means of transport during
`
`
`
`
`
`
`broadcasting. A transcoding device as described in the
`
`
`
`
`
`
`
`
`
`
`
`
`
`opening paragraph is disclosed in European Patent Appli-
`
`
`
`
`
`
`
`
`cation n° EP 0690 392 (PHF 94001) and is depicted in FIG.
`
`
`
`
`
`
`
`
`
`1. Said device (100) for transcoding encoded digital signals
`
`
`
`
`
`
`
`
`(SI) which are representative of a sequence of images,
`
`
`
`
`
`
`comprises a decoding channel (11,12) followed by an encod-
`ing channel (13,14,15). Aprediction channel is connected in
`
`
`
`
`
`
`cascade between these two channels, and said prediction
`
`
`
`
`
`
`
`
`channel comprises, in series, between two snbtractors (101,
`
`
`
`
`
`
`
`102), an inverse discrete cosine transform circuit IDCT (16),
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`a picture memory MEM (17), a circuit for motion-compen-
`sation MC (18) in view of displacement vectors (V) which
`
`
`
`
`
`
`are representative of the motion of each image, and a
`
`
`
`
`
`
`
`
`
`discrete cosine transform circuit DCT (19).
`
`
`
`
`
`
`SUMMARY OF THE INVENTION
`
`
`
`
`
`
`
`
`
`It is an object of the invention to provide a transcoding
`
`
`
`
`
`
`method and a corresponding device that allows a better
`
`
`
`
`
`
`quality of pictures for low bitrate applications. The present
`
`
`
`
`
`
`
`
`invention takes the following aspect into consideration.
`
`
`
`
`
`
`
`With the advent of home digital Video, recording of
`
`
`
`
`
`
`
`
`
`MPEG broadcasts, transcoders can be used in consumer
`
`
`
`
`
`
`
`devices to implement long-play modes or to guarantee the
`
`
`
`
`
`
`recording time. However, the input signal to be transcoded
`
`
`
`
`
`
`
`has often been encoded at a variable bitrate with a low
`
`
`
`
`
`
`
`
`average bit—rate. This is due to the generalization of statis—
`
`
`
`
`
`
`
`tical multiplexing that allows broadcasters to put a lot of
`
`
`
`
`
`
`
`
`Video programs in a multiplex in order to save the band-
`
`
`
`
`
`
`
`width. It is likely that a coarser re-quantization of the input
`
`
`
`
`
`
`
`signal, using a prior art transcoding method, will lead to
`
`
`
`
`
`
`
`
`conspicuous quantization artifacts. In consequence, such a
`
`
`
`
`
`
`transcoding method is not adapted to low bitrate applica-
`
`
`
`
`
`
`
`tions.
`
`To overcome this drawback, the transcoding method in
`
`
`
`
`
`
`
`accordance with the invention is characterized in that it
`
`
`
`
`
`
`
`
`
`
`
`
`
`further comprises a filtering step between the dequantizing
`sub-step and the quantizing sub-step.
`
`
`
`
`
`
`10
`
`15
`
`
`
`20
`
`
`
`30
`
`35
`
`
`
`40
`
`
`
`45
`
`
`
`
`
`
`
`60
`
`
`
`65
`
`
`
`Page 6 ofll
`
`2
`
`The transcoding method in accordance with the invention
`
`
`
`
`
`
`
`allows filters to be implemented at negligible cost in the
`
`
`
`
`
`
`
`
`prior art transcodcr. These filters can be tuned to control the
`
`
`
`
`
`
`
`
`
`static and dynamic resolution and also to effect noise reduc-
`
`
`
`
`
`
`
`
`
`tion. For the same number of bits, the filtered transformed
`
`
`
`
`
`
`
`
`
`signal is encoded with a smaller quantization scale, thus
`
`
`
`
`
`
`
`reducing visual artifacts such as blocking,
`ringing and
`
`
`
`
`
`
`
`
`mosquito noise.
`
`
`In a first embodiment of the invention, the method of
`
`
`
`
`
`
`
`
`
`
`
`
`
`transcoding comprises a step of predicting a transformed
`1110tion-c0111pensated signal from a transformed encoding
`
`
`
`
`
`error derived from the encoding step, said prediction step
`
`
`
`
`
`
`
`
`
`being situated between the encoding and decoding steps, and
`
`
`
`
`
`
`
`
`is characterized in that
`the filtering step is a temporal
`
`
`
`
`
`
`
`
`
`filtering step for receiving the transformed motion-compen-
`
`
`
`
`
`
`
`sated signal and the first transformed signal and for deliv-
`
`
`
`
`
`
`
`
`
`
`ering a filtered transformed signal to the quantizing sub—step.
`
`
`
`
`
`
`
`Such a temporal filtering step allows noise reduction to be
`
`
`
`
`
`
`
`performed using, for example, a recursive filter. In conse-
`
`
`
`
`
`
`
`
`quence, bits are only spent on the useful information con-
`
`
`
`
`
`
`
`
`
`tained in the picture and the picture quality is thus increased.
`
`
`
`
`
`
`
`
`In a second embodiment of the invention, the method of
`
`
`
`
`
`
`
`transcoding also comprises a prediction step and is charac-
`
`
`
`
`
`
`
`terized in that the filtering step is a spatial filtering step for
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`receiving the first transformed signal and for producing a
`filtered transformed signal, said filtered transformed signal
`
`
`
`
`
`
`
`and the transformed motion-compensated signal being
`
`
`
`
`
`
`delivered to the quantizing sub-step.
`
`
`
`
`Such a spatial filtering allows a reduction of the sharpness
`
`
`
`
`
`of the picture and decreases the possible source of ringng
`
`
`
`
`
`
`
`
`and mosquito noise,
`
`
`
`The present invention also relates to a corresponding
`
`
`
`
`
`
`device for carrying out such a transcoding method.
`
`
`
`
`
`
`
`The present invention finally relates to a computer pro-
`
`
`
`
`
`
`
`
`gram product for a receiver, such as a digital video recorder
`
`
`
`
`
`
`
`
`or a set—top—box, which comprises a set of instructions,
`
`
`
`
`
`
`
`which, when loaded into the receiver, causes the receiver to
`
`
`
`
`
`
`
`
`
`carry out the method of transcoding.
`
`
`
`
`
`These and other aspects of the invention will be apparent
`
`
`
`
`
`
`
`from and will be elucidated with reference to the embodi—
`
`
`
`
`
`
`
`ments described hereinafter.
`
`
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`The present invention will now be described in more
`
`
`
`
`
`
`detail, by way of example, with reference to the accompa-
`
`
`
`
`
`
`nying drawings, wherein:
`
`
`
`
`
`
`
`FIG. 1 is a block diagram corresponding to a transcoding
`device according to the prior art,
`
`
`
`
`
`FIG. 2 is a block diagram corresponding to a first embodi-
`
`
`
`
`
`ment of a transcoding device according to the invention, said
`
`
`
`
`
`
`device comprising a temporal filtcr circuit,
`
`
`
`
`
`FIG. 3 is a block diagram corresponding to a second
`
`
`
`
`embodiment of a transcoding device according to the inven-
`
`
`
`
`
`tion, said device comprising a spatial filter circuit,
`
`
`
`
`
`
`
`FIG. 4 a block diagram corresponding to a third embodi-
`
`
`
`
`
`ment of a transcoding device according to the invention, said
`
`
`
`
`
`
`device also comprising a spatial filter circuit, and
`
`
`
`
`
`
`
`FIG. 5 a block diagram corresponding to a fourth embodi—
`
`
`
`
`
`ment of a transcoding device according to the invention, said
`
`
`
`
`
`
`device also comprising a spatial filter circuit and, possibly,
`
`
`
`
`
`
`
`a temporal filter circuit.
`
`
`
`
`DETAILED DESCRIPTION OF TIIE
`
`
`INVENTION
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`The present invention relates to an improved method of
`
`
`
`
`
`
`
`and a corresponding device for transcoding Video encoded
`
`
`
`
`
`
`
`
`Page 6 of 11
`
`

`

`
`
`US 7,012,960 B2
`
`3
`
`to MPEG-2 encoded
`signals. It relates, more especially,
`
`
`
`
`
`
`
`
`signals but it will be apparent to a person skilled in the art
`
`
`
`
`
`
`
`
`that said transcoding method also remains applicable to any
`
`
`
`
`
`
`
`
`type of video signals encoded via a block-based technique
`
`
`
`
`
`
`
`such as, for example, those provided by MPEG-1, MPEG-4,
`
`
`
`
`
`
`
`H—261 or H—263 standards.
`
`
`
`A transcoding device allows a primary encoded signal
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`(Sl), previously encoded with a first quantization scale and
`comprising a sequence of pictures, to be converted into a
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`secondary encoded signal (S2), encoded with a second
`
`
`quantization scale.
`Such a transcoding device comprises at least the follow-
`
`
`
`
`
`
`ing elements:
`
`
`a decoding unit, comprising a variable length decoder VLD
`
`
`
`
`
`
`
`and a first dequantizer IQ for decoding a current picture
`
`
`
`
`
`of the primary encoded signal and for delivering a first
`
`
`
`
`
`
`
`
`transformed signal,
`
`
`an encoding unit, comprising a quantizer Q, a variable
`
`
`
`
`
`
`length encoder VLC for obtaining the secondary encoded
`
`
`
`
`
`
`signal, and a second dequantizer IQ,
`
`
`
`
`
`a prediction unit, between the encoding unit and the decod-
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`ing unit, and comprising in series:
`an inverse discrete transform circuit IDCT (an Inverse
`
`
`
`
`
`
`Discrete Cosine Transform in the case of MPEG),
`
`
`
`
`
`
`
`
`
`
`a picture memory MEM,
`a circuit MC, for motion-compensation in view of dis-
`
`
`
`
`
`placement vectors which are representative of the
`
`
`
`
`
`
`motion of each picture,
`
`
`
`a discrete transform circuit DCT, for predicting a trans-
`
`
`
`
`
`
`
`
`
`
`
`
`(Rmc)
`formed motion—compensated signal
`from a
`transformed encoding error
`(Re) derived from the
`
`
`
`
`
`
`
`encoding unit,
`
`
`for determining a sum of the transformed
`an adder,
`
`
`
`
`
`
`motion-compensated signal and a transformed signal
`
`
`
`
`(R1 or Rf),
`
`
`a subtractor, for determining the transformed encoding
`
`
`
`
`
`
`error from a difference between said sum and a second
`
`
`
`
`
`
`
`
`
`
`
`
`transformed signal (R2) produced by the encoding unit,
`a filter circuit, between the first dequantizer and the quan-
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`tizer, for delivering a filtered transformed signal (Rf).
`Said filter circuit can be a temporal or a spatial filter
`
`
`
`
`
`
`circuit adapted to control the static and dynamic resolution
`
`
`
`
`
`
`
`and to perform noise reduction in a picture. The different
`
`
`
`
`
`
`implementations of such filters are described in the follow-
`
`
`
`
`
`
`ing FIGS. 2 to 5.
`
`
`It will be apparent to a person skilled in the art that the
`
`
`
`
`
`
`
`
`
`result of the transcoding device is unchanged if the adder is
`
`
`
`
`
`
`
`replaced by another subtractor adapted to determine a dif-
`
`
`
`
`
`
`ference between a transformed signal (R1 or Rf) and the
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`transformed motion-compensated signal (Rmc) and if the
`first cited subtractor is adapted to determine the transformed
`
`
`
`
`
`
`
`encoding error (Re) from a difference between the second
`
`
`
`
`
`
`
`
`transformed signal (R2) and the output of the other subtrac-
`
`
`
`
`
`
`
`
`tor.
`
`In a first embodiment of the invention,
`the transcoder
`
`
`
`
`
`
`
`
`
`
`
`
`implements a motion-compensated temporal filter. Temporal
`filtering allows to reduce signals which are not correlated
`
`
`
`
`
`
`
`
`from frame to frame. It can very effectively reduce noise
`
`
`
`
`
`
`
`
`when combined with motion-compensation, as motion-com-
`
`
`
`
`
`pensation tries to correlate the image content from frame to
`
`
`
`
`
`
`
`
`frame. In this embodiment, a recursive filter is implemented
`
`
`
`
`
`
`since it provides a better selectivity at lower cost.
`
`
`
`
`
`
`
`
`
`
`
`
`A naive transcoding chain with a motion—compensated
`recursive temporal filter usually comprises in cascade:
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Page 7 ofll
`
`4
`
`a decoder, for producing motion-compensated blocks D1 of
`
`
`
`
`
`
`decoded pictures from an input stream,
`
`
`
`
`
`a recursive temporal filter, for producing filtered blocks Df
`
`
`
`
`
`
`
`
`of decoded pictures, and
`
`
`
`
`an encoder, for producing an output stream and motion-
`
`
`
`
`
`
`
`
`
`
`
`
`compensated blocks D2 of locally decoded pictures after
`
`encoding.
`To reduce costs, the motion-compensation in the encoder
`
`
`
`
`
`
`
`is re-used in the recursive temporal filter. Thus, the signal D2
`
`
`
`
`
`
`
`
`
`is fed back to said filter instead of Df. The filtering equation
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`of a motion-compensated block Df(n,m) is then:
`Df(n,m)=(1—(L)-D1(n,m)+a-MC(D2(p(n)), V(n,m)),
`
`
`
`
`
`
`
`10
`
`15
`
`
`
`20
`
`
`
`30
`
`35
`
`
`
`40
`
`
`
`45
`
`
`
`
`
`
`
`60
`
`
`
`65
`
`
`
`
`
`(1)
`
`
`
`where:
`
`n is the index of the current picture,
`
`
`
`
`
`
`m is the index of a block of said current picture,
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`V(n,m) is the motion associated with block m, of picture n,
`
`
`
`
`
`
`
`
`p(n) is the index of the anchor picture associated with image
`1'1,
`
`
`
`
`
`
`MC is the motion-compensation operator, and
`at is a positive scalar smaller than one that tunes the filter
`
`
`
`
`
`
`
`
`
`
`
`response.
`
`
`
`
`
`
`
`
`
`An expression similar to equation (1) can be drawn for
`bidirectional motion-compensation. However, without loss
`
`
`
`
`
`of generality, we shall restrict the demonstration to the
`
`
`
`
`
`
`
`
`
`unidirectional case. Note that intra—encoded blocks cannot
`
`
`
`
`
`
`
`be filtered since no prediction is formed for them. Yet,
`
`
`
`
`
`
`
`
`
`intra-encoded blocks in non-intra pictures most often cor-
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`respond to newly exposed regions that could not possibly be
`
`
`temporally filtered.
`The naive transcoding chain can be simplified using the
`
`
`
`
`
`
`
`hypothesis that
`the motion-compensation information is
`
`
`
`
`
`unchanged. To this end,
`the motion-compensated block
`
`
`
`
`
`
`D1(n,m) is expressed as follows:
`
`
`
`D1(mm)=M"R1(nym)'M+MC(D1(P(n)), V(rt mi),
`
`
`
`
`
`(2)
`
`where:
`
`M is the 8x8 discrete cosine transform matrix,
`
`
`
`
`
`
`
`
`
`
`
`
`
`M” is the corresponding transposed matrix, and
`R1(n,m) is the residue retrieved from the input bit-stream
`
`
`
`
`
`
`
`after variable length decoding VLC and dequantization
`
`
`
`
`
`IQ.
`
`
`
`
`
`
`
`
`M is defined by equation (3) and is such that MM’=I:
`
`
`
`
`
`
`ifi=0,
`
`
`M.,,..{fi/4
`cos(i/r(2j + l)/16)/2 otherwise.
`
`
`
`
`
`
`
`
`Then, the filtered block is encoded using the same motion-
`
`
`
`
`
`
`
`
`
`
`
`
`
`compensation information. Let Rf(n,m) be the correspond-
`ing residue:
`
`
`Rflny m)=M-Qf(n,m)-M—M-MC(D2 (p(n)), V(n,m))AI’.
`
`(4)
`
`
`
`
`
`The residue is then quantized and dequantized again to
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`compute the locally decoded pictures D2. Let R2(n,m) be
`the quantized and dequantized residue:
`
`
`
`
`
`R2(n,m)=M-D2(n,m)-M‘—M'1WC(D2(p(n)), V(n,m))-M‘.
`
`(5)
`
`The equations (1) and (4) are combined, so that Rf is
`
`
`
`
`
`
`
`
`
`derived directly from D1 and D2:
`
`
`
`
`
`Rflrt m)=(1 —U.) [Al-D101, mt) -M’—M-MC(D2(p(n)) , V(n,
`m))-1W].
`
`
`
`
`
`
`
`Combining the equation (2) with equation (6) gives:
`Rflnm)=(1—O.)[R1(n,m)+M-MC(D1(p(n)), V(n,m))-
`
`
`M’—M'MC(D2(p(n)), V(n, m))M].
`
`
`
`(6)
`
`
`
`(7)
`
`
`
`
`
`Page 7 of 11
`
`

`

`5
`
`
`
`
`
`
`Since motion-compensation is performed identically from
`D1 and from D2, the motion-compensation operator MC can
`
`
`
`
`
`operate on the picture difference, i.e., on the error signal due
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`to the transcoding operation. Defining 6D=D1—D2, equation
`(7) is rewritten as follows:
`
`
`
`thn, m)‘=(1—0t)[R1(n,m)+A[-MC(5D(p(n)), V(n,m))
`M]-
`
`
`The error signal 6D can be derived from the prediction
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`errors, combining equations (5) and (6):
`
`(8)
`
`6D(n, m) = M‘A[
`
`
`
`Rfin, m)
`
`lid
`7 R2(n, m)] - M.
`
`
`
`
`
`
`(9)
`
`
`
`
`
`
`
`
`
`
`
`Equations (8) and (9) define the transcoder structure
`depicted in FIG. 2. Said transcoder (200) comprises:
`
`
`
`
`
`
`a decoding channel, comprising a variable length decoder
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`VLD (11) and a first dequantizer IQ (12) for decoding a
`
`
`
`
`
`
`
`
`current picture of a primary encoded signal (51) and for
`
`
`
`
`
`producing a first transformed signal (R1),
`
`
`
`
`
`
`an encoding channel, comprising a quantizer Q (13), a
`
`
`
`
`
`
`
`
`variable length encoder VLC (14) for obtaining the sec-
`
`
`
`
`
`
`
`ondary encoded signal (S2), and a second dequantizer IQ
`
`
`
`
`
`
`
`(15) for delivering a second transformed signal (R2),
`a prediction channel, comprising in series:
`
`
`
`
`
`
`
`
`
`
`
`a subtractor (201), for determining a transformed encod—
`
`
`
`
`
`
`
`
`ing error (Re) and whose negative input receives the
`second transformed signal,
`
`
`
`an inverse discrete cosine transform circuit IDCT (16),
`
`
`
`
`
`
`
`
`
`
`
`a picture memory MEM (17),
`
`
`
`
`a circuit for motion-compensation MC (18),
`a discrete cosine transform circuit DCT (19), for predict-
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`ing a transformed motion-compensated signal (Rnic),
`an adder (202), for delivering a sum of the transformed
`
`
`
`
`
`
`
`
`motion—compensated signal and the first transformed
`
`
`
`
`
`
`
`
`
`
`
`
`
`signal (R1) to the positive input of the subtractor,
`
`
`
`
`
`
`
`
`
`
`a temporal filter circuit Wt (21), for receiving said sum and
`
`
`
`
`
`
`
`
`for delivering the filtered transformed signal (Rf) to the
`
`
`quantizer Q (13).
`In an advantageous variant of the invention, the strength
`
`
`
`
`
`
`
`filter
`of the motion-compensated recursive temporal
`is
`
`
`
`
`
`
`
`adjusted separately for each transformed coefficient Rf[i],
`
`
`
`
`
`
`
`i.e., for each DCT sub-band. The transformed coefficient of
`
`
`
`
`
`
`
`
`rank i is multiplied by W[i]=1—o.[i] such as:
`
`
`
`
`
`
`Rf[i]=W[i](R1[i]+Rmc[i])
`
`
`
`
`
`
`(10)
`
`
`
`
`
`Thus, the noise reduction can be tuned to the spectral
`
`
`
`
`
`
`
`shape of the noise. It can also be decided not to filter low
`
`
`
`
`
`
`
`
`frequencies in order to avoid visible artifact in case of a bad
`
`
`
`
`
`
`motion—compensation and in order to reduce the noise.
`
`
`
`
`
`
`In the second and third embodiments of the invention, the
`
`
`
`
`
`
`
`
`transcoder implements a spatial filter. Spatial filtering is not
`
`
`
`
`
`
`
`so efficient
`to reduce the noise as motion—compensated
`
`
`
`
`
`
`
`
`temporal filtering is. Yet, it can prevent blocking artifacts at
`
`
`
`
`
`
`
`
`low bit-rate, smoothing down sharp edges that would oth-
`
`
`
`
`
`
`
`
`
`erwise create ringing effects. It can also simplify complex
`
`
`
`
`
`
`
`
`patterns that would otherwise be randomly distorted from
`
`
`
`
`
`
`
`one picture to the other, resulting in the so-called mosquito
`
`
`
`
`
`
`
`
`noise.
`
`Let us consider again the naive transcoding chain. The
`
`
`
`
`
`
`
`
`pixel domain filter shall have the same granularity that the
`
`
`
`
`
`
`
`
`
`
`granularity the decoder has. Thus we consider a block-wise
`
`
`
`
`
`
`
`filter. Let D1(n,m) be block In of picture n. The filtered block
`
`
`
`
`
`
`
`
`
`
`
`D1(n,m) is computed as follows:
`Df(n,m)=FV(n)-D1(n,m)-Fh’(n)
`
`(11)
`
`Page 8 ofll
`
`5
`
`
`
`10
`
`15
`
`
`
`20
`
`
`
`30
`
`35
`
`
`
`40
`
`
`
`45
`
`
`
`
`
`
`
`60
`
`
`
`65
`
`
`
`
`
`US 7,012,960 B2
`
`6
`
`
`
`
`
`
`
`
`
`where Fv(n) and Fh(n) are matrices which define respec-
`tively the vertical and horizontal filtering within the
`
`
`
`
`
`
`
`block.
`
`
`
`
`
`
`
`
`
`Combining the equation (11) with the equation (2), we
`find:
`
`
`Df(n, m:)=Fv(n) -M"R1(n, m)-M-Fh‘(n) +Fv(n) -MC(D1(p
`(11)), V(n7 m)) T'h‘(n)
`
`(1 2)
`
`then
`If the filter is the same for a group of pictures,
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Fv(n)=Fv(p(n)) and Fh(n) =Fh(p(n)). Thus, the following
`
`
`
`
`
`
`
`approximation can be given for equation (12) based on the
`assumption that block-wise filtering commutes with motion-
`
`
`
`
`
`
`
`compensation:
`Df(n, m)=Fv(n) -M‘-R1(11, m)uM-Fht(n)+AlC(Df(p(n)),
`VGA "0)
`
`(13)
`
`
`
`
`
`
`
`
`
`
`
`
`(14)
`
`It follows that the block-wise filter can be applied to
`
`
`
`
`
`
`
`
`
`residue R1(n,m) after an inverse discrete cosine transform
`
`
`
`
`
`
`
`IDCT. To implement the spatial filter in the transcoder, the
`
`
`
`
`
`
`
`
`
`
`
`
`
`residue R1(n,m) needs to be substituted by:
`Rfln, r11):M-Fv(n)-M"R1(n,m)-Z\l-l‘:I/t‘(n)-MI
`
`
`
`
`
`
`
`
`
`Even if the matrices M-Fv(n)-M’ and M~Fh’(n)~M’ can be
`pre-computcd,
`their computing seems to involve many
`
`
`
`
`
`
`operations. Said computing can be simplified for a class of
`
`
`
`
`
`
`
`block-wise filters for which the two matrices are diagonal.
`
`
`
`
`
`
`
`
`
`Such filters are symmetric filters with an even number of
`
`
`
`
`
`
`
`
`In our embodiment, we consider normalized 3-tap
`taps.
`
`
`
`
`
`
`
`
`symmetric filters, since they are more suitable for small
`
`
`
`
`
`
`
`
`
`blocks. Such filters have a single parameter, denoted a. The
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`corresponding pixel domain filtering matrix, (Fig-)Ogizkg, is
`defined by:
`
`
`
`fori=j=lto 6,
`a
`
`
`
`
`
`
`
`for i = j: l,
`l
`F
`1
`
`"J_2+a l+a fori=j=0and7,
`
`
`otherwise.
`0
`
`
`
`Then,
`
`
`
`
`.
`1
`
`Maw]:
`2 + a
`'
`
`
`
`
`2cos in 8 +1;
`
`
`( /)
`O
`
`
`
`
`for i: '
`
`‘1
`otherwrse.
`
`
`(15)
`
`
`
`(16)
`
`
`
`Thus, to implement filtering with horizontal parameter ah
`
`
`
`
`
`
`and vertical parameter av, the residue R1(n,m) needs to be
`
`
`
`
`
`
`
`i,]'<8
`
`
`
`
`
`weighted (component-wise) by (WSW-)0:
`defined as fol-
`lows:
`
`
`.
`2cos(i7r/8) + av 2cos(j7r/8) + a»,
`
`
`
`
`WSW- :— -—
`2+av
`2+5”,
`
`
`
`
`£17)
`
`
`
`
`
`
`FIG. 3 shows a transcoder with spatial pre—filtering
`
`
`
`
`
`according to the second embodiment of the invention. Said
`
`
`
`
`
`
`
`
`
`transcoder (300) comprises:
`a decoding channel, comprising a variable length decoder
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`VLD (11) and a first dequantizer IQ (12) for producing a
`first transformed signal (R1),
`
`
`
`
`
`
`
`
`
`
`
`
`
`for receiving said first
`a spatial filter circuit Ws (31),
`transformed signal and for producing the filtered trans-
`
`
`
`
`
`
`
`
`
`
`formed signal (Rf),
`
`
`
`
`
`
`an encoding channel, comprising a quantizer Q (13), a
`
`
`
`
`
`
`
`variable length encoder VLC (l4) and a second dequan-
`
`
`
`
`
`
`
`tizer IQ (15) for producing a second transformed signal
`(R2),
`
`
`
`
`
`Page 8 of 11
`
`

`

`
`
`US 7,012,960 B2
`
`7
`
`a prediction channel, comprising in series:
`
`
`
`
`
`
`
`
`
`
`
`a subtractor (201), for determining a transformed encod-
`
`
`
`
`
`
`
`
`ing error (Re) and whose negative input receives the
`second transformed signal,
`
`
`
`an inverse discrete cosine transform circuit IDCT (16),
`
`
`
`
`
`
`
`
`
`
`
`a picture memory MEM (17),
`
`
`
`
`a circuit for motion-compensation MC (18),
`
`
`
`
`
`
`
`a discrete cosine transform circuit DCT (19) for predict-
`
`
`
`
`ing a transformed motion-compensated signal (Rmc),
`and
`
`
`
`
`
`
`
`
`
`an adder (302), for delivering a sum of said transformed
`motion-compensated signal and the filtered trans-
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`formed signal (Rf) to the positive input of the subtrac-
`tor.
`
`FIG. 4 is a transcoder according to the third embodiment
`
`
`
`
`
`of the invention, with spatial post-filtering whose weight
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`factors are Wsiyf. Said transcoder (400) comprises:
`a decoding channel, comprising a variable length decoder
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`VLD (11) and a first dequantizer IQ (12) for producing a
`
`
`
`
`first transformed signal (R1),
`
`
`
`
`
`
`an encoding channel, comprising a quantizer Q (13), a
`
`
`
`
`
`
`
`
`variable length encoder VLC (14) and a second dequan-
`
`
`
`
`
`
`
`
`
`tizer IO (15), and further comprising an inverse filter
`
`
`
`
`
`
`
`circuit (42) for producing a second transformed signal
`(R2),
`
`a prediction channel, comprising in series:
`
`
`
`
`
`
`
`
`
`
`
`a subtractor (201), for determining a transformed encod-
`
`
`
`
`
`
`
`
`ing error (Re) and whose negative input receives the
`second transformed signal,
`
`
`
`an inverse discrete cosine transform circuit IDCT (16),
`
`
`
`
`
`
`
`
`
`
`
`a picture memory MEM (17),
`
`
`
`
`a circuit for motion—compensation MC (18),
`a discrete cosine transform circuit DCT (19), for predict-
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`ing a transformed motion-compensated signal (Rmc),
`an adder (202), for delivering a sum of said transformed
`
`
`
`
`
`
`
`
`motion—compensated signal and the first transformed
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`signal (R1) to the positive input of the subtractor, and
`
`
`
`
`
`
`
`
`
`
`a spatial filter circuit Ws (41), for receiving said sum and for
`
`
`
`
`
`
`
`delivering a filtered transformed signal (Rf) to the encod-
`ing channel.
`
`
`
`
`
`
`
`
`
`Compared to pre-filtering,
`the spatial filtering is per-
`formed in the encoding part of the transcoder.
`
`
`
`
`
`
`FIG. 5 is a transcoder according to the fourth embodiment
`
`
`
`
`
`of the invention, with spatial post-filtering. Said transcoder
`
`
`
`
`
`
`
`
`(500) comprises:
`a decoding channel comprising a variable length decoder
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`VLD (11) and a first dequantizer IQ (12) for delivering
`a first transformed signal (RI),
`
`
`
`
`
`
`
`
`
`
`
`an encoding channel comprising a quantizer Q (13), a
`variable length encoder VLC (14) and a second
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`dequantizer IQ (15) for delivering a second trans—
`
`
`
`formed signal (R2),
`a prediction channel comprising in series a subtractor
`
`
`
`
`
`
`
`
`
`
`
`
`(201) for determining a transformed encoding error
`(Re) and whose negative input receives the second
`
`
`
`
`
`
`
`
`transformed signal, an inverse discrete cosine trans-
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`form circuit IDCT (16), a picture memory MEM (17),
`
`
`
`
`
`a circuit for motion—compensation MC (18), a discrete
`cosine transform circuit DCT (19) for predicting a
`
`
`
`
`
`
`
`
`
`
`
`
`transformed motion-compensated signal (Rmc), and an
`adder (202) for delivering a sum of said transformed
`
`
`
`
`
`
`
`motion—compensated signal and the first transformed
`
`
`
`
`
`
`
`
`
`
`
`
`
`signal (R1) to the positive input of the subtractor.
`Said transcoder further comprises a switch (52), which
`
`
`
`
`
`
`
`
`
`
`
`
`
`has at least two positions. In a first position (a) of the switch,
`
`
`
`
`
`
`
`a spatial filter circuit Ws (51) is adapted to receive the output
`of the adder and to deliver a filtered transformed signal (Rf)
`
`
`
`
`
`
`
`
`8
`
`
`
`
`
`
`
`
`
`
`
`to the quantizing circuit (13). In that case, and contrary to
`FIGS. 3 and 4, the spatial filter circuit is not applied to every
`
`
`
`
`
`
`
`
`
`macroblocks contained in the current picture, but is only
`
`
`
`
`
`
`
`applied to intra-coded m