MXPA00007331A - Signal processing apparatus and method - Google Patents

Abstract

A signal-processing arrangement comprising an examining circuit (EXAM), and adjustable filter (FIL), and a signal processor (PRC). The signal processor (PRC) may be, for example, a video encoder for encoding a sequence of pictures in accordance with an MPEG standard. The examining circuit (EXAM) examines a signal (S) to be processed so as to obtain a distortion indication (IND) the distortion indication (IND) indicates to which extent distortion (DIST) would be introduced if the signal (S) were processed (PRC[S]) by means of the processor (PRC). An adjustable filter (FIL) filters the signal (S) in dependence on the distrotion indication (IND) so as to obtain a filtered signal (SF). The signal processor (PRC) processes the filtered signal (SF). Thus, the signal-processing arrangement filters the signal to be processed in a pro-active manner so as to counter distortion which might otherwise be introduced by processing the signal. Accordingly, a satisfactory signal quality can be obtained. For example, in a video-encoding application, the filter can reduce details contained in a series of pictures. This allows the series of pictures to be coded with a sufficient precision without introducing block effects which might otherwise occur if the pictures were not filtered. Moreover, since the filter is pro-actively adjusted, it filters the series of pictures more evenly than if the filter were retro-actively adjusted. Accordingly, there will be relatively little variation in resolution from one picture to another, which contributes to a satisfactory overall picture quality.

Description

APPARATUS AND METHOD FOR PROCESSING SIGNALS FIELD OF THE INVENTION The invention relates to the processing of signals such as, for example, the coding of a sequence of images according to a standard defined by the moving image expert group (MPEG).

BACKGROUND TECHNIQUE JP-A-09, 009, 260 describes an image encoding device, dynamic. What follows is a literal quote from the summary in English. For data from converted AD image frames the low frequency components are passed through a two-dimensional block unit in a variable low-pass spatial filter, a high-frequency coding processing is performed in a compressor 13, they are stored alternately in a FIFO memory and a storage amount of compressed data in the FIFO memory is verified by means of a FIFO verification device. The FIFO storage information in the verification data and the • Block activity information for each unit of the compressor block is fed to a band controller and cutoff frequency for each block set is instructed to the variable low pass spatial filter of the band controller.

BRIEF DESCRIPTION OF THE INVENTION An object of the invention is to provide signal processing that allows a better quality. The invention takes the following aspects into consideration. The processing of a signal can cause distortion. For example, encoding a sequence of images is a form of signal processing that can cause distortion. In order not to exceed a certain limit in terms of the amount of coded data obtained, it may be necessary to encode certain images with relatively low accuracy. If an image is encoded with relatively low precision, the image may be distorted in the sense that block effects occur in the frame at the end of decoding. It is possible to post the distortion caused by the processing of a signal, filtering the signal before being processed. However, if a signal is filtered, it distorts too much in the sense that certain components of the signal are attenuated or even removed. In this way, if a signal is filtered before being processed, in effect, a compromise can be made between the distortion caused by the filtering and the distortion caused by the processing. For example, if an image is filtered, the filtered image thus obtained will comprise fewer details than the original image. As a result, the filtered image can be encoded more accurately than the original image with respect to the time to certain limits in terms of the amount of encoded data obtained. However, the filtered image will have a lower resolution than the original image. In this way, the filtering of an image before being encoded implies a compromise between the accuracy of the encoding and the resolution of the image because both affect the overall quality of the image. In the prior art, the low-pass spatial filter, which filters the image data before being encoded, seems to be adjusted as a function of the amount of the encoded data contained in the FIFO output buffer or intermediate circuit. Consequently, this initially does not prevent a relatively large amount of coded data from occurring. If a relatively large amount of data actually occurs, the image data not yet encoded will be distorted to a relatively greater degree. This is because these image data have already been filtered to a relatively large degree, or have been encoded with relatively low precision, or both, to prevent a limit from being exceeded in terms of the amount of coded data obtained. In this way, the image data to be encoded will have a relatively poor quality, because the resolution is low or because they are some block effects or both. According to the invention, an examining circuit examines a signal to be processed by means of a signal processor to obtain an indication of distortion. The distortion indication indicates in what degree the distortion would be introduced if the signal as such were processed by the signal processor. An adjustable filter filters the signal depending on the distortion indication to obtain a filtered signal. The signal processor processes the filtered signal. Thus, in the invention, the adjustable filter is adjusted in a proactive manner, whereas in the prior art, the low-pass spatial filter is adjusted in a retroactive fashion. Accordingly, in the invention, there is no substantial delay in creating compromise between the distortion caused by the processing and the distortion caused by the filtering, whereas in the prior art, there is inherently some delay in creating such a commitment. In this way, the invention allows a better compromise between the two types of distortions mentioned above. As a result, the invention allows a better quality.

For example, the invention can be applied to encode a sequence of images. Assume that the examining circuit examines an image to be encoded or a portion of it. By doing this, you can indicate that if the image as such were encoded, you would get a relatively large amount of encoded data. As explained here above, this potentially causes distortion of one or more images to be encoded later. Since the invention allows the image to be filtered depending on its indication, it can prevent the encoding of the image from actually producing a large amount of encoded data. Consequently, this contributes to the quality of one or more subsequent images. Thus, in effect, the invention allows a quality exchange between the examined image and one or more of the subsequent images. Accordingly, the invention allows a satisfactory and homogeneous image quality. The invention and the additional features, which may be optionally used to implement the invention with advantage, are evident from and will be elucidated with reference to the drawings described hereinafter.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a conceptual diagram illustrating the basic features of the invention as claimed in claim 1. Figure 2 is a block diagram illustrating an example of a video encoder; Figure 3 is a block diagram illustrating a video coding arrangement according to the invention; Figure 4 is a graph illustrating a first filter control feature for the video coding array illustrated in Figure 3. Figure 5 is a graph illustrating a second filter control feature for the video coding array illustrated in Figure 3.

DETAILED DESCRIPTION OF THE DRAWINGS First, some observations will be made about the use of the reference signs. Similar entities were denoted by an identical alphabetic code through the drawings. The different similar entities can be shown in a single drawing. In that case, a number was added to the alphabetic code to distinguish similar entities from each other. The numbering will be in parentheses if the number of similar entities is an operating parameter. In the description and in the claims, any numbering in a reference sign will be omitted from it if appropriate. Figure 1 illustrates the basic characteristics of the invention in solid lines. A signal processing array comprises an EXAM examiner circuit, a FIL preprocessor, and a PRC signal processor. The EXAM examining circuit examines an S signal to be processed to obtain an indication of distortion IND. The indication of distortion IND indicates in what degree the DIST distortion would be introduced if the signal S was processed PRC [S] by means of the processor PRC. The preprocessor FIL preprocesses the signal S depending on the indication of distortion IND to obtain a preprocessed signal SF. The PRC signal processor processes the preprocessed signal SF. The features illustrated in Figure 1 can be applied, for example, to encode a sequence of images. Such an application, the signal S is a data stream representing the sequence of images to be encoded. The PRC signal processor is a video encoder which can distort one or more images to a relatively large degree. This will be explained in more detail here later.

Figure 2 illustrates an example of a video encoder. The video encoder provides an MPEG data stream OUTPUT having a certain bit rate R in response to an image data stream ENTRY representing a sequence of images. The input of the image data stream can be obtained for example, by converting an analog video signal to a digital video signal. The video encoder comprises the following functional blocks: an IBUF input buffer, a SUB subtractor, a MEM memory, a MEC motion estimator and compensator, a discrete cosine DCT transformer, a discrete inverse cosine IDCT transformer, a Q quantizer, an IQ inverse quantizer, an ADD adder, a variable length encoder VLC, an OBUF output buffer, and a CON driver. The video encoder illustrated in Figure 2 encodes most images in the following manner. The IBUF input buffer temporarily stores an image to be encoded. The image to be encoded is, in effect, subdivided into macroblocks of 16 by 16 pixels. It is assumed that the MEM memory contains a decoded version of a previously encoded image which will be subsequently referred to as the previous image. For each macroblock in the image to be encoded, the MEC motion estimator and compensator looks for a similar macro block of 16 by 16 pixels in the previous image. The MEC motion estimator and compensator supplies the similar macroblock that has found the SUB subtractor. The substractor SUB provides a differential macroblock by subtracting, from the macroblock to be encoded, the similar macroblock provided by the MEC motion estimator and compensator. The discrete cosine transformer DCT transforms the differential macroblock into a block of frequency coefficients. Quantizer Q divides each frequency coefficient by a value which is proportional to a quantization parameter QP and rounds the result of this division to the nearest integer. Consequently, quantizer Q provides a block of quantized frequency coefficients. The inverse quantizer IQ and the inverse discrete cosine transformer ICDT provide a decoded version of the differential macroblock based on the block of quantized frequency coefficients. The ADD adder adds the similar macroblock in the previous image, which has found the MEC motion estimator and compensator, to the decoded version of the differential macroblock. Accordingly, the adder ADD provides a decoded version of the macroblock that is being encoded, a decoded version which is stored in the MEM memory. The MEM memory in this way will comprise a decoded version of the image that is currently being decoded once all the macroblocks included in this image have been encoded. Accordingly, this is achieved even when an image is encoded, a decoded version of the preceding image contained in the MEM memory. The variable length encoder VLC transforms the block of quantized frequency coefficients into a series of variable length code. The block of quantized frequency coefficients will generally comprise a number of frequency coefficients having the value of zero, the number is greater than the quantization parameter QP has a higher value. As a result, the series of variable length codes comprises relatively few bits if the quantization parameter QP has a higher value, and vice versa. The variable length codes are temporarily stored in the OBUF output buffer. The OBUF output buffer produces variable length codes belonging to the macroblocks that have already been encoded, at the bit rate R. The controller CON controls the quantization parameter QP based on the following criteria. The amount of encoded data contained in the output buffer OBUF, which quantity will later be referred to as the fullness of the output buffer F, will be between a minimum and a maximum level. This implies that, if the OBUF output buffer is relatively full, the quantization parameter QP will have a relatively high value. If the OBUF output buffer is relatively empty, the quantization parameter QP will generally have a relatively low value. In more detail, the controller CON calculates an objective amount of encoded data for an image to be encoded, or a portion thereof. The target amount of encoded data depends on the fullness of the output buffer F. That is, if the OBUF output buffer is relatively full, the target will be relatively low, and vice versa. The controller CON controls a value for the quantization parameter QP based on the target amount of the encoded data and recent coding experiences. Those recent coding experiences include a quantity of data comprised of a recently encoded portion of the image, and the value of the quantization parameter QP that was applied to the encoding of that portion. Assume that the newly coded portion of the image comprises a relatively large amount of coded data. The controller NO deduces from this that the quantization parameter QP must have a value which is somewhat less than the recent value, to prevent the target quantity of coded data from being exceeded. The video encoder illustrated in Figure 2 can distort an image to a relatively large degree if the image, or a portion thereof, is encoded with the quantization parameter QP having a relatively high value. If the quantization parameter QP has a relatively high value, the number of different values that a relatively low quantized frequency coefficient can have. In addition, in practice, many of the frequency coefficients will be rounded to a value of zero. As a result, at the end of the coding, the image will be distorted to a relatively large degree. For example, the objects, or portions thereof, which comprise relatively many details will be presented by means of relatively large, uniform blocks. This phenomenon will be referred to here later as block effects. Block effects can occur, for example, when there is a scene change in the image sequence. In general, the first image of a new scene will be different from the last scene of the preceding image in a relatively large degree. Accordingly, if the first image of the new scene is coded, many blocks of frequency coefficients will comprise many frequency coefficients having a substantial value. If those blocks were quantized with the quantization parameter QP having a nominal value, a relatively large amount of coded data would be obtained. To prevent this, the controller CON will give the quantization parameter QP a high value. As explained here above, this can cause block effects. Figure 3 illustrates an example of a video coding arrangement according to the invention. The video coding arrangement recaptures the features illustrated in Figure 1. It provides an MPEG MDS data stream in response to a PDS image data stream representing a sequence of images. The video coding array comprises a PENC video pseudocoder, an FCON filter controller, a DEL delay circuit, and an adjustable FIL filter and an ENC video encoder. The ENC video encoder and the PENC video pseudo-encoder are similar to the video encoder illustrated in Figure 2 and discussed here above. However, the controller CON included in the video pseudo-encoder does not control the quantization parameter QP in the same manner as described above with reference to Figure 2. This will be explained in more detail here below.

The video coding arrangement illustrated in Figure 3 operates basically as follows. The pseudo video encoder PENC encodes an image to derive an indication of complexity Cl indicating, how difficult it would be to encode the image. The FCON filter controller controls the FIL adjustable filter depending on the Cl complexity indication. The DEL delay circuit compensates for the time it takes for the PENC video encoder to encode the image as well as the time it takes for the FCON filter controller to control FIL adjustable filter. The adjustable FIL filter filters the image to be encoded with a cutoff frequency FC determined by the complexity indication Cl. The ENC video encoder encodes the filtered image. There are several ways in which the PENC video pseudo-encoder can provide the complexity indication Cl. For example, the pseudo video encoder can encode the image with the quantization parameter QP having a certain fixed value. If the amount of encoded data is relatively large, it can be said that the image is difficult to encode, and vice versa. In this way, the amount of encoded data obtained by the coding of the image can constitute the indication of complexity Cl. The PENC video encoder may also encode the image with the quantization parameter QP that is being controlled based on a quantity of objective encoded data set in the ENC encoder. If the quantization parameter QP is given a relatively high value, to fill the target, it can be said that the image is difficult to encode as and vice versa. In this way, the quantization parameter QP can constitute the complexity indication Cl. It is also possible to use the product of the quantization parameter and the amount of encoded data obtained as an indication of complexity. In that case, it does not matter how the pseudo video encoder PENC sets a value for the quantization parameter QP. The complexity indication Cl is, in effect, an indication of distortion. If the complexity indication Cl has a high value, which means that the image is difficult to code, the following may occur. The ENC video encoder can give the quantization parameter QP a relatively high value, to avoid producing a relatively large amount of coded data. In that case, the block effects may occur in the image at the end of the decoding, as explained here above. The ENC video encoder can also give the quantization parameter QP a value which is sufficiently low to avoid such block effects. However, in that case, a relatively large amount of coded data will be produced, which will affect the amount of one or more subsequent images as explained here above. Thus, if the complexity indication Cl has a high value, the image itself will be distorted, with one or more of the subsequent images or both, if the ENC encoder were to encode the image as such. Figure 4 illustrates a filter control characteristic which can be applied by the FCON filter controller. Figure 4 is a graph having a horizontal axis representing the complexity indication Cl and a vertical axis representing the cutoff frequency FC of the adjustable filter FIL. The graph illustrates that, if the complexity indication Cl is below a threshold value Cth, the cutoff frequency has a relatively high value that does not vary substantially as a function of complexity Cl. Consequently, if an image is relatively easy to encode, it will be filtered to a relatively small degree only and, consequently the details of the image will be preserved substantially. However, if the complexity indication Cl is above the threshold value, the cutoff frequency decreases as the complexity indication Cl increases in value. Consequently, if an image is relatively difficult to encode, it will be filtered to a degree that depends on how hard it is to encode the image. That is, if an image is relatively difficult to encode, the details of the image will be removed to a degree that depends on how hard it is to encode the image. This allows the ENC video encoder to encode the image so filtered with the quantization parameter QP having a value which is sufficiently low to avoid block effects in the image itself, while the amount of encoded data obtained is sufficiently low to allow satisfactory quality of subsequent images. Figure 5 illustrates another filter control feature which can be optionally applied in combination with the control feature illustrated in Figure 4. Figure 5 is a graph having a horizontal axis representing an image number and a vertical number which represents the FC cutoff frequency. The image number N represents the image for which the indication of complexity would be established as described here above. It is assumed that the Cl complexity indication of this image has a relatively high value, which means that the image is difficult to encode. As a result, the spatial filter FIL has a relatively low cutoff frequency FC when the image number N is filtered. Figure 5 illustrates that, as a consequence, subsequent images having numbers N + 1, N + 2, N + 3, are also filtered with the cutoff frequency that increases gradually as the image number increases. The filter control feature illustrated in Figure 5 is advantageous, for example, when a scene change occurs. Assume that the image number N is the first image of a new scene. Since the first image of the new scene is filtered with the relatively low cutoff frequency FC, relatively many details will be removed. That is to say, that the end of the decoding, the first image of the new scene will have a relatively low resolution. Since the cutoff frequency FC gradually increases with the image number in the new scene, the resolution will gradually become better with each subsequent image. This evolution in resolution equals the properties of human vision. Human beings tend first to pay attention to the basic shape and color of the new objects and only later pay attention to the details of the new objects. Thus, in the case of a change of state, the characteristic of the filter control illustrated in Figure 5 will affect the perceived resolution only to a relatively small degree. What is more, the control feature allows the first images of a new scene to be coded with the quantization parameter QP having a value which is high enough to avoid block effects, without producing a relatively large amount of coded data. In this way, the control characteristic illustrated in Figure 5 contributes to a satisfactory total cavity. The drawings and their description above illustrate rather than limit the invention. It will be evident that there are numerous alternatives that will fall within the scope of the appended claims. In this regard, the following observations are made. There are numerous ways to physically diffuse functions or functional elements over several units. In this regard, the drawings are very schematic, each one representing only one possible embodiment of the invention. In this way, although a drawing shows different functional elements as different blocks, its meaning does not exclude that some functional elements, or all the functional elements, can be implemented as a single physical unit. Although, the video coding arrangement illustrated in Figure 3 uses a PENC pseudo-encoder to provide a distortion indication in the form of Cl complexity indication, other circuits may be used to provide a distortion indication. For example, a circuit can be used, which establishes a sum of absolute differences between the pixels of an image to be encoded and the corresponding pixels of a previous image. If the sum of the absolute differences is high, the image is difficult to code and, consequently, the image itself may be distorted or one or more of the subsequent images, or both. In this way, the sum of the absolute differences is an indication of distortion. Although in the video coding arrangement illustrated in Figure 3, the FIL adjustable filter filters images and is controlled on a picture-by-picture basis, it can also filter portions of images and be adjusted on a portion-by-portion basis. . With respect to the adjustable FIL filter shown in Figure 5, the following was also noted. In principle, any type of filter can be used. For example, the adjustable FIL filter can be a spatial filter or a horizontal filter. That is, it can be a two-dimensional filter or a one-dimensional filter. The FIL adjustable filter can be adjusted, for example, as a medium filter, a gauss filter or a medium filter. The adjustable filter FIL can also be implemented as a domain filter of a wave train. The domain filter of a wave train can be based on the unlock algorithm for JPEG compressed images. In that case, the filter performs edge detection and uniformity to preserve the edges detected in the wavelet domain: Edge detection may be based on a threshold that is determined in the PENC video pseudocoder shown in the Figure 3. Consequently, the threshold adapts to the complexity of the current image to be encoded. The waveform domain filter counteracts block effects by filtering only those areas of an image in which these effects may occur. In this way, a filter in the wavelength domain will reduce the resolution of an image to a lesser degree than a low pass filter. When applied, preferably filtering the wave train domain frame by frame, or better field by field, a JPEG algorithm may also be used to counteract block effects. Such an algorithm provides a satisfactory and relatively simple unlocking. In view of the observations made here above it will be clear that the term "filter" is not limited to signal processing characterized by response to a certain frequency. The term "filter" should be broadly construed to include several types of preprocessing which will decrease the distortion that would otherwise be introduced by a processor.

Any sign of reference in the claims should not be a limitation of the invention in question.

Claims (7)

CHAPTER CLAIMING Having described the invention, it is considered as a novelty and, therefore, what is claimed is contained in the following CLAIMS:

1. A method for processing a signal (S) by means of a signal processor (PRC), the method is characterized in that it comprises the steps of: - examining (EXAM) the signal (S) to be processed to obtain a distortion indication ( IND) which indicates in what degree the distortion (DIST) would be introduced if the signal (S) were processed (PRC [SJ) by means of the processor (PRC); - preprocessing (FIL) the signal (S) depending on the and indication of distortion (IND) to obtain a preprocessed signal (SF); and - processing (PRC [SFJ] the preprocessed signal (SF) by means of the processor (PRC). An array for processing signals, comprising a signal processor (PRC), characterized in that the array for processing signals further comprises: - an examination circuit (EXAM) for examining a signal (S) to be processed to obtain an indication of distortion (IND) that indicates in what degree the distortion (DIST) would be introduced if the signal (S) were processed (PRC [S]) by means of the processor (PRC); - a preprocessor (FIL) to preprocess the signal (S) depending on the distortion indication (IND) to obtain a preprocessed signal (SF), the signal processor (PRC) is coupled to the preprocessor (FIL) to process (PRC [ SF]) the preprocessed signal (SF): The arrangement for processing signals according to claim 2, characterized in that the preprocessor (FIL) comprises an adjustable filter whose filter characteristics are adjusted depending on the distortion indication (IND) . 4. A method for encoding video, characterized in that it comprises the steps of: - examining (PENC) video data to be encoded to obtain an indication of complexity (Cl) which indicates how difficult it is to encode the video data; - filtering (FIL) the video data depending on the indication of complexity (Cl) to obtain filtered video data; and - encode (ENC) the filtered video data. An arrangement for encoding video, characterized in that it comprises: - an examination circuit (FIL) for examining video data to be encoded to obtain an indication of complexity (Cl) which indicates how difficult it is to encode the video data; - an adjustable filter (FIL) to filter the video data depending on the indication of complexity (Cl) to obtain filtered video data; and - a video encoder (ENC) for encoding the filtered video data. 6. A method for encoding video, characterized in that it comprises the steps of: - examining (PENC) video data to be encoded to obtain an indication of complexity (Cl) indicating how difficult it is to encode the video data; - filtering (FIL) the video data depending on the indication of complexity (Cl) to obtain filtered video data; and - encoding (IBUF, SUB, DCT, Q, IQ, IDCT, ADD, MEM, MEC, VLC) the filtered video data depending on the quantization parameter (QP) to obtain coded video data; - write coded video data in a buffer (OBUF) and read the encoded video of the buffer (03UF) at a certain bit rate (R); and controlling (WITH) the quantization parameter (QP) based on the amount of encoded video data (F) contained in the buffer memory (OBUF). An arrangement for encoding video, characterized in that it comprises the steps of: - an examination circuit (FIL) for examining video data to be encoded to obtain an indication of complexity (Cl) which indicates how difficult it is to encode the data of video; - an adjustable filter (FIL) for filtering the video data depending on the indication of complexity (Cl) to obtain filtered video data; - a coding circuit (IBUF, SUB, DCT, Q, IQ, IDCT, ADD, MEM, MEC, VLC) for encoding the filtered video data depending on the quantization parameter (QP) to obtain coded video data; a buffer (OBUF) for temporarily storing the encoded video data and for sending the encoded video data at a certain bit rate (R); and - a controller (CON) for controlling the quantization parameter (QP) based on the amount of encoded video data (F) contained in the buffer (OBUF). SUMMARY OF THE INVENTION An array that processes signals comprising an examining circuit (EXAM), and an adjustable filter (FIL), and a signal processor (PRC). The signal processor (PRC) can be, for example, a video encoder for encoding a sequence of images according to an MPEG standard. The examining circuit (EXAM) examines a signal (S) to be processed to obtain an indication of distortion (IND), the indication of distortion (IND) indicates to what degree the distortion (DIST) would be introduced if the signal (S) were processed (PRC [S]) by means of the processor (PRC). An adjustable filter (FIL) filters the signal (S) depending on the indication of the distortion (DIST) to obtain a filtered signal (SF). The signal processor (PRC) processes the filtered signal (SF). In this way, the signal processor array filters the signal to be proactively processed to account for the distortion that would otherwise be introduced by signal processing. As a result, a satisfactory signal quality can be obtained. For example, in a video encoding application, the filter can reduce the details, contained in a series of images. This allows the series of images to be encoded with sufficient precision without introducing block effects that could otherwise occur if the images were not filtered. In addition, since the filter is proactively adjusted, it filters the series of images more evenly than if the filter were retroactively adjusted. As a result, there will be relatively little variation in resolution from one image to another, which contributes to a satisfactory overall image quality.