Classifications

• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
  • H04N7/00—Television systems
  • H04N7/01—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level
  • H04N7/0135—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level involving interpolation processes
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
  • H04N7/00—Television systems
  • H04N7/01—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level
  • H04N7/0117—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level involving conversion of the spatial resolution of the incoming video signal
  • H04N7/012—Conversion between an interlaced and a progressive signal
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
  • H04N7/00—Television systems
  • H04N7/01—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level
  • H04N7/0125—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level one of the standards being a high definition standard
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
  • H04N9/00—Details of colour television systems
  • H04N9/64—Circuits for processing colour signals
• • G—PHYSICS
  • G06—COMPUTING; CALCULATING OR COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T2207/00—Indexing scheme for image analysis or image enhancement
  • G06T2207/20—Special algorithmic details
  • G06T2207/20024—Filtering details
  • G06T2207/20032—Median filtering

Definitions

• the invention relates to a signal processing system for video signals.
• a transcoder was proposed as an example of such a signal processing system, which uses a video signal with e.g. B. 1250 lines, an image change frequency of z. B. 50 Hz and interlace method in a video signal with the same number of lines, a doubled image change frequency and interlace method (interlace) transcoded to eliminate large area and interlaced flicker.
• This transcoding can be carried out in a motion-adaptive manner.
• Adaptive to motion means that pixels are calculated differently with static image content than with dynamic image content.
• Transcoding means that a television signal is converted from one standard to another standard.
• the luminance pixels of reproduced fields can be calculated from received pixels using a two-out-of-six median filter, as suggested in P 38 03 605.
• a median filter sorts six pixels each according to the size of their numerical value and forms the arithmetic mean of the two pixels sorted in the middle of the ranking.
• the chrominance pixels of reproduced fields are formed by time averaging of received chrominance pixels which have the same spatial position in two neighboring received fields.
• double the line frequency is also required.
• 100 Hz frame rate and interlace method results in a line frequency of 62.5 KHz. It is planned to manufacture television sets with such a line frequency.
• a vertical-temporal alias ie a moving disturbance pattern, can arise from the image reproduction in the intermediate line method even at a frame rate of 100 Hz, in particular with dynamic image contents with high vertical frequencies.
• the invention is based on the object of specifying a signal processing system for video signals which, at an increased frame rate, avoids interference patterns caused by vertical-temporal aliases.
• a line frequency of 93.8 KHz for playback without interlacing and of 46.9 KHz for playback with interlacing is required compared to 125 KHz or 62.5 KHz. If you take into account that when playing of a video signal without interlacing is only about 70% of the number of lines of a video signal with interlacing required to obtain the same vertical resolution, the number of lines can be z. B. 1250 can be reduced to 900 without loss of vertical resolution. In the case of a video signal with 900 lines, 75 Hz frame rate and playback without interlacing, this leads to a line frequency of 67.5 KHz.
• This value almost corresponds to the value of 62.5 KHz, which is required for the function of the planned television sets with 1250 lines and 100 Hz frame rate. Because playback without interlacing vertical-temporal alias and interlacing. Particularly with dynamic picture content, greatly reduced, the reproduction of a television signal with 900 lines without interlacing and 75 Hz frame rate with comparable effort for the deflection circuit leads to better picture quality than that of a television signal with 1250 lines with interlacing and 100 Hz image change reguenz.
• the solution according to the invention consists in that a signal processing system generates four successive progressive output video images from three input video images which follow one another in the temporal direction.
• This type of signal processing also called transcoding, can be carried out in a motion-adaptive manner.
• input uminance or chrominance pixels are repeated in the corresponding spatial and temporal position for the output pixels.
• intermediate values are first generated in the middle of the time between the corresponding input pixels for the output luminance pixels to be calculated in images with a position lying temporally between the input images. From these intermediate values and Pixels of temporally adjacent input images are then proportionally composed of the pixels for the output signal with 1.5 times the image change frequency in accordance with the inversely proportional time intervals. For the calculation of the output chrominance pixels, the generation of intermediate values in the middle between the input images can be omitted. The chrominance pixels are then composed of parts of pixels of temporally adjacent input images in accordance with the inversely proportional time intervals.
• Every fourth output image contains corresponding pixels of every third input image.
• the television signal can first be motion-adaptively as proposed in P 38 03 835 into a television signal with 900 lines without Interlace with 50 Hz frame rate can be converted.
• a television signal with 900 lines can be generated from this television signal in a motion-adaptive manner without interlacing with a 75 Hz frame rate.
• a transcoder For signal processing of the type according to the invention, e.g. a transcoder.
• FIG. 1 block diagram of a television receiver with a transcoder
• Fig. 3 block diagram of a transcoder for digital video signals
• FIG. 5 block diagram of a transcoder according to the invention
• Fig. 6 is a timing diagram of the transcoder according to the invention.
• FIG. 1 shows a television tuner 10, an IF amplifier 11 and a demodulator 12 which provides an audio signal, a luminance signal and a chrominance signal. These signals are e.g. B. transmitted sequentially.
• the audio signal is processed in circuit 161, amplified in circuit 162 and fed to a loudspeaker 163.
• the luminance signal is digitized in the A / D converter 131 and the chrominance signal in the A / D converter 132. Both are passed on to a transcoder 14.
• the transcoder converts an input signal of e.g. B. 1250 lines with line jump and 50 Hz picture change frequency in an output signal of 1250 lines with line jump and 100 Hz picture change frequency.
• the transcoder converts an input signal from e.g. B. 1250 lines with a line jump and 50 Hz frame rate according to P 38 03 835 motion adaptive to a television signal with 900 lines without interlace and 50 Hz frame rate. This television signal is then motion-adaptively transcoded into an output signal of 900 lines without interlacing with a 75 Hz frame rate.
• the luminance and chrominance output signals of the transcoder 14 are supplied to a matrix 17 via the D / A converters 151 and 152, where they are converted into RGB signals and displayed on the display 19 via the RGB amplifier 18.
• Fig. 2 shows a method proposed in P 38 31 524, as from pixels of a television signal with interlace and z.
• pixels 211, 232, 251 of fields of an incoming television signal and pixels 211, 222, 231, 242, 251 of fields of a generated television signal with double image reproduction frequency over time 20, in the horizontal or line direction 201 and shown in the course in the vertical direction 202.
• the pixels 222, 231, 242 and the other pixels of corresponding lines in the respective field are not present in the incoming signal and must therefore be interpolated. This happens in a motion-adaptive way.
• the luminance or chrominance pixels 222 become receiving luminance or chrominance pixels 232, for the luminance or chrominance pixels 231 the received luminance or chrominance pixels 211 and for the luminance or chrominance pixels 242 the received luminance or Chrominance pixels 232 are used. This applies accordingly to the other pixels of the respective fields to be interpolated.
• the pixels 222 are then interpolated from pixels 212 and 232 and the pixels 242 from pixels 232 and 252.
• the luminance pixel 223 is calculated from the luminance pixels 213 and 233.
• the six pixels 213 and 233 are sorted according to the size of their numerical value and the two sorted in the middle of the ranking are selected. The arithmetic mean is formed from them.
• This mean value is then reproduced as pixel 223 and together with the corresponding other pixels as a second field.
• the pixels 242 are reproduced as a fourth field.
• the first received field is reproduced as a first field and the third received field is reproduced as a fifth.
• the interpolated field which consists of the pixels 231 and corresponding ones, is reproduced as a third field.
• FIG. 3 shows a transcoder as proposed in P 38 31 524, which uses the method according to FIG. 2.
• Two incoming fields are stored in the form of one frame in the memories 311 and 312. While the fields arriving in a memory are storing, the fields stored and combined into full images are fed pixel by pixel to a changeover switch 321 and from there to another changeover switch 301 and line memories 331 connected in parallel.
• a further changeover switch 341 conducts image points in parallel from the line memories 331 to the vertical filters 351 connected in parallel.
• B. four, vertically adjacent pixels of a field are calculated in the middle of the four pixels in each field an ⁇ pixel position. Such pixels are shown in Fig. 2 as 212, 231 and 252.
• a motion detector 30 continuously decides for each pixel whether pixels received from static switch 321 for static image content or pixels received from the vertical filters 351 or interpolated in the field from the vertical filters 351 to a memory 361 and to a median value generator 371 can be forwarded.
• a memory 361 At the input and output of the frame memory 361 are base values (213 and 233 in FIG. 2) for the median value generator . tapped.
• the changeover switch 391 either pixels (211, 231 or 251 in FIG. 2) from the memory 361 or pixels delayed by a field memory 381 (222 or 242 in FIG. 2) from the median value generator 371 forwarded to the exit.
• the median value generator 371 is supplied with six pixels (213 and 233 in FIG. 2) for each pixel (223 in FIG. 2) that is output.
• the image point (223 in FIG. 2) is calculated as described for FIG. 2.
• the pixel (223 in FIG. 2) is controlled by the motion detector 30 only from the arithmetic mean of the middle of the three pixels (213 in FIG. 2) and the middle of the other three pixels (233 in Fig. 2) formed.
• the other branch of the transcoder for the chrominance works.
• the circuits 313 and 314 correspond to the circuits 311 and 312, 322 corresponds to 321, 332 corresponds to 331, 342 corresponds to 341, 352 corresponds to 351, 302 corresponds to 301, 362 corresponds to 361, 372 corresponds to 371, 382 corresponds to 381, 392 ⁇ speaks 391.
• the parallel vertical filter 352 can e.g. .B. perform an averaging. Correspondingly fewer parallel vertical filters 352 and line memories 332 are required.
• the circuit 372 works for static and dynamic picture parts like the circuit 371 for static picture parts, i. H. it only forms an arithmetic mean. To reduce the storage space required, only active pixels are saved.
• FIG. 4 shows a method of how from luminance or chrominance pixels of a television signal without interlacing and z. B. 50 Hz frame rate, a television signal can be formed without a line jump with 1.5 times the frame rate (75 Hz).
• the pixels 461, 462, 471, 472 and the other pixels of corresponding lines in the respective full image are not present in the incoming signal and must therefore be interpolated. This can be done in a motion-adaptive manner.
• an intermediate value in the temporal position 423 or 443 is first determined. This happens e.g. B. by a median, as described for Fig. 2.
• the z. B. six pixels 413 and 433 sorted according to the size of their numerical value and the two sorted in the middle of the ranking selected. The arithmetic mean is formed from them. This mean value gives the numerical value for the intermediate value 423.
• the numerical values of 413 are e.g. B. (127, 132, 125) and the numerical values - of 433 are e.g. B. (130, 195, 220).
• the sorted ranking is then (220, 195, 132, 130, 127, 125) and the arithmetic mean of (132, 130) is formed.
• the intermediate value then receives the value 131.
• the six pixels 413 and 433, the pixels adjacent to the left and right of 413 and 433 in the line can also be added, so that z. B. ten pixels can be used.
• the intermediate value 423 can be calculated from the arithmetic mean of the middle of the three pixels 413 and the middle of the three pixels 433, or the middle of the pixels 413 or 433 can be used directly for pixels 463 or 473.
• the corresponding arithmetic mean can always be formed from pixels of the same color component as for the luminance signal with static picture content.
• the median value can be formed in accordance with the median value formation for the corresponding luminance signals, but with different input pixels. These are for the pixel 423 z. B. the middle picture elements of 413 and 433 and the picture elements of 413 and 433 adjacent to the left and right in the line of the same color component.
• the U component of a chrominance pixel is formed from two of six U components of received chrominance pixels or the V component of a chrominance pixel is formed from two of six V components of received chrominance pixels.
• the pixel 463 or 473 of the luminance or chrominance signal can then, for. B. can be calculated as follows: The single value of the central pixel of 433 and the • double value of the intermediate value 423 or 443 are added and the sum divided by three. The result is the value for pixels 463 or 473.
• the other pixels 461, 462 or 471, 472 of the outgoing luminance or chrominance signal are calculated accordingly.
• the pixels 461, 462 or 471, 472 of the luminance or chrominance signal can be obtained as follows:
• the pixels 411, 412 are used for the pixels 461, 462 and for the pixels 471 , 472, the pixels 431, 432 are used.
• a first received frame 411, 412 is reproduced as a * first frame 411, 412, a frame interpolated from the first 411, 412 and second 431, 432 received as a second frame 461, 462 is reproduced as a third frame 471, 472, a full image received from the second 431, 432 and third 451, 452 is inter- reproduced polished frame and as a fourth frame 451, 452 a third received frame 451, 452 is reproduced.
• the transcoder 50 shows a transcoder, to which a transcoder 50 according to P 38 03 835 is connected upstream at the luminance input.
• the trans coder 50 generates from a television signal with e.g. B. 1250 lines with skip and 50 Hz frame rate adaptive to motion a television signal with 900 lines without skip and 50 Hz frame rate.
• the pixel-wise motion detector signal from the transcoder 50 is also used in the transcoder 51 for chrominance signals and for the circuits 521 and 522.
• the transcoder 51 can be constructed like the transcoder 50. However, it does not need its own motion detector. A full luminance image is stored in the memory 511.
• the frame stored in 511 is supplied to the changeover switch 531 (411, 412 in FIG. 4).
• the incoming luminance signal (431, 432 in FIG. 4) and the previously stored signal (411, 412 in FIG. 4) from 511 and the motion detector signal are fed to a median value generator 521.
• This median value generator 521 operates as described for FIG. 4.
• With dynamic image content z. For example, six pixels (413 and 433 in FIG. 4) two are selected and their arithmetic mean value is fed to the changeover switch 531 and, in the case of static image content, the arithmetic mean value of two image points (the mean of 413 and 433 in FIG.
• the pixels coming from 521 representing a second full image to be forwarded are then read out of 511 and form a third full image to be forwarded.
• the stored pixels (431 and 432 in FIG. 4) from 511 and incoming pixels to be stored in 511 (451 and 452 in FIG. 4) are fed to the changeover switch 531 via the median value generator 521 and form a fourth relay. of the full screen.
• This process that is to say alternately reading out 511 and reading out 511 and incoming pixels, combined with formation of median values, is continued continuously. This produces a television signal at the output of the changeover switch 531 without interlacing with a doubled frame rate.
• This signal is fed to an adder 561 via a frame memory 541 and multiplier 551 and via a frame memory 542 and a multiplier 552.
• the memories 541 and 542 are loaded alternately with the incoming signal at a frame rate of 100 Hz and discharged at an image frequency of 75 Hz for an outgoing signal.
• the memories 541 and 542 can be of the "FIFO readable multiple" type. FIFO means: first in, first out.
• blocks 61 are shown which contain the read-in memory content, e.g. B. pixels “(412)” from Fig. 4, in the respective memory, for. B. "541" of Fig. 5 show.
• blocks 62 are shown which contain the read-out memory content, e.g. B. pixels “(412)” from FIG. 4, the respective memory, e.g. B. "541 + 542" from FIG. 5, as well as assignment arrows 63, which indicate from which blocks 61 the respective pixels come, and multiplication factors 64, which show the respective weighting factor for the respective blocks 61 of FIG Specify pixels.
• the multiplier 541 changes e.g. B. for each read out frame between the factors 1 and 1/3 and the multiplier ⁇ 542 z. B. between the factors 0 and 2/3, as can be seen from Fig. 6.
• the full image is read again for 33 ms.
• Multiplier 541 operates e.g. B. with the factors 1 and 1/3, the multiplier 542 z. B. with the factors 0 and 2/3, so that the summed factors give 1 each.
• the other branch of the transcoder for chrominance operates in accordance with the branch of the transcoder described in FIG. 5 for the luminance.
• the memory 512 corresponds to memory 511
• median value generator 522 corresponds, as described for FIG. 4, median value generator 521
• changeover switch 532 corresponds to changeover switch 531
• memory 543 and multiplier 553 correspond to memory 541 and multiplier 551
• memory 544 and multiplier 554 correspond to memory 542 and multiplier 552
• adder 562 corresponds to adder 561.
• the transcoder according to FIG. 5 can also add an interlaced picture signal by attaching a further FIFO memory. testify by the fact that at the output only every second line is reproduced with a corresponding field position and halved readout clock.
• the transcoder according to FIG. 5 can also have an input signal without interlacing, e.g. B. 900 lines, 50 Hz, process if the transcoders 50 and 51 in FIG. 5 are omitted and an additional motion detector is arranged, as it is e.g. B. is proposed in P 38 09 249.
• an input signal without interlacing e.g. B. 900 lines, 50 Hz
• Such a transcoder can also image signals with line jump, the same number of lines and z. B. 50 Hz image change frequency as input signal if the input signal with interlacing is converted into an input signal without interlacing.
• the motion detector can then be used together for this motion-adaptive conversion and for controlling the median value formers 521 and 522 from FIG. 5.

Landscapes

• Engineering & Computer Science (AREA)
• Multimedia (AREA)
• Signal Processing (AREA)
• Computer Graphics (AREA)
• Television Systems (AREA)
• Color Television Systems (AREA)

Applications Claiming Priority (2)

Publications (1)

Family

ID=6379969

Family Applications (1)

Country Status (10)

Families Citing this family (3)

Family Cites Families (5)

• 1989
  • 1989-05-03 DE DE3914550A patent/DE3914550A1/de not_active Withdrawn
• 1990
  • 1990-04-28 CA CA002049988A patent/CA2049988A1/en not_active Abandoned
  • 1990-04-28 AU AU55474/90A patent/AU5547490A/en not_active Abandoned
  • 1990-04-28 WO PCT/EP1990/000689 patent/WO1990013969A1/de not_active Application Discontinuation
  • 1990-04-28 EP EP90906943A patent/EP0470984A1/de not_active Withdrawn
  • 1990-04-28 HU HU903341A patent/HUT60080A/hu unknown
  • 1990-04-28 JP JP2506573A patent/JPH04505079A/ja active Pending
  • 1990-04-28 KR KR1019910701519A patent/KR920702147A/ko not_active Application Discontinuation
  • 1990-05-02 DD DD90340312A patent/DD298868A5/de not_active IP Right Cessation
• 1991
  • 1991-11-01 FI FI915177A patent/FI915177A0/fi unknown

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events