DE102004049106B4 - Method and device for measuring and compensating signal distortions - Google Patents

Abstract

Method for measuring and compensating signal distortions caused by non-constant group delay along a transmission path, in particular a video transmission path, in particular teletext signal distortions,
characterized in that
An input signal, in particular in video signal (VIDEO_IN), with a reference signal (P _ref ) in a predetermined number (n, m) of series-connected stages (tap1, tap2, tap3, .... tapn; tap ^- 1, tap ^- 2, tap ^- 3, ... tap ^- m) sequentially each time by a group delay _value (τ _g1 , τ _g2 , τ _g3 , ... τ _gn , τ ^- _g1 , τ ^- _g2 , τ ^- _g3 , ... .τ ^- _gm , τ _g ) of a predetermined size is compensated for
Each of the compensated reference signals (P _{ref, 1} , P _{, 2} ) obtained after each stage (tap1, tap2, tap3, ... tapn; tap ^- 1, tap ^- 2, tap ^- 3, ... tap ^- m), P _{ref, n;} P ^- _{ref, 1,} P ^- _{ref, 2,} P ^- _{ref, 3,} ... P ^- _{ref, m)} with the ideal course of the reference signal (P _ref) is compared,
- As a reference signal (P _ref ), a signal is used whose ideal course is symmetrical, in particular rectangular or triangular, and that
- the compensated signal as an output signal (COM PENS_VIDEO_OUT) is output which is the compensated reference signal (P _{ref, 1,} P _{ref, 2,} P _{ref, 3,} ..., P _{ref, n;} P ^- _{ref, 1,} P ^- _{ref, 2} , P ^- _{ref, 3} , ... P ^- _{ref, m} ), ...

Description

The The invention relates to a method and a corresponding device for measuring and compensating signal distortions according to the respective Preamble of the claims 1 or 14.

According to the state The technology is essentially two process principles for the measurement and Compensation of non-constant group delay of a video transmission link caused teletext signal distortions, namely algorithms according to the method of least squares for equalization of a video signal via adaptive non-recursive filters (FIR filters) and a measurement of the time shift of a reference pulse and compensation, as for example in the teletext chip SDA6000 the company Micronas is realized.

Examples of the equalization of a video signal by least squares algorithms are the patents US 5,793,821 and US 6,313,882 refer to. The use of least squares algorithms to equalize a video signal involves a significant amount of hardware for matrix calculations. In addition, transient phenomena of the system are disturbing visually perceptible.

at the measurement of the time shift of a reference pulse temporal shifts of a digitized pulse in the transmission channel due to group delay distortions measured and the video signal after detection of such temporal shift of the digitized pulse with a fixed prescribed characteristic of the group maturity corrected. This method is very inaccurate, as is common only then detects a time shift of a digitized pulse if the teletext signal is already so heavily distorted that subsequent error correction procedures are no longer effective.

As a further prior art, the DE 33 19 011 C2 , the DE 197 41 434 C2 and the DE 35 15 283 A1 called.

The The object of the invention is a comparatively simple but reliable Method and a corresponding device for measurement and compensation of signal distortions.

These The object is achieved by a method having the features of the patent claim 1 and a device with the features of claim 14 solved.

advantageous versions and further developments of the invention are specified in the subclaims.

The Invention is based on the idea, not the distortion of the signal premier Group time to measure directly, but first one Compensation of the input signal with a known group delay perform and by evaluating the quality of the received signal on the leading to distortion To conclude group runtime. The inventive method for measurement and compensation of non-constant group delay along a transmission path, in particular a video transmission link, caused signal distortions, in particular teletext signal distortions, is characterized in that in a first step an input signal, in particular a video signal, with a reference signal in one predetermined number of stages connected in series sequentially each compensated by a group delay value of predetermined size will, that in a second step, each of the received after each stage compensated reference signals with the ideal course of the reference signal compared that a signal is used as the reference signal, its ideal course symmetrical, in particular rectangular or is triangular, and that compensated for that in a third step Signal is output as an output signal, which compensates for this Reference signal includes, whose course the ideal course of the reference signal the next comes and has the least symmetry deviation.

With this method can already detected the slightest distortion in the input signal and with be compensated for different degree of compensation, since non-linear Group delay distortions not from a strongly disturbed signal have to be extracted because of the complexity the signal determination of a fault is difficult and hardly predictable is. Rather, here is just a selection of one to issue to be brought signal from a manageable Number of potential output signals are taken, the Select the right signal for a known waveform is returned. This is also the quality the output signal improved.

The corresponding device according to the invention comprises compensation means which compensate the input signal, in particular the aforesaid video signal, with the reference signal in the predetermined number of stages connected in series, each time by a group delay value of predetermined magnitude, comparison means which obtain each of the values obtained after each stage compensated reference signals with the ideal course of Compare reference signal, wherein the reference signal is a signal whose ideal course is symmetrical, in particular rectangular or triangular, and output means which outputs the compensated signal as an output signal comprising the compensated reference signal whose course comes closest to the ideal course of the reference signal ,

When Compensation means may in particular be a network with a predetermined Number of all-pass stages are used, which serially connected in series are and at their respective outputs the o.a. Signals can be tapped are.

Around to achieve the highest level of flexibility; can Means be provided, with which the number of active Allpass stages configurable is. For example, you can by means of a multiplexer measuring outputs at different positions selected by the Allpass network become.

According to the invention as Reference signal uses a signal whose ideal course is symmetrical, especially rectangular or triangular, is. The use of a Symmetric reference signal has the advantage of causing distortions Immediately reflected in symmetry deviations. As an output signal can then easily output the compensated signal which comprises the compensated reference signal whose course having the least symmetry deviation.

In particularly advantageous embodiment, in particular this variant is used according to the invention as a reference signal a defined pulse sequence in a teletext signal, in particular the first three consecutive data bits of the World Standard Teletext (WST) framing codes, used. The use of a defined, So known, pulse train has the advantage that an assignment of Comparison and reference signal is very easy, since both the temporal Expiration, as well as the amplitude profile are already known.

In Design of the latter variant, it has to be special Cheap when the pulse train representing the reference signal is within a measurement window, which consists of sync signals of an external Synchronization separation is derived. The exact start position of the pulse train is derived from a rising edge. This procedure presents sure the measurement window is actually the one for symmetry evaluation needed Pulse sequence has.

Farther is provided in a variant embodiment, that the number of stages in which compensation is carried out changed cyclically is determined by the number at the end of the respective teletext line the successive stages of the compensation network incrementally elevated becomes. This measure its purpose is to gradually adjust the degree of equalization of the input signal to increase.

Prefers it is envisaged that the number of stages in which a compensation is carried out, with reaching a maximum number to a predetermined number, in particular one, reset becomes.

The The invention is not limited to merely steps are provided with positive group delay compensation, rather sees the invention in a part of the stages to carry out a positive group delay compensation and in another part the stages the execution a negative group delay compensation.

Usually it is envisaged that in some of the stages, preferably in all Steps with the same group delay value is compensated. The usage of stages with identical group delay values allows one largely modular design.

The Evaluation of the generated potential output signals and the selection the signal intended for the output is achieved according to the invention in a simpler manner Way in that for each comparison obtained after each stage compensated reference signals with the ideal course of the undistorted Reference signal each one representing a degree of similarity Size determined and that from those representing a respective degree of similarity determined sizes determined which compensated reference signal has the course, the the ideal course of the non-distorted reference signal comes closest. From several successive measurement results of the described Procedure, an average is formed.

The respective averages can for example, in a multi-stage FIR method by means of appropriate FIR filters are determined. FIR procedures are, in particular also results from the introduction to the description, known per se and such FIR filters can be easily incorporated into a device according to the invention integrate.

In order to prevent that at low and possibly tolerable distortions alternately a compensated and a non-compensated signal is applied to the output, the invention provides that the compensated signal comprising the compensated reference signal whose course comes closest to the ideal course of the reference signal , only output as output signal when the group delay required to compensate the compensated signal for output exceeds a predetermined threshold.

This Threshold value can preferably be set from the outside to adjust the compensation behavior to the respective input signal can.

There the accuracy of the measurement with the noise in the input signal in principle decreases, is provided according to the invention in a preferred embodiment, that the threshold is increased with increasing signal noise.

main application the method according to the invention or the device according to the invention is in the equalization of teletext signals in a single-chip Video ASIC.

The The invention will now be described in more detail with reference to the drawings. Show it

1 a determination of a symmetry deviation in an axisymmetric pulse

2 An embodiment of a device according to the invention for measuring and compensating caused by non-constant group delay along a video transmission link teletext signal distortion

3 an identification sequence of a teletext line with a distorted framing code pulse (FC) with different compensation

• a) positiver Gruppenlaufzeit
• b) negativer Gruppenlaufzeit

4 a phase characteristic and an associated group delay as a function of the frequency in a single all-pass filter stage

• a) positive group maturity
• b) negative group maturity

This in 2 illustrated block diagram of a device according to the invention for measuring and compensation caused by non-constant group delay along a video transmission link teletext signal distortion shows as essential components of a compensation network 1 to an equalization of teletext signals and a measuring block 2 for evaluating a distortion of a reference signal occurring in a teletext data stream and for determining the optimal degree of compensation.

The compensation circuit 1 consists of a network of all-pass tap1, tap2, tap3, ... tapn; tap ^- 1, tap ^- 2, tap ^- 3 ... tap ^- m configurable length for positive group delay τ _1g , τ _g2 , ... τ _gn as well as for negative group delay τ ^- _g1 , τ ^- _g2 , ... τ ^- _gm . All allpass stages are constructed identically (τ _g1 = τ _g2 = τ _g3 ... = τ _gn = τ ^- _g1 = τ ^- _g2 = τ ^- _g3 = ... = τ ^- _gm = τ _g ) and have the same Transfer function G (z). These all-pass stages, like all all-passes, have frequency-dependent phase shift and group delay but a constant gain over the frequency. Due to the series connection, the group delay adds up to the number n + m of the set stages n, m.

die Konstante a die Charakteristik der Übertragungsfunktion bestimmt und vorzugsweise a = 2 gewählt ist. 4 shows the transfer function G (z) of a single stage for positive and negative group delay. The transfer function of an all-pass filter with a negative history of the group delay τ ^- _{g is} for m sequential stages

the constant a determines the characteristic of the transfer function and preferably a = 2 is selected.

For positive group delay compensation for n sequential stages, the transfer function calculates too

The compensation network 1 the device according to the 2 consists of a cascade of all-pass taps tap1, tap2, ... tapn, ... tap-n, tap ^- 1 ..., tap ^- m for positive and negative distortions. This compensation network 1 On the input side a video signal VIDEO-IN can be fed. There are tap options on the Allpass network 1 both for the measurement of a distortion var_compensated_video as well as for a simultaneously running in operation compensation of the input signal compens_video_out. For this purpose, each Allpassstufe tap1, tap2, tapn ...., ^{TAP1, TAP2} ... tap ^- m with the terminals COMPENS_VIDEO_OUT or VAR_COMPENS_VIDEO connectable outputs A _{M, 1, M A, 2,} A _{M, 3} ... A _{M, n} , A _{M, -1} , A _{M, -2} , ... A _{M, -m} , A _{K, 1} , A _{K, 2} , A _{K, n} ... A _{K, - m} up.

The measuring block 2 consists of a comparator 3 , the input side each one of the compensation network 1 generated compensated signals can be fed and a control device 4 which controls measurement and compensation process.

To determine the degree of distortion, a defined pulse sequence in the teletext signal is used as a reference. In the present embodiment, the first three bits of the World Standard Teletext Framing Code (Note: to identify the WST - world standard teletext - is a "framing code" consisting of 8 bits inserted in fixed position at the beginning of each line of data (ETS 300 706, EIA 516, BTA T001) While the first four bits are always defined as hexadecimal OEh (series of three consecutively transmitted data bits ), the last four bits are dependent on the used standard and country) are used as a reference These bits are used by the comparator 3 temporally interpreted as a pulse and the distortion of the pulse used as an indicator for the group delay. The distortion is caused by non-linear phase shift of the harmonics of the pulse (Fourier) and influences the symmetry of the pulse and the eye level of the transmitted data (eye diagram). The degree of asymmetry of the pulse shows no linear relationship to the degree of group delay distortions, but the measuring and compensation circuit checks the signal at different degrees of compensation for best symmetry (grd_camp) and thus determines the optimal compensation for the operation Kompensationungsgrad (compensation control ). As a measure of the symmetry deviation of the reference pulse, the difference dy of the pulse amplitudes on both sides of an assumed symmetry axis symmetry is used in the exemplary embodiment ( 1 ).

The distorted video signal VIDEO_IN is from the all-pass network 1 filtered with different group delay characteristics. The number n + m of the passed stages tap1, tap2, tap3 ..., tapn, tap ^- 1, tap ^- 2, tap ^- 3, ... tap ^- m of the network 1 determines the degree of compensation. In the measuring block 2 becomes the video signal to video line more or less equalized reference signal P _{ref, 1} , P _{ref, 2} , P _{ref, 3} , ... P _{ref, n} ... P _{ref, -1} , P _{ref, -2} , P _{ref , -3} , P _{ref, m} checked for symmetry by determining the difference dy. The degree of compensation with the best symmetry dy = dy _min is stored.

The measurement window in the video line is derived in the presented embodiment from synchronization signals of the external synchronization separation and from the rising edges (timing / measurement window). The length of the filter chain used for the measurement, ie the number of n + Allpassstufen tap1, tap2 ... tapn, ^{TAP1, TAP2,} tap m ^- 3, ... tap ^- m is at the end of each valid teletext line increases incrementally (meas_ctrl). The cascade length m + n varies cyclically and reaching a maximum number n _max + m _max of the possible all-pass stages tap1, tap2 ... tapn _max, ^{tap1, tap2} ... tap ^- m _max in compensation -Network 1 set back to one again.

Per Measuring cycles become n + m measurements with n positive and m negative steps carried out. There the number of teletext lines in the vertical blanking interval (VBI) and varies with different transmitters, the measurements are distributed to different teletext lines in the VBI.

The Measurement results are averaged with a k-stage FIR filter. exceeds the for the optimal compensation measured and averaged group delay (= number the allpass stages with best pulse symmetry or smallest distortion) an adjustable Threshold var. Threshold, the compensated input signal COMPENS_VIDEO switched to the outside instead of the uncompensated input signal VIDEO_IN (COMPENS_VIDEO_OUT). With the variable threshold var.threshold and an internal hysteresis prevents the compensation for small distortions constantly is activated and deactivated. As the accuracy of the measurement with the noise in the video signal inherently decreases, there is the Possibility to Threshold var.threshold from the outside with increasing signal noise to increase.

The transient response at station change is determined by the number of used allpass stages and the number of steps of the averaging and remains at n _max + m _max = 16 pos. + 16 neg steps (about 2 fields) and k = 8-step averaging with 320 ms still below the perceivable visually disturbing threshold.

main application The invention relates to a circuit for equalizing teletext signals in a single-chip video ASIC.

1

compensator

2

measuring device

3

comparator

4

control device

τ _g1

Group delay

τ _g2

Group delay

τ _g3

Group delay

,

,

,

,

,

,

τ _gn

Group delay

τ ^- _g1

negative Group delay

τ ^- _g2

negative Group delay

,

,

,

,

,

,

τ ^- _gm

negative Group delay

P _ref

reference pulse

P _{ref, 1}

by first allpass stage compensated reference pulse

,

,

,

,

,

,

P _{ref, n}

by nth allpass stage compensated reference pulse

n

number positive levels

m

number negative stages

n _max

maximum Number of positive stages

m _max

maximum Number of negative levels

k

number the FIR filter stages

tap1

first Allpass level z. Compensation positive group runtime. compensation positive group runtime

tap2

second Allpass level z. Compensation of positive group delay

tap3

third Allpass level z. Compensation of positive group delay

,

,

,

,

,

,

tapn

nth Allpass level z. Compensation of positive group delay

tap ^- 1

first Allpass level z. Compensation of negative group delay

tap ^- 2

second Allpass level z. Compensation of negative group delay

tap ^- 3

third Allpass level z. Compensation of negative group delay

,

,

,

,

,

,

tap ^- m

mth Allpass level z. Compensation of negative group delay

A _{M, 1}

measuring output the first all-pass level

A _{M, 2}

measuring output the second allpass level

A _{M, 3}

measuring output the third Allpass level

,

,

,

,

,

,

A _{M, n}

measuring output the nth Allpass level

A _{M, -1}

measuring output d. first allpass stage with negative group delay compensation

A _{M, -2}

measuring output d. second all-pass stage with negative group delay compensation

,

,

,

,

,

,

A _{M, -m}

measuring output d. m-th Allpass stage with negative group delay compensation

A _{K, 1}

Compensation output of first allpass level

A _{K, 2}

Compensation output of second allpass level

A _{K, 3}

Compensation output of third allpass level

,

,

,

,

,

,

A _{K, n}

Compensation output of nth Allpass level

A _{K, -1}

Compensation output of first allpass stage with negative group delay compensation

A _{K, -2}

Compensation output of second all-pass stage with negative group delay compensation

A _{K, -3}

Compensation output of third all-pass stage with negative group delay compensation

,

,

,

,

,

,

A _{K, -m}

Compensation output of m-th Allpass stage with negative group delay compensation

dy

amplitude difference

dy _min

minimum amplitude difference

VIDEO_IN

Video input

Symmetry

axis of symmetry

COMPENS_VIDEO

compensated video signal

var.thereshold

threshold

COMPENS_VIDEO_OUT

Video output

timing

control device

measurement window

control device

meas_ctrl

control device

grd_comp

control device

compensation_control

control device

Claims (27)

Method for measuring and compensating signal distortions caused by non-constant group delay along a transmission path, in particular a video transmission path, in particular teletext signal distortions, characterized in that - an input signal, in particular in video signal (VIDEO_IN), having a reference signal (P _ref ) in a given number (n, m) of series connected stages (tap1, tap2, tap3, .... tapn; tap ^- 1, tap ^- 2, tap ^- 3, ... tap ^- m) sequentially one at a time Group _{delay value} (τ _g1 , τ _g2 , τ _g3 , ... τ _gn , τ ^- _g1 , τ ^- _g2 , τ ^- _g3 , .... τ ^- _gm , τ _g ) is compensated predetermined size, that - each of the after each level (tap1, tap2, tap3, ... tapn; tap ^- 1, tap ^- 2, tap ^- 3, ... tap ^- m) of compensated reference signals (P _{ref, 1} , P _{, 2} , P _{ref, n} P ^- _{ref, 1} , P ^- _{ref, 2} , P ^- _{ref, 3} , ... P ^- _{ref, m} ) are compared with the ideal course of the reference signal (P _ref ) d, - as a reference signal (P _ref ), a signal is used whose ideal course is symmetrical, in particular rectangular or triangular, and that - the compensated signal is output as an output signal (COM-PENS_VIDEO_OUT), which the compensated reference signal (P _{ref, 1} , P _{ref, 2} , P _{ref, 3} , ... P _{ref, n} ; P ^- _{ref, 1} , P ^- _{ref, 2} , P ^- _{ref, 3} ,... P ^- _{ref, m} ) whose course is closest to the ideal course of the reference signal (P _ref ) and has the least symmetry deviation. Method according to Claim 1, characterized in that a defined pulse sequence (P _ref ) in a teletext signal, in particular the first three consecutive data bits of the World Standard Teletext (WST) framing code, is used as the reference signal. Method according to claim 2, characterized in that in that the pulse sequence representing the reference signal is within a Measuring window, which consists of synchronizing signals of an external synchronization separation is derived, recognized and the exact start position of the pulse train is derived from a rising edge. Method according to one of the preceding claims, characterized in that the number (n, m) of the stages (tap1, tap2, tap3, .... tapn; tap ^- 1, tap ^- 2, tap ^- 3), in which a compensation is performed is cyclically changed, in particular by means of a multiplexer measurement outputs are selected at different positions of a compensation network. Method according to claim 4, characterized in that the number (n, m) of the stages (tap1, tap2, tap3, .... tapn; tap ^- 1, tap ^- 2, tap ^- 3, ... tap ^- n) , in which a compensation is performed, with reaching a maximum number (n, m) to a predetermined number, in particular one, is reset. Method according to one of the preceding claims, characterized in that in a part of the stages (tap1, tap2, tap3, .... tapn) a positive group delay compensation (τ _g1 , τ _g2 , τ _tg3 , ... τ _gn ) and in a part of the stages (tap ^- 1, tap ^- 2, tap ^- 3, ... tap ^- m) a negative group _{delay compensation} (τ ^- _g1 , τ ^- _g2 , τ ^- _g3 , ... τ ^- _gm ) is performed. Method according to one of the preceding claims, characterized in that in a part of the stages, preferably in all stages (tap1, tap2, tap3, .... tapn; tap ^- 1, tap ^- 2, tap ^- 3, ... tap ^- m), with the same group delay value (τ _g ) is compensated. Method according to one of the preceding claims, characterized in that for each comparison, the values obtained after each stage (tap1, tap2, tap3, ... tapn; tap ^- 1, tap ^- 2, tap ^- 3, ... tap ^- m) compensated reference signal (P _{ref, 1,} P _{ref, 2,} P _{ref, 3,} ..., P _{ref, n,} P ^- _{ref, 1,} P ^- _{ref, 2,} P ^- _{ref, 3,} ... P ^- _{ref , m)} with the ideal curve of the non-distorted reference signal (P _ref) in each case a degree of similarity representing variable (dy _min) is determined and that from a respective degree of similarity representing the determined variables (dy _min) is determined which compensated reference signal (P _{ref, 1} , P _{ref, 2} , P _{ref, 3} ,... P _{ref, n} ; P ^- _{ref, 1} , P ^- _{ref, 2} , P ^- _{ref, 3} ,... P ^- _{ref, m} ) has the course which comes closest to the ideal course of the non-distorted reference signal (P _ref ), and that an average value is formed from a plurality of successive measurement results of the described method. Method according to claim 8, characterized in that that the respective averages are determined in a multistage FIR procedure become. Method according to one of the preceding claims, characterized in that the compensated signal containing the compensated reference signal (P _{ref, 1} , P _{ref, 2} , P _{ref, 3} , ... P _{ref, n} ; P ^- _{ref, 1} , P ^- _{ref, 2} , P ^- _{ref, 3} ,... P ^- _{ref, m} ), the course of which comes closest to the ideal course of the reference signal (P _ref ), is only output as an output signal (COMPEMS_VIDEO_OUT), if the Compensation of the compensation signal provided for output required group delay (τ _g1 , τ _g2 , τ _g3 , ... τ _gn , τ ^- _g1 , τ ^- _g2 , τ ^- _g3 , .... τ ^- _gm , τ _g ) a predetermined Threshold (var. Threshold). Method according to claim 10, characterized in that that the threshold (var. Threshold) is set from the outside. A method according to claim 11, characterized in that the threshold (var. Threshold) is increased with increasing signal noise. Use of a method according to one of the preceding claims in the equalization of teletext signals in a single-chip video ASIC. Device for measuring and compensating signal distortions caused by non-constant group delay along a transmission path, in particular a video transmission path, in particular teletext signal distortions, characterized in that - compensation means ( 1 ), which are an input signal, in particular a video signal (VIDEO_IN), with a reference signal (P _ref ) in a predetermined number (n, m) connected in series Sutfen (tap1, tap2, tap3, ... tapn; tap ^- 1, tap ^- 2, tap ^- 3, ... tap ^- m) sequentially each time by a group delay _value (τ _g1 , τ _g2 , τ _g3 , ... τ _gn , τ ^- _g1 , τ ^- _g2 , τ ^- _g3 , .... τ ^- _gm , τ _g ) of predetermined size compensate that - comparison means ( 2 ) are provided, each of which receives the compensated reference signals (P. 1) obtained after each stage (tap1, tap2, tap3, ... tapn; tap ^- 1, tap ^- 2, tap ^- 3, ... tap ^- 3, tap ^- m) _{ref, 1,} P _{ref, 2,} P _{ref, 3,} ..., P _{ref, n;} P ^- _{ref, 1,} P ^- _{ref, 2,} P ^- _{ref, 3,} ... P ^- _{ref, m)} compare the ideal course of the reference signal (P _ref ), - the reference signal (P _ref ) is a signal whose ideal course is symmetrical, in particular rectangular or triangular, and that - output means ( 1 . 2 ), which output just the compensated signal as an output signal (COMPENS_VIDEO_OUT), which the compensated reference signal (P _{ref, 1} , P _{ref, 2} , P _{ref, 3} , ... P _{ref, n} ; P ^- _{ref, 1} , P ^- _{ref, 2} , P ^- _{ref, 3} ,... P ^- _{ref, m} ) whose course is closest to the ideal course of the reference signal (P _ref ) and has the least symmetry deviation. Apparatus according to claim 14, characterized in that the compensation means ( 1 ) comprises a predetermined number (n, m) of all-pass stages (tap1, tap2, tap3, ... tapn; tap ^- 1, tap ^- 2, tap ^- 3, ... tap ^- m). Apparatus according to claim 14 or 15, characterized in that means are provided with which the number (n, m) of the active all-pass stages (tap1, tap2, tap3, ... tapn; tap ^- 1, tap ^- 2, tap ^- 3, ... tap ^- m) is configurable, in particular a multiplexer for selecting measurement outputs at different positions of a compensation network. Device according to one of Claims 14 to 16, characterized in that the allpass stages (tap1, tap2, tap3, ... tapn; tap ^- 1, tap ^- 2, tap ^- 3, ... tap ^- m) allpass Stages (tap1, tap2, tap3, ... tapn) for positive group delay (τ _g1 , τ _g2 , τ _g3 , ... τ _gn ) and all-pass tap ^- 1, tap ^- 2, tap ^- 3, .. tap ^- m) for negative group _delay (τ ^- _g1 , τ ^- _g2 , τ ^- _g3 , ... τ ^- _gm ). Device according to one of Claims 14 to 17, characterized in that the reference signal is a defined pulse sequence (P _ref ) in a teletext signal, in particular the first three consecutive data bits of the World Standard Teletext (WST) framing code. Device according to claim 18, characterized in that in that means are provided which represent the reference signal Pulse sequence within a measuring window, the sync signals of a external synchronization separation is derived, can recognize where the exact starting position of the pulse train from a rising Edge is derived. Device according to one of Claims 14 to 19, characterized in that means are provided which determine the number (n, m) of the stages (tap1, tap2, tap3, ... tapn; tap ^- 1, tap ^- 2, tap ^- 3 , ... tap ^- m), in which a compensation is performed, change cyclically, in particular a multiplexer for selecting measurement outputs at different positions of a compensation network. Apparatus according to claim 20, characterized in that means are provided which determine the number (n, m) of the stages (tap1, tap2, tap3, ... tapn; tap ^- 1, tap ^- 2, tap ^- 3, ... Tap ^- m), in which a compensation is performed, with reaching a maximum number (n, m) to a predetermined number, in particular one, reset, as such means, in particular the multiplexer for the selection of measurement outputs at different positions of the compensation network is provided. Device according to one of Claims 14 to 21, characterized in that similarity-determining means are provided for each comparison after each stage (tap1, tap2, tap3, ... tapn; tap ^- 1, tap ^- 2, tap ^- 3 , ... tap ^- m) obtained compensated reference signals (P _{ref, 1} , P _{ref, 2} , P _{ref, 3} , ... P _{ref, n} , P ^- _{ref, 1} , P ^- _{ref, 2} , P ^- _{ref , 3} ,... P ^- _{ref, m} ) determine a variable (dy _min ) representing a degree of similarity with the ideal profile of the non-distorted reference signal (P _ref ), in particular a comparison circuit for determining the symmetry of the compensated reference signal (P _{ref, 1} , P _{ref, 2} , P _{ref, 3} , ... P _{ref, n} ; P ^- _{ref, 1} , P ^- _{ref, 2} , P ^- _{ref, 3} , ... P ^- _{ref, m} ), and that the Determining degree of similarity determining means from said determined quantities (dy _min ) representing respective degrees of similarity, which compensated reference signal P _{ref, 1} , P _{ref, 2} , P _{ref, 3} , ... P _{ref, n} ; P ^- _{ref, 1} , P ^- _{ref, 2} , P ^- _{ref, 3} , ... P ^- _{ref, m} ) has the course closest to the ideal course of the non-distorted reference signal (P _ref ), as well as the results several successive degrees of similarity form an average value. Device according to claim 22, characterized in that in that averaging means comprise a multi-stage FIR filter. Device according to one of claims 14 to 23, characterized in that the output means ( 1 . 2 ) are provided, the compensated signal, the compensated reference signal (P _{ref, 1} , P _{ref, 2} , P _{ref, 3} , ... P _{ref, n} ; P ^- _{ref, 1} , P ^- _{ref, 2} , P ^- _{ref, 3} , ... P ^- _{ref, m} ), the course of which comes closest to the ideal course of the reference signal (P _ref ), only to output as output signal (COMPENS_VIDEO_OUT) if the compensation signal provided for the output is compensated required group delay (τ _g1 , τ _g2 , τ _g3 , ... τ _gn , τ ^- _g1 , τ ^- _g2 , τ ^- _g3 , ... τ _gm ) _exceeds a predetermined threshold (var. Threshold). Device according to claim 24, characterized in that that adjusting means are provided, via which the threshold value (bar., threshold) from the outside can be adjusted, in particular programmable registers. Device according to claim 25, characterized in that in that the setting means are provided to set the threshold value (var. threshold) with increasing signal noise. Use of a device according to one of claims 14 to 26 in the equalization of teletext signals in a signal chip Video ASIC.