EP1677277A2 - Control procedure for a matrix display screen - Google Patents

Abstract

The method involves applying a line selection voltage (VLS) to a line and a voltage to a column related to a grey level, during a line selection time (Tl) that is subdivided into groups according to a non-linear law with a transfer function close to the inverse of a gamma correction transfer function. The column voltage takes a value of one of two voltages throughout one group or switches from one voltage to the other during the group. The grey level is numerically coded according to the gamma transfer function based on the perception of the eye. The column voltage is selected among voltages (Vc0-Vc4) that are distributed in pairs. An independent claim is also included for a device for controlling a matrix display screen.

Description

TECHNICAL AREA

The present invention relates to a method of controlling a matrix display screen for displaying images having different gray levels. The images to display can be in black and white or in colors, in the latter case, the expression "gray level" means half-tone of color.

STATE OF THE PRIOR ART

The display matrix screens to which the present invention more particularly applies are flat screens with electron emission or with electron sources.

These include screens with hot cathodes, photoemissive cathodes and field effect microtip cathodes, as described in document referenced [1] at the end of the description, nanofield screens field effect, as described in the document referenced [2], plane electron source screens of the graphite or diamond carbon type, as described in the referenced document [3].

Figure 1 schematically illustrates in section a matrix display screen that uses a field emission electron source. This matrix display screen comprises one or more anode electrodes 1, cathode electrodes 2 which are electron sources, gate electrodes 3, isolated from the cathode electrodes 2 but which cooperate with them. These anode electrodes 1 on the one hand, cathode 2 and gate 3 on the other hand are located on different supports 5, 6 which when assembled delimit a space 7 in which the vacuum prevails. The cathode electrodes 2 emit a flow of electrons as a function of the electric field created by the potential difference V _GK imposed between the gate electrodes 3 and the cathode electrodes 2.

Polarization means are intended to apply a voltage V _G to the gate electrodes 3 and a voltage V _K to the cathode electrodes 2. The voltage V _GK represents the control voltage which is equal to (V _G - V _K ).

A voltage source V _A makes it possible to apply a high voltage to the anode electrodes 1. This voltage (of the order of several hundred volts) is substantially higher than that applied to the cathode electrodes to attract the electrons emitted by the electrodes. cathode electrodes.

Electrons emitted by the cathode electrodes 2 are accelerated and collected by the anode electrodes 1 subjected to the high voltage VA. If a layer of phosphor material 8 ("phosphor") is deposited on the anode electrodes 1, the kinetic energy of the electrons is converted into light.

Referring to Fig. 2, there is shown a display device with an electron emission matrix display screen and the associated control electronics. In the type of matrix display screen shown, the electrodes of cathode and gate extend in substantially perpendicular directions to form an electron emitter matrix array. The cathode electrodes generally form columns and grid electrodes of the rows of the matrix display screen. An image dot is at the intersection between a cathode electrode also called column electrode c1 to cm and a gate electrode also called line electrode 11 to 1p. The matrix display screen has p rows and m columns. The document referenced [4] describes how the addressing of each image point of the display screen and therefore the control of the luminance thereof.

The device of FIG. 2 comprises a line scan generator 9. This line scan generator 9 is connected to a voltage source VLS (for example equal to 80V) and to a voltage source VLNS which is generally grounded. The device also comprises a control generator 12, for controlling the columns, connected to two voltage sources 13 (for example equal to 40V) and 14 (for example ground), making it possible to apply one or other of these voltages, according to the information to be displayed, on the electrodes of columns c1 to cm. The generators 10 and 12, often called "drivers", are connected to a controller 16 which synchronizes the assembly in relation to data and control signals provided.

More precisely, all the lines 11, ... 1p are connected to the line scan generator 9 and they are each in turn selected sequentially for a line selection time Tls by applying to them the voltage VLS which is called the selection voltage. During this line selection time Tls, the unselected lines are brought to the voltage VLNS which is called the non-selection voltage. All the columns c1, .., cm are connected to the control generator 12.

During this line selection time Tls, the columns c1, ... cm are brought to a voltage corresponding to the information to be displayed by the image points at the intersection of the latter and the selected line. The voltage of the unselected lines VLNS is such that the voltages present on the columns do not affect the display on these lines.

To obtain a given gray level on an image point Pi, where j is at the crossing of the line li and of the column cj, it is possible to act on the value of the potential difference Vli-Vcj that applies between the line li and column cj as well as the duration t during which it applies, this duration having to remain less than or equal to the line selection time Tls.

FIG. 3 shows that the _cathode response curve (Vli-Vcj) of an image point Pi, j comprises an emission threshold Vs. As long as the line / column potential difference remains below this threshold Vs, it There is almost no electron emission.

When controlling an image point Pi, j, a modulation of the column voltage is therefore chosen Vcj such that when the voltage Vli equal to the non-selection voltage VLNS is applied to the line li, the potential difference between VLNS and the column voltage Vcj is always less than the threshold Vs and such that when one applies to the line li the voltage Vli equal to the selection voltage VLS, the column voltage Vcj makes it possible to reach a zone where the line / column potential difference vli-Vcj is beyond the threshold Vs.

We can polarize rows and columns in different ways to arrive at the same result because it is the potential difference line / column Vli-Vcj that drives the emission of electrons. FIGS. 4A and 4B show two different constructions which are equivalent at least for the selection time.

In FIG. 4A, the maximum potential difference (and therefore the white), is obtained by applying a column voltage V _Con close to VLNS and the minimum potential difference (for black) is obtained with a column voltage V _Coff intermediate between VLNS and VLS.

In contrast, in FIG. 4B, the minimum potential difference (and hence the black) is obtained by applying a column voltage V _Coff close to VLNS and the maximum potential difference is obtained with a column voltage V _Con lower than VLNS.

In the following descriptions, the examples will be given according to the voltage system of FIG. 4A. The invention applies of course to both systems.

In the following, the various control methods of the columns will be of interest in order to obtain suitable gray levels.

The analog modulation of the signal to be applied to the columns is the most obvious solution for displaying gray levels on an electron source display. This solution described in the referenced document [4] whose references are at the end of the description, however, is not very viable in the current state of the art at least for the control of complex display screens. Indeed, a display screen for HDTV type HDTV 1080p includes 1080 refreshed lines with a frequency of 60 Hz and 1920 columns. The line selection time Tls is approximately 15 μs. If we tolerate a setup time of about 1%, this represents a rise time of about 150 ns for a circuit that should also modulate a forty volts, a scan speed (English name of slew rate) 60V / μs. This performance must be performed at the least cost for the 1920 columns with severe constraints on the power consumption of these circuits.

We then turned to a two-level digital addressing with temporal modulation of the voltage during the line time Tls, PWM method for "Pulse Width Modulation". In such an addressing mode, all the columns c1 to cn are switched during a more or less long fraction of the line selection time Tls to a potential allowing to extract electrons and during the remainder of this same line selection time Tls switched to another potential blocking this extraction. This fraction is a function of the luminance to be obtained. The gray level displayed is directly proportional to the quantity of charges emitted. The document [5] whose references are at the end of the description describes this type of temporal modulation.

A disadvantage of this type of addressing is that the frequencies associated with these pulses do not pass without damage in the complex display screens. Indeed, the time constants specific to the display screen depend both on the resistivity of the row and column electrodes and the line / column capacity to be loaded at each voltage change. The reduction of these values is a strong technological constraint, difficult to achieve to display a large number of gray levels on large complex display screens. This method also comes up against the problem of the capacitive consumption generated by both the levels of the voltages to be switched and the frequency of these switches.

The method described in the patent application bearing the number [7], whose complete references are found at the end of the description, recommends the use of circuits making it possible to switch several voltage levels at linearly selected moments during line selection time. This makes it possible to minimize the switching levels to be applied to columns to reduce capacitive consumption, while displaying a large number of gray levels. The luminance response of an electron source display screen being quasi-linear with the excitation time, the method cited above is adapted to an electron source linearly encoding the gray levels.

The methods mentioned above describe solutions based on a linear coding of gray levels. Indeed, matrix display screens have mostly response types that follow a linear law whether with respect to a voltage or a duration of excitation.

It is known to those skilled in the art that it takes 12 to 14 bits to properly transmit a video signal in digital form if samples are used in linear proportion of light intensity.

The European patent application bearing the number [6] and whose references are at the end of the description proposes to perform a correction of the luminances from linearly coded data by varying the light emission amplitude over time. line selection, thus giving equal reference deviations, different illumination differences. It also proposes to obtain the same result by using pulse width modulation (PWM) with a nonlinear distribution of times. This corrects the answer for high levels of illumination but can not restore details for low levels.

If one wanted in the document referenced [7], to use a source whose grayscale coding is not linear, it would be necessary to encode the signal of the source by means for example of a table of conversion (known under the acronym for LUT for Look Up Table) to obtain a numerical code that varies linearly with the luminous intensity to be transmitted, which must lead to a coding of at least 12 bits for a good rendering of the gray levels.

STATEMENT OF THE INVENTION

It is an object of the invention to improve the control method described in [7] to enable the display of a gray scale continuum according to the response curve of the human eye without significantly increasing the number of samples. treat. A larger number of samples means samples of shorter durations and thus clashes with the time constants of the display screen. The method of the invention does not apply exclusively to microtip screens but to any type of electron emission matrix display screen.

Another object of the invention is to make the response of an electron source screen compatible with the nonlinear coding called gamma correction.

Another object of the invention is to reduce the number of bits to be used for controlling a matrix display screen.

Another object of the invention is to propose a method for controlling a matrix display screen whose capacitive consumption is minimized.

In order to achieve this, the present invention proposes, in order to display the gray levels, instead of using samples in linear proportion of the light intensity, to use samples in a non-linear proportion by opting for a coding which follows the curve of response of the human eye because the eye is more sensitive to luminance differences at low level of illumination than at high level. His perception of luminance follows a nonlinear law known as gamma correction which has been modeled in particular by the International Commission on Illumination.

More specifically, the present invention relates to a method of controlling a matrix display screen with lines and columns whose intersections form a picture dot, including applying to a line a line selection voltage during a line selection time and simultaneously with a column of a control voltage corresponding to a gray level to be displayed by the associated image point. The gray level is selected from 2 ^q levels and numerically coded according to a nonlinear law consistent with the perception of luminance by a human eye. The different voltages to be applied to the columns are chosen in a strictly increasing sequence of (2 ⁿ + 1) voltages with n being ≥ 1, these voltages being distributed in N = 2 ⁿ consecutive pairs of voltages, each pair to display a range of gray levels. The line selection time is subdivided into one or more groups of 2 ^(qn) time intervals with (qn) integer> 1, each group having the same duration. The distribution of these 2 ^(qn) time intervals in a group is, for a pair of voltages, according to a nonlinear law having a transfer function close to the inverse of that previously used to code the gray levels to the corresponding gray level range. The voltage applied to a column takes, for a pair of voltages and a group of time intervals, either one of the values of the voltage torque during the whole duration of the group, or switches at least once, during the duration of the group, at the end of a time interval, of a value of the pair of voltages to the other.

et $$\mathrm{G}=\left[{\left(1,099\times {\mathrm{P}}_{\mathrm{G}}\right)}^{0,45}-0,099\right]\left(2^{\mathrm{q}}-1\right)\mathrm{\hspace{0.17em}}\mathrm{pour}\mathrm{\hspace{0.17em}}{\mathrm{P}}_{\mathrm{G}}>0,018,$$

P_G étant un poids relatif (normé à un) attribué au niveau de gris.The law of coding of a gray level G is such that: $$\mathrm{BOY\; WUT}=\left({\mathrm{two}}^{\mathrm{q}}-1\right)\times 4,5\times {\mathrm{P}}_{\mathrm{BOY\; WUT}}\mathrm{}\mathrm{for}\mathrm{}{\mathrm{P}}_{\mathrm{BOY\; WUT}}\le 0,018$$

and $$\mathrm{BOY\; WUT}=\left[{\left(1,099\times {\mathrm{P}}_{\mathrm{BOY\; WUT}}\right)}^{0,45}-0,099\right]\left({\mathrm{two}}^{\mathrm{q}}-1\right)\mathrm{}\mathrm{for}\mathrm{}{\mathrm{P}}_{\mathrm{BOY\; WUT}}>0,018,$$

P _G being a relative weight (standardized to one) attributed to the gray level.

avec τ durée d'un groupe de 2^q-n intervalles de temps, P_G poids du niveau de gris à afficher, P_Gsup et P_Ginf poids des niveaux de gris correspondants respectivement aux bornes supérieures et inférieures de la plage de niveaux de gris associée au couple de tensions.For a given pair of voltages, a range of gray levels, a group of time intervals, and a gray level to be displayed, the torque voltage leading to the highest luminance can be applied for a given time interval by : $$\mathrm{.delta.t}=\mathrm{\tau }({\mathrm{P}}_{\mathrm{BOY\; WUT}}-{\mathrm{P}}_{\mathrm{Ginf}})/({\mathrm{P}}_{\mathrm{GSUP}}-{\mathrm{P}}_{\mathrm{Ginf}})$$

with τ duration of a group of 2 ^qn time intervals, P _G weight of the gray level to be displayed, P _Gsup and P _Ginf weight of the respective gray levels respectively at the upper and lower bounds of the gray level range associated with the voltage pair.

When the line selection time is subdivided into two groups of time slots, the time slots in the two groups can be distributed symmetrically with respect to the middle of the line selection time, so as to limit current calls of switching common to all columns.

It is preferable that during the line selection time, the line selection voltage is transient free.

The present invention is particularly applicable to electron source flat screens.

• une source de donnée numériques apte à fournir des mots binaires codés sur q bits selon la loi non linéaire conforme à la perception de la luminance par un oeil humain et représentant les codes des 2^q niveaux de gris à afficher,
• un contrôleur d'écran recevant des signaux de synchronisation de la source de données et gérant des signaux propres à piloter un générateur de balayage des lignes et un générateur de tensions de pilotage des colonnes qui reçoit pour chaque colonne les codes des niveaux de gris à afficher et qui permet de générer, à partir d'un générateur de tensions discrètes, les tensions des couples de tension, et de commuter si nécessaire d'une tension à l'autre dans un couple.

The present invention also relates to a device for controlling a matrix display screen according to the method previously described which comprises:

• a digital data source capable of providing binary words encoded on q bits according to the non-linear law according to the perception of luminance by a human eye and representing the codes of the 2 ^q gray levels to be displayed,
• a screen controller receiving synchronization signals from the data source and managing signals suitable for driving a line scan generator and a column driving voltage generator which receives for each column the gray level codes to be displayed which makes it possible to generate, from a generator of discrete voltages, the voltages of the voltage torques, and switch if necessary from one voltage to the other in a couple.

The binary words are subdivided into two subwords one of n bits corresponding to the most significant bits and the other of q-n bits corresponding to the least significant bits. The column driving voltage generator may comprise for each column, a combinatorial logic stage controlling a series of analog switches to preselect a voltage pair delivered by the discrete voltage generator from the n most significant bits of a word binary and a signal from a counter initialized at each line selection time and to switch from one of the voltages to the other torque when the counter has reached the value corresponding to qn low-order bits of the word binary.

The counter receives a series of non-linearly distributed pulses corresponding to the pair of voltages coming from a pulse generator connected to the screen controller via a multiplexer also receiving in address the n most significant bits of the binary word delivered by the data source.

The combinational logic stage can receive the n most significant bits of the binary word delivered by the data source via a shift register associated with memory latches.

The combinatorial logic stage can be connected to the counter via a comparator which compares the signal from the counter with the qn bits of least significance of the binary word delivered by the data source.

BRIEF DESCRIPTION OF THE DRAWINGS

• la figure 1 représente en coupe un écran de visualisation matriciel auquel peut s'appliquer le procédé de l'invention ;
• la figure 2 illustre schématiquement un dispositif de commande d'un écran de visualisation matriciel connu auquel peut s'appliquer le procédé de l'invention ;
• la figure 3 montre l'allure du courant cathode en fonction de la tension appliquée à la ligne li et à la colonne cj ;
• les figures 4A, 4B illustrent les tensions à appliquer à une ligne et une colonne dans deux configurations différentes pour commander un point image ;
• la figure 5A présente la courbe dite de correction de gamma qui transforme une intensité lumineuse ou luminance en un signal vidéo et la figure 5B est la courbe inverse ;
• les figures 6A, 6B montrent deux exemples de signaux à appliquer à une ligne et à une colonne d'un écran de visualisation matriciel pour afficher un niveau de gris donné selon le procédé de l'invention ;
• La figure 7 illustre sous forme de chronogrammes (7A à 7F) les signaux mis en jeux pour la commande d'un écran de visualisation matriciel dans l'exemple de la figure 6B avec prise en compte des temps morts ;
• les figures 8A, 8B illustrent de manière générale et en détail, un dispositif de commande d'un écran de visualisation matriciel selon le procédé de l'invention.

The present invention will be better understood on reading the description of exemplary embodiments given, purely by way of indication and in no way limiting, with reference to the appended drawings in which:

• FIG. 1 represents in section a matrix display screen to which the method of the invention can be applied;
• FIG. 2 diagrammatically illustrates a control device of a known matrix display screen to which the method of the invention can be applied;
• Figure 3 shows the shape of the cathode current as a function of the voltage applied to the line li and to the column cj;
• FIGS. 4A, 4B illustrate the voltages to be applied to a line and a column in two different configurations for controlling a picture point;
• FIG. 5A shows the so-called gamma correction curve which transforms a luminous intensity or luminance into a video signal and FIG. 5B is the inverse curve;
• FIGS. 6A, 6B show two examples of signals to be applied to a row and a column of a matrix display screen to display a given gray level according to the method of the invention;
• FIG. 7 illustrates in the form of timing diagrams (7A to 7F) the signals used for controlling a matrix display screen in the example of FIG. 6B, taking into account dead times;
• FIGS. 8A, 8B generally and in detail illustrate a device for controlling a matrix display screen according to the method of the invention.

Identical, similar or equivalent parts of the different figures described below bear the same numerical references so as to facilitate the passage from one figure to another.

The different parts shown in the figures are not necessarily in a uniform scale, to make the figures more readable.

DETAILED PRESENTATION OF PARTICULAR EMBODIMENTS

An example of the process of the invention will now be described in more detail.

In this method, an addressing mode is used presenting both the modulation possibilities in terms of time and voltage offered by the electro-optical response of a screen using an electron source. Beyond the threshold, the luminance obtained varies according to a quasi-exponential law with the voltage and linearly with the duration of excitation.

The display screen to which the invention applies, is a matrix display screen to p lines 11 to 1p and m columns c1 to cm. This matrix display screen is electron source.

In the method of the invention, during the line selection time Tls, a selected line voltage VLS is applied to a selected line li which in the example described will be at a high level since in this example the operation will be performed according to the mode of Figure 4A. During this line selection time T1s, control voltages will be applied to the columns c1 to cm such that the image point Pi, j at the intersection between the selected line li and the column concerned cj will display a desired gray level G . We choose to have 2 ^q gray levels with q integer. The number q is the number of bits that will be used to code the gray level. Eight-bit color coding allows good grayscale rendering.

To obtain these 2 ^q gray levels, we choose to have a set of 2 ⁿ + 1 (n ≥ 1) voltages Vc0 to Vc2 ⁿ or else VcN if N = 2 ⁿ successive to be applied to columns c1 to cm of the matrix display screen. This set of tensions to apply to the columns forms a strictly increasing sequence. The voltages of the sequence are grouped into N = 2 ⁿ pairs of voltages, each being formed by two consecutive voltages of the sequence (Vci, Vci + 1). One of the voltages Vci + 1 makes it possible to display, at one image point, a higher luminance than is allowed by the other voltage Vci.

It can be written that in the increasing series of tensions, the tensions of a couple are such that:
Vci = Vci + 1 + δvi and the sign of δvi depends on the chosen voltage system as explained in FIGS. 4A, 4B. The potential difference of VLS with respect to Vci is lower than that with respect to Vci + 1. The voltage deviations δvi may be equal or different as explained in document [7]. Schematically illustrated voltage differences substantially equal on the luminance / voltage curve of Figure 5B.

Thus 2 ⁿ = N pairs of voltages are defined such that (Vc0, Vc1), (Vc1, Vc2) ... (Vci-1, Vci), ......, (VcN-1, VcN), Vc0 represents the voltage whose potential difference with respect to the line selection voltage VLS corresponds to the Vs electron emission threshold. The application of a voltage of one of the couples on a column immediately causes an emission more or less electrons according to the luminance / voltage curve shown in Figure 5B.

The 2 ^q gray levels are divided into 2 ^qn ranges n of q gray levels. At each pair of voltages Vci, Vci + 1 is associated with one of the ranges n of q gray levels.

The line selection time Tls is subdivided into one or more groups T of 2 ^(qn) time intervals [ti + 1, ti] between times t0 and t2 ^qn , with qn greater than one. Each group T has the same duration. Within a group T, the time intervals t0 to t2 ^qn have different durations for each pair of voltages (Vci, Vci + 1). The time intervals t0 to t2 ^qn of a group T have durations which follow a nonlinear law which has a transfer function close to the inverse of that previously used to code the gray levels for the corresponding range n of gray levels.

A given gray level G is in one of the ranges n of gray levels and to display this gray level G, we will use the pair of voltages (Vci, Vci + 1) which is associated with the said range n of levels of gray.

For a group T of given time intervals, one of the torque voltages (Vci, Vci + 1), for example Vci, is applied to a column Ci for the duration of the group, or, initially, one of the voltages, for example Vci, is applied and then at the end of a time interval of the group T, at least once, the previous voltage Vci is switched to the other voltage Vci + 1 of the pair.

The voltage to be applied to a column ci then takes the first value Vci during an integer number of time intervals [ti, ti-1] of the group T and switches to the second value Vci + 1 if necessary during the rest of the time of the group if, for example, only one switching is planned.

The time intervals between the instants t2 ^qn and t0 in the group of time intervals T have durations that vary from one pair of voltages to another according to a nonlinear law having a transfer function close to the inverse of the one used to code the gray levels.

The first pair of voltages (Vc0, Vc1) makes it possible to display the range n of the 2 ^q / 2 ⁿ first levels of gray, the second pair of voltages (Vc1, Vc2) makes it possible to display the range n of the following 2 ^q / 2 ⁿ gray levels and so on with the pair of voltages (VcN-1, VcN) which makes it possible to show the range n of the last 2 ^q / 2 ⁿ gray levels.

pour P_G ≤0,018 et $$\mathrm{G}=\left[{\left(1,099\times {\mathrm{P}}_{\mathrm{G}}\right)}^{0,45}-0,099\right]\left(2^{\mathrm{q}}-1\right)$$

A gray level G, expressed in decimal base, can be expressed in accordance with a nonlinear law consistent with the perception of luminance by the human eye. This law can be expressed as follows: $$\mathrm{BOY\; WUT}=\left({\mathrm{two}}^{\mathrm{q}}-1\right)\times 4,5\times {\mathrm{P}}_{\mathrm{BOY\; WUT}}$$

for P _G ≤0.018 and $$\mathrm{BOY\; WUT}=\left[{\left(1,099\times {\mathrm{P}}_{\mathrm{BOY\; WUT}}\right)}^{0,45}-0,099\right]\left({\mathrm{two}}^{\mathrm{q}}-1\right)$$

For P _G > 0.018

P _G represents the normalized weight to one that has the coded gray level.

These relations are modeled on the transfer function of the so-called gamma correction law which transforms a linearly varying intensity of light into a nonlinear video signal so that the human eye has a perception as faithful as possible. This nonlinear law has been modeled and is standardized as the document [8] whose references are at the end of the document. It is shown in Figure 5A. The substantially linear portion corresponds to the levels close to black corresponding to 8.1% of the signal corresponding to the white and intensity less than or equal to 1.8% of the luminance of the white.

τ est la durée d'un groupe de 2^q-n intervalles de temps, P_G poids relatif (normé à un) du niveau de gris à afficher, P_Gsup et P_Ginf poids des niveaux de gris correspondants respectivement aux bornes supérieures et inférieures de la plage de niveaux de gris associée au couple de tensions Vci, Vci+1. Lorsqu'il n'y a qu'un seul groupe dans le temps de sélection de ligne, τ est égal à Tls. Le poids P_G est calculé en prenant l'inverse des formules (1) et (2) soit : $${\mathrm{P}}_{\mathrm{G}}=\frac{G}{\left(2^q-1\right)\times 4,5}\mathrm{\hspace{0.17em}}\mathrm{pour}\mathrm{\hspace{0.17em}}{\mathrm{P}}_{\mathrm{G}}\le 0,018$$

Et $$P_G={\left[\frac{\frac{G}{2^q-1}+0,099}{1,099}\right]}^{1/0,45}\hspace{0.17em}\mathrm{pour}\hspace{0.17em}{\mathrm{P}}_{\mathrm{G}}>0,018$$

According to the G code of the gray level to display, assuming that the voltage is switched once to apply this gray level, the application time on a column of the voltage to display the highest luminance in the pair V _ci , V _{ci + 1 is} expressed by: $$\mathrm{\Delta }\mathrm{t}=\mathrm{\tau }\left({\mathrm{P}}_{\mathrm{BOY\; WUT}}-{\mathrm{P}}_{\mathrm{BOY\; WUT}\inf }\right)/\left({\mathrm{P}}_{\mathrm{BOY\; WUT}\sup }-{\mathrm{P}}_{\mathrm{BOY\; WUT}\inf }\right)$$

τ is the duration of a group of 2 ^qn time intervals, P _G relative weight (normalized to one) of the gray level to be displayed, P _Gsup and P _{Ginf the} weight of the gray levels corresponding respectively to the upper and lower limits of the range of gray levels associated with the voltage pair Vci, Vci + 1. When there is only one group in the line selection time, τ is equal to Tls. The weight P _G is calculated by taking the inverse formula (1) and (2): $${\mathrm{P}}_{\mathrm{BOY\; WUT}}=\frac{\mathrm{BOY\; WUT}}{\left({\mathrm{two}}^q-1\right)\times 4,5}\mathrm{}\mathrm{for}\mathrm{}{\mathrm{P}}_{\mathrm{BOY\; WUT}}\le 0,018$$

And $$P_{\mathrm{BOY\; WUT}}={\left[\frac{\frac{\mathrm{BOY\; WUT}}{{\mathrm{two}}^q-1}+0,099}{1,099}\right]}^{1/0,45}\mathrm{for}{\mathrm{P}}_{\mathrm{BOY\; WUT}}>0,018$$

The weights P _Gsup and P _Ginf are calculated in the same way from the gray levels Gsup and Ginf which limit the range of gray levels associated with the pair of voltages. We can write that:

Ginf = G (ix2 ^qn ) and Gsup = G ((i + 1) × 2 ^qn ) with i integer varying from 0 to N-1.

This time Δt is therefore coded according to a nonlinear law having a transfer function close to the inverse of that (1), (2) used to code the gray levels G for the range n of gray levels considered.

It is assumed, in the example described, that the line selection time Tls has only one group of time slots and that the voltage Vc0 is first applied so as to display the lowest luminance then then that one switches to the voltage Vc1 allowing the strongest luminance during the rest of the time of Tls.

The first voltage pair (Vc0, Vc1) is used to display the first (2 ^qn ) gray levels. The application time of the voltage Vc1 allowing the emission, according to the G code of the gray level to be displayed, must be such that: $$\Delta t=Tls\times \frac{P_{\mathrm{BOY\; WUT}}}{P_{\mathrm{BOY\; WUT}\sup }}$$

We will now give a first concrete example with the same hypotheses.

One wishes to display on a point image a level of gray G = 118. This level of gray is coded on eight bits is q = 8. There are ²⁸ = 256 gray levels. We have 2 ⁿ + 1 voltage levels with for example n = 2 or 5 voltage levels called Vc0, Vc1, Vc2, Vc3, Vc4 forming N = 4 couples (Vc0, Vc1), (Vc1, Vc2), (Vc2 , Vc3), (Vc3, Vc4) whose rank ranges from 1 to 4. The line selection time Tls is subdivided into a group T of 2 ^qn = 2 ⁶ = 64 time intervals whose terminals go from t64 to t0 . These time intervals (by example t62-t63) or samples have different durations for each pair of voltages and these durations are to be calculated according to formula (3). In fact, this formula makes it possible to place the terminals of these time intervals with respect to the line selection time Tls.

The binary code corresponding to the gray level G = 118 in base 10 is:

The first n high-order bits give information on the voltage pair that will be used to control the column that must be addressed to display the gray level G. In our example, the n most significant bits are equal to 01. This is the pair (Vc1, Vc2) of rank 2, which will serve. If these n most significant bits had been 00, we would have taken the pair (Vc0, Vc1) of rank 1. If these n bits of weight had been 11, we would have used the pair (Vc2, Vc3) of rank 3. More generally the Y value in decimal of the n most significant bits leads to the use of the pair of tensions of rank Y + 1. The pair (Vc1 Vc2) of the example displays the range of gray levels whose terminals are 64 and 128.

The inverse P ₁₁₈ of the code 118 of the gray level considered will be calculated. $$P_{118}={\left(\frac{\frac{118}{256-1}+0,099}{1,099}\right)}^{1/0,45}=0,225.$$

P ₁₁₈ was calculated with formula (2 ').

The q-n low-order bits 110101 correspond to 53 in base 10. They indicate when the switching of the voltage Vc1 to the voltage Vc2 will occur. It will occur at time t53 which is the upper bound of the time interval [t54, t53].

• P_Ginf = P₆₄= 0,0786
• P_Gsup = P₁₂₈ = 0,2615
• P₆₄ et P₁₂₈ ont été calculés avec la formule (2'). $$\mathrm{\Delta }\mathrm{t}=Tlsx\left(\frac{0,225-0,0786}{0,2615-0,0786}\right)=0,8xTls$$
  

One can also calculate the application time of the determined torque voltage (Vc1, Vc2) for displaying the highest luminance is Vc2.

• P _Ginf = P ₆₄ = 0.0786
• P _Gsup = P ₁₂₈ = 0.2615
• P ₆₄ and P ₁₂₈ were calculated with formula (2 '). $$\mathrm{\Delta }\mathrm{t}=Tlsx\left(\frac{0,225-0,0786}{0,2615-0,0786}\right)=0,8xTls$$
  

The voltage Vc2 will be applied for 80% of the line selection time Tls between times t53 and t0 of the line selection time Tls while the voltage vc1 will only be applied for 20% of the line selection time Tls between t64 and t53 instants.

FIG. 6A shows a first simplified example of the signal to be applied to a line for selecting it and the signal to be applied to a column so that the image point at the intersection of the line and the column displays a gray level. coded 11 in decimal. During the line selection time Tls between t8 and t0, a selected line sees the voltage applied to it go from VLNS to VLS at the beginning of time t8 and return to VLNS at the end of time t0. We assume that we choose q = 5 and that we have 2 ⁵ = 32 gray levels.

We assume that n = 2 and that we therefore have 2 ^{n + 1} = 5 voltage values Vc0, Vc1, vc2, Vc3, Vc4 distributed in N = 4 pairs (Vc0, Vc1), (Vc1, vc2), (vc2, Vc3), (Vc3, Vc4). These voltages are ranked in descending order from Vc0 to Vc4 in this example. The voltage Vc0 is such that VLS-Vc0 corresponds to the electron emission threshold of the matrix display screen used. With the voltage Vc0 electrons are emitted but at a very low level which may be insufficient depending on the external light atmosphere to have a perception of the associated luminance.

The voltage Vc4 makes it possible to display the highest luminance.

The binary code of the gray level coded G = 11 is:
01 011

The n = 2 most significant bits 01 indicate the voltage torque to be used for displaying this gray level. The voltage pair to be used is the second (Vc1 Vc2).

The low-order bits 011 indicate the moment when one of the voltages Vc1 will be switched to the other Vc2.

The bits of low weight 011 correspond to 3 in decimal base. During the line selection time Tls, there are 2 ^qn = 8 time intervals bounded by the instants t8, t7, t6, t5, t4, t3, t2, t1 and t0. For each voltage pair, the instants t7 to t1 are positioned between t8 and t0 according to the calculations of Δt explained above.

In this example, the successive time intervals have decreasing durations but of course they could have had increasing durations. It could be envisaged, alternatively, that the distribution of the time intervals corresponds to that of the part comprised between t0 and t8 of FIG. 6B, with the difference that the time interval [t8, t0] would be equal to the selection time line T1s and not halfway.

Returning to the example illustrated in FIG. 6A, it has been seen that the low-order bits 011 represent 3 at the base ten, this implies that the switching of the voltage Vc1 to the voltage Vc2 will be done at time t3, at the result of the time interval which goes from t4 to t3.

In the other example from FIG. 6B, switching of the voltage Vc2 to the voltage Vc1 would be done at time t3, at the end of the time interval which goes from t2 to t3.

The different switching times possible from one voltage to the other can be controlled by pulse signals from a screen controller 21 in response for example to a conversion table 11 as illustrated in FIG. 8B. These pulses can be used by counting means 13 which, depending on the low-order bits of the gray level to be displayed, and therefore the appropriate torque of the voltages will determine the instant or times when a switching must take place. At every tension will be associated a specific signal carrying all possible switching times. Only one or a few of these moments will be activated according to the gray level to be displayed, the others will be neutralized.

The dashed grid represents the different switching times possible for each of the voltage pairs and the number entered in each box represents the gray level obtained, expressed in decimal base, if a switching occurs in a voltage voltage pair allowing display the lowest luminance towards the voltage to display the highest luminance.

The bold line shows the pace of the signal to be applied to a column to obtain on a point image, located at the intersection of this column and a selected line, the gray level encoded 11. The boxes bearing the numbers 0, 8, 16, 24, i.e. of the form 2 ^k , represent the gray levels obtained without switching. To obtain the gray level 0, the voltage remains during the entire line selection time T1s to Vc0, for the gray level 8, the voltage remains at Vc1 and so on.

It is of course possible to provide in the line selection time Tls several groups of time slots. FIG. 6B illustrates an advantageous variant of implementation of the invention in which the line selection time T1s is subdivided into two equal groups T1, T2 of 2 ⁿ time intervals. In these two groups the distribution of the 2 ^qn time intervals is symmetrical with respect to the middle of the line selection time Tls. The first group T1 is bounded by instants t0 to t8 and the second group is bounded by instants t8 '(= t8) and t0'.

To display the coded gray level 11, as previously we use the voltage pair (Vc1, Vc2). During the first group T1 of time intervals, the voltage Vc2 applies from time t0 to time t3, then there is switching to voltage Vc1. The voltage Vc1 applies from time t3 to time t8. During the second time interval T2, the voltage Vc1 applies from time t8 'to time t3', then there is switching to voltage Vc2. The voltage Vc2 applies from time t3 'to time t0'. The signal that applies to the columns is symmetrical with respect to the instant t8 or t8 ', these two moments being merged.

In FIGS. 6A and 6B, the instants referenced correspond to those of the voltage pair Vc3, Vc4.

This variant makes it possible to obtain a distribution of rise and fall times which avoids the common switching current calls to all the columns.

To overcome unavoidable distortions due to the rise and fall times of the signals to be applied to the lines of the display screen, it is possible that the line selection time Tls taken into account for the control of the signals on the columns corresponds to the duration during which the line selection signal is well established, that is, has reached its VLS level and therefore, that tm end dead times corresponding to the rising and falling edges of the signal as it passes from the VLNS non-selection level to the VLS selection level and vice versa should be ignored. The line selection time is then free of end dead time and therefore of voltage transient.

FIG. 7 illustrates in the form of timing diagrams (7A to 7F) the signals used to control a matrix display screen in the example of FIG. 6B, taking into account dead times.

The timing diagram 7A is a line clock signal HL whose rising edge is intended to trigger a transition of the signal to be applied to the lines: transition from VLNS to VLS or vice versa.

The signal to be applied on the lines is illustrated by the timing diagram 7D. It is generated by the line scan generator. The solid lines show the signal to be applied on a line that is selected and the dashed line signals illustrate the signals that applied to the previous line and will apply to the next line.

The timing diagram 7B illustrates an LC pulse signal which controls the loading of data relating to the columns to be controlled. At each rising edge of the pulse, the column driving voltage generator switches the column of a given image point of the voltage it had at the previous line selection time to the voltage it must take at the beginning of the current line selection time.

Timechart 7C illustrates the signal CC which will allow the signal to be applied to the columns to overcome the end dead times tm. The positive part of this signal corresponds to the time Tls used to code the gray levels. The signal applies to the pulse generator 11 (described later in FIG. 8) to validate the useful time of the series of pulses as illustrated by the timing diagram 7F.

Timechart 7E gives the appearance of the signal to be applied to a column to display the gray level shown in Figure 6B. From the beginning of the line selection time free of dead end time tm, the voltage applied to the column is Vc2, the column keeps this voltage for a duration D1t then switches to the voltage Vc1, it keeps this voltage and then switches to the voltage Vc2 with an advance of D1t with respect to the end of the line selection time T1s free of the end dead time tm.

The voltage applied to this same column will then be used to control a point image at the cross of this column and the next line which is now selected. It is applied once the end dead time tm has passed a voltage Vcj + 1 for a duration D2t then switches to the voltage Vcj, it keeps this voltage Vcj and then switches to the voltage Vcj + 1 with an advance of D2t with respect to the end of the line selection time Tls free of the end dead time tm.

The timing diagram 7F illustrates a series of pulses taken from the 2 ⁿ series which are associated with each pair of voltages (Vci, Vci + 1). The pulses are produced by a pulse generator 11 to generate the switching between the two voltages via a multiplexer 12, a counter 13 and a comparator 14 described later in FIG. 8B.

We will now be interested in a control device of a matrix display screen 25 for displaying gray levels with reference to FIGS. 8A, 8B. The screen 25 has lines li and columns cj which intersect and whose intersections form image points Pi, j.

The control device comprises a row scanning generator 22 and a control voltage generator of the columns 23. The control voltage generator of the columns 23 is connected to a digital data source 20 capable of supplying binary words representing the codes 2 ^q gray levels to be displayed, a screen controller 21 and a generator of 2 ⁿ + 1 (or N + 1) discrete voltages 24. These binary words are coded on q bits according to the response curve of the human eye. Screen controller 21 is also connected to line scanner 22 (not visible in FIG. 8B). The screen controller 21 receives synchronization signals from the data source 20, it manages and supplies signals suitable for driving the line scan generator 22 and the pilot voltage generator of the columns 23.

The binary words provided by the data source 20 have q bits, they decompose in two subwords one of which is formed of n high-order bits and the other of low-order bits. The sub word formed of the n bits serves to determine the pair of voltages to be used according to the gray level to be displayed. The sub word consisting of qn low-weight bits reflects a moment of switching from one torque voltage to the other.

The control voltage generator of the columns 23 has as much output as columns. Each output Cout of the control voltage generator of the columns 23 is controlled by a control circuit 16 comprising N + 1 analog switches CA whose outputs are all connected to the column of the considered channel, the validation inputs of these switches being driven by a combinatorial logic stage 15 receiving, on the one hand, the subword formed of the n most significant bits representing the gray level to be displayed and, on the other hand, by a signal coming from a counter 13 initialized at each time of line Tls and providing the index of the addressing sequence within the line selection time Tls. This counter 13 itself receives a series of 2 ^qn pulses per line time Tls, this series of non-linearly distributed pulses being chosen from 2 ⁿ series according to the n high-order bits of the gray level to be displayed. The series of pulses are provided, for example, for example by a conversion table type circuit (known under the name of look-up table or LUT) associated with a counter reset at each start of line selection time Tls. The times Δt between the beginning of the line selection time and each of the pulses provided are consistent with those expressed by formulas (1), (2), (3). The combinational logic stage 15 makes it possible to select a pair of voltages from the n most significant bits and then to switch from the voltage Vci to Vci + 1 when the counter 13 reaches the value corresponding to the qn least significant bits of the binary word. corresponding to the gray level to display.

More specifically, as shown in FIG. 8B, the screen controller 21 supplies the pulse generator 11 with the control signals LC (timing chart 7B) and CC (timing chart 7C) of FIG. 7. The digital data source 20 provides the q-bit data words according to the response source of the human eye and the necessary synchronization signals. The data are conventionally received by a device 10 including for each column a shift register associated with memory latches for the different channels to be controlled. Moreover, each output Cout of the control voltage generator of the columns 23 must be able to switch between two voltages supplied by the generator of N + 1 discrete voltages 24.

The pulse generator 11 produces on its 2 ⁿ outputs series of non-linearly distributed pulses corresponding to the instants t0 to t2 ^qn (timing diagram 7F) for respectively each pair of voltages (Vci, Vci + 1). It is connected at the output to the multiplexer 12 also receiving in address the n most significant bits of the binary word of q bits representing the gray level to display delivered by the data source 20. The multiplexer 12 makes it possible to switch to the counter 13 the series of pulses corresponding to the pair of voltages determined by the n most significant bits. The output of the multiplexer 12 serves as a clock at the counter 13 which is reset at the beginning of the line selection time by the screen controller 21. The counter 13 of qn bits is connected at the output to a comparator 14 which provides a comparing the word of qn bits supplied by the counter 13 and the qn least significant bits of the binary word provided by the data source 20. The output of the comparator 14 changes state when the value counted by the counter 13 becomes equal or greater than the value from the data source 20. The output of the comparator 14 is sent to the combinatorial logic stage 15 which also receives the n most significant bits, to control the passage of the voltage Vci (selected by the bits of high weight) at the voltage Vc + 1. This passage is provided by the control circuit 16 comprising the analog switches CA whose outputs form an output Cout of the control voltage generator of the columns 23, this output being to be connected to a column of the display screen 25, a single CA switches being closed at a time.

An advantage of the method of the invention compared to conventional methods with temporal modulation of the voltage is that the columns receive less information to manage for the same image quality but also that the use of time is optimized. For example, in a system with two voltage levels (classical PWM), to transcribe an image digital video in HDTV 1080p quality, it takes at least 12 bits for a line time of about 15μs. A linear division of time gives a sample interval of 15/2 ¹² # 3.7ns. Switching to 8 bits increases this time by a factor of 16.

avec G=254, G étant le code du niveau de gris à afficher et γ valant environ 2,5.
Pour un temps de sélection de ligne Tls de 15µs, Ψ vaut environ 150ns. Tous les autres niveaux de gris G (variant de 0 à 254) sont obtenus avec des durées plus longues puisque les nuances de rang inférieurs sont obtenues en prolongeant cette diminution du temps d'excitation. Le choix de G valant 255 donne Ψ= 0 et la tension VcN+1 est maintenue pendant tout le temps de sélection de ligne Tls. Dans le procédé de l'invention, la durée d'impulsion la plus courte est donc 40 fois plus longue que celle à mettre en oeuvre avec un procédé linéaire.More precisely, in the voltage system of FIG. 4A, the white corresponding to the level 255 for an 8-bit coding is obtained on the image point considered, during the line selection time Tls, by applying the voltage VLS on the line and the minimum voltage, for example VLNS = 0 volts on the column. As a first approximation, to obtain the level 254 gray, it is then necessary to reduce the excitation time of the column by: $$\mathrm{\Psi }=Tls\left[1-{\left(\frac{\mathrm{BOY\; WUT}}{255}\right)}^{\gamma }\right]$$

with G = 254, where G is the gray level code to be displayed and γ is approximately 2.5.
For a line selection time Tls of 15μs, Ψ is about 150ns. All other gray levels G (ranging from 0 to 254) are obtained with longer durations since lower rank grades are obtained by prolonging this decrease in excitation time. The choice of G equaling 255 gives Ψ = 0 and the voltage VcN + 1 is maintained throughout the line selection time Tls. In the process of the invention, the shortest pulse duration is therefore 40 times longer than that to be implemented with a linear process.

In summary, the method according to the invention makes it possible to increase the time between switching while decreasing the capacitive consumption by minimizing the transitions of voltages applied to the columns during the passage of a line addressed to the next. Depending on the gray level to be displayed, in a voltage pair, switching at times distributed on well-defined non-linear scales is provided. The low-order bits of the gray level will drive the time response for the selected level while the high-order bits will be used for the choice of voltage pairs. The values of the voltages will be preset so as to obtain a uniform luminance response from one voltage to the next in the following. Since the voltage response of an electron source is close to the inverse of the gamma correction transfer function, the voltage staging can be established with successive differences relatively close to one another.

The method of the invention is not limited to display screens displaying video type images, it can be applied to the control of a display screen not displaying video signals such as personal computer screens. . In this case, there are signals to be applied to the columns which are linearly coded. The codes corresponding to the gamma correction can be extracted from a LUT conversion table in order to display all the shades of gray distinctly by reducing the number of samples (and therefore with longer sample durations).

Although several embodiments of the present invention have been shown and described in detail, it will be understood that different changes and modifications can be made without departing from the scope of the invention.

CITES DOCUMENTS

1. [1] "Fluorescent screens with microtips" R. Baptist (The electric wave, November-December 1991, volume 71, No. 6, pages 36-42).
2. [2] K. Sakai et al., "Flat panel displays based on surface conduction electron emitters". (Proceedings of the 16 ^th International Research Research Conference, ref.18.3L., Pp. 569-572).
3. [3] "Carbon Nanotubes FED elements" by S. Uemura et al. (SID 1998 Digest, pages 1052-1055).
4. [4] "Microtips displays addressing" by T. Leroux et al. (SID 91 Digest, pages 437-439).
5. [5] FR-A-2,633,764.
6. [6] EP-A-1 005 012.
7. [7] EP-A-0 635 819.
8. [8] ITU-R Recommendation BT.709, Basic Parameter Values for the HDTV Standard for the International Exchange Program (1990) [formerly CCIR Rec. 709]. (Geneva: ITU, 1990).
9. [9] "The rehabilitation of gamma" by Charles Poynton. (Human Vision and Electronic Imaging III, Proceedings of SPIE / IS & T Conference 3299, San Jose, Calif., Jan. 26-30, 1998 (Bellingham, Wash .: SPIE, 1998)).

Claims (10)

A method of controlling a matrix display screen with lines (li) and columns (cj) whose intersections form an image point (Pi, j), comprising applying to a line a selection voltage line (TLS) and simultaneously with a column of a voltage corresponding to a gray level (G) to be displayed at the associated image point, characterized in that , said gray level being chosen from 2 ^q levels and numerically coded according to a nonlinear law consistent with the perception of luminance by a human eye, the different voltages (Vci, Vci + 1) to be applied to the columns are chosen in a strictly increasing sequence of (2 ⁿ +1) voltages with n ≥1, these voltages being distributed in N = 2 ⁿ consecutive pairs of voltages, each pair making it possible to display a range of gray levels and in that the line selection time (Tls) is subdivided into one or more groups of 2 ^(qn) time intervals with (qn) integer> 1, each group having the same duration, the distribution of these 2 ^(qn) time intervals in a group being, for a pair of voltages, according to a nonlinear law presenting a transfer function close to the inverse of that previously used to code the gray levels for the corresponding gray level range, the voltage applied to a column, for a pair of voltages and a group of time slots, either taking one of the values of the voltage pair during the whole duration of the group, either switching at least once, during the duration of the group, at the end of a time interval, from one value to another of the pair of voltages. Process according to claim 1,
characterized in that the law of coding a gray level G is such that: $$\mathrm{BOY\; WUT}=\left({\mathrm{two}}^{\mathrm{q}}-1\right)\times 4,5\times {\mathrm{P}}_{\mathrm{BOY\; WUT}}$$

for P _G ≤0.018 and $$\mathrm{BOY\; WUT}=\left[{\left(1,099\times {\mathrm{P}}_{\mathrm{BOY\; WUT}}\right)}^{0,45}-0,099\right]\left({\mathrm{two}}^{\mathrm{q}}-1\right)$$

for P _G > 0.018, where P _G is a relative weight assigned to the gray level. Process according to claim 2,
characterized in that for a given pair of voltages, a range of gray levels, a given group of time intervals and a gray level to be displayed, the torque voltage leading to the highest luminance is applied during an interval of time given by:
Δ t = τ (P _G -P _Ginf ) / (P _Gsup -P _Ginf ) with τ duration of a group of 2 ^qn time intervals, P _G weight of the gray level to display, P _Gsup and P _Ginf weight of corresponding gray levels at the upper and lower limits of the range of gray levels associated with the voltage pair. Method according to one of Claims 1 to 3, characterized in that when the line selection time is subdivided into two groups of time intervals, the time intervals in the two groups are distributed symmetrically with respect to the middle of the line selection time. Method according to one of claims 1 to 4, characterized in that during the line selection time the line selection voltage is free of voltage transients. Method according to one of claims 1 to 5, characterized in that the screen is a flat screen with electron source. Device for controlling a matrix display screen according to the method of claims 1 to 6, characterized in that it comprises: a digital data source (20) capable of providing binary words encoded on q bits according to the non-linear law according to the perception of luminance by a human eye and representing the codes of the 2 ^q gray levels to be displayed, a screen controller (21) receiving synchronization signals from the data source (20) and managing signals suitable for driving a line scanning generator (22) and a column driving voltage generator (23) which receives for each column the codes of the gray levels to be displayed and which makes it possible to generate, from a discrete voltage generator (24), the voltages of the voltage torques, and to switch, if necessary, a voltage to the voltage other in a couple and in that the binary words are subdivided into two subwords one of n bits corresponding to the most significant bits and the other of qn bits corresponding to the least significant bits, the control voltage generator of the columns (23) comprising for each column a combinational logic stage (15) controlling a series (16) of analog switches (CA) to preselect a voltage pair delivered by the discrete voltage generator (24) from the n most significant bits of a binary word and a signal from a counter (13) initialized at each line selection time and to switch from one of the voltages to the other torque when the counter (13) has reached the value corresponding to the qn bits of low weight of the binary word. Control device according to claim 7, characterized in that the counter (13) receives a series of non-linearly distributed pulses corresponding to the torque of voltages coming from a pulse generator (11) connected to the screen controller via a multiplexer (12) also receiving in address the n most significant bits of the binary word delivered by the data source (20). Control device according to one of claims 7 or 8, characterized in that the combinational logic stage (15) receives the n most significant bits of the binary word delivered by the data source (20) via a shift register associated with memory latches (10). Control device according to one of Claims 7 to 9, characterized in that the combinational logic stage (15) is connected to the counter (13) via a comparator (14) which compares the signal from the counter ( 13) and the qn least significant bits of the binary word delivered by the data source (20).

Images