CN1755791A - Gamma correction method and circuit and image processing equipment and display device - Google Patents
Gamma correction method and circuit and image processing equipment and display device Download PDFInfo
- Publication number
- CN1755791A CN1755791A CN 200510104193 CN200510104193A CN1755791A CN 1755791 A CN1755791 A CN 1755791A CN 200510104193 CN200510104193 CN 200510104193 CN 200510104193 A CN200510104193 A CN 200510104193A CN 1755791 A CN1755791 A CN 1755791A
- Authority
- CN
- China
- Prior art keywords
- value
- tout
- bit
- tone
- signal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Abstract
A kind of gamma correction method, comprise: contrary gamma correction is handled, be used to import X bit tone input signal Tin value, and output has Y bit (the tonal signaling Tout of X＜Y) of the tone-light characteristic of the linear characteristic of being converted into non-linear tone-light characteristic; And format conversion processing, be used to import Tout, and when with the mantissa of the exponential sum M bit signal Tm indication of constant A, N bit signal Tn indication Tout being expressed as A^Tn * Tm, output is by one group of (N+M) bit tone output signal that forms near Tn and the Tm of Tout.
Description
Technical field
The present invention relates to a kind of gamma correction (gamma correction) method, program and equipment, use the image processing method of this gamma correction method, program and equipment, and display device.
Background technology
Have such as the various display devices of cathode ray tube (CRT), LCD (LCD) and plasma display panel (PDP) and to be every kind of peculiar intrinsic gamma characteristic of type of display and colorimetric properties.
The gamma characteristic of display is the relation between display input signals level and the output brightness.On the other hand, colorimetric properties are indicated the three primary colours of each display or the colourity of more primary colours (for example RGB+ white).Because general use three primary colours RGB in display, so colorimetric properties are generally indicated RGB colourity.
These characteristics depend on the type of display device, and therefore in CRT, LCD and each system of PDP, gamma characteristic is different to a great extent, and colorimetric properties are also different.
On the other hand, definite such as gamma characteristic of the received image signal of TV signal and sRGB (relation between signal level and the brightness that shows thus) and colorimetric properties by predetermined specifications.Therefore, carry out gamma correction processing and color space color space transformation and handle, so that between vision signal and display device, realize the coupling of gamma characteristic and colorimetric properties.
Consider the demonstration on the CRT, TV signal is subjected to the gamma correction and the chroma conversion that are complementary with the CRT gamma characteristic usually.Therefore, use in the television receiver of CRT at great majority, the CRT side does not need special gamma correction to handle.Yet, show for the TV signal on LCD or the PDP, must implement gamma correction (gamma characteristic of vision signal being converted to the processing of the gamma characteristic of display device) in general sense at least, to keep high image quality.The difference of gamma characteristic is greater than colorimetric properties.Because the difference of gamma characteristic relates to the generation of false contouring and the color and taste of medium tone (color taste) conversion, if therefore do not obtain coupling by the gamma correction of carrying out in general sense, then picture quality will worsen significantly.
At this, gamma correction in general sense is assumed that, be " the contrary gamma correction " that utilization is used for that linear tone-light characteristic is converted to " gamma correction (sense stricto) " of nonlinear characteristic and is used for non-linear tone-light characteristic is converted to linear characteristic, make the required a series of processing of gamma characteristic coupling of vision signal and display device.After this, in this manual, sense stricto gamma correction is represented as " positive gamma correction ", and gamma correction in general sense is " gamma correction " by direct representation.
And, abide by image and the picture (in other words) that the sRGB standard produces in order to reappear exactly for realizing corresponding color rendering, after contrary gamma correction, also need color space transformation.More particularly, need carry out 3 * 3 matrix operations to the rgb signal of input, disclosed in " Color Image Reproduction (coloured image reproduction) " (pp.33-39, formula 3.11) (list of references 3) as Tajima.
Yet, utilizing digital processing to carry out under the situation of contrary gamma correction processing, the output figure place must be greater than the input figure place, to keep its precision.If output figure place deficiency will be then because quantization error will cause deterioration of image quality (for example false contouring).
On the other hand, when the output figure place increases, the computing circuit that is provided with in the circuit scale of contrary gamma-correction circuit and the back one-level will become greatly the problem that causes the display device manufacturing cost to uprise with figure place (bit number).
And, in order to carry out corresponding color rendering, after finishing contrary gamma correction, need to carry out color space transformation.In this case, when the output figure place that produces when against gamma correction increases, the circuit scale of color space converting circuit also will increase.
Solve the background technology that figure place increases problem as being used to, JP 10-126648 (list of references 1) discloses " Input Signal Level Application Type GammaCorrection Circuit in Liquid Crystal Display Apparatus (the applied gamma-correction circuit of the incoming signal level in the liquid crystal display) ".List of references 1 relates to and a kind ofly uses less figure place and do not cause the gamma-correction circuit of deterioration of image quality in conversion process.
The applied gamma-correction circuit of incoming signal level has a kind of like this structure in the liquid crystal display of describing in the list of references 1, this structure comprises and lays respectively at before the digital gamma correcting circuit and afterwards modulus (AD) converter and digital-to-analogue (DA) converter, and comprises in addition and be positioned at before the AD converter and DA converter variable gain control amplifier afterwards.
The amplification degree of these variable gain control amplifiers has complementary relationship.List of references 1 changes amplification degree by the level according to analog input signal, has improved the tone characteristic of dark space well.Therefore, under the situation that does not increase figure place, realized the high precision gamma correction.
As other background technology,, in 741 (lists of references 2) example that utilizes floating number to come the tonal signaling of indicated number equipment has been described for example in U.S. Patent No. 5,528.
Yet list of references 1 is the fact of simulating signal based on input signal.Therefore, when input signal is digital signal, further need the DA converter, this makes circuit complicate equally.TV signal (high-definition television (HDTV) etc.) and picture signal (sRGB) almost are transfused to as digital signal, and often the combine digital picture signal is handled.Therefore, preferably adopt a kind of can be under the situation of not using AD converter and DA converter, realize that the numeral of contrary gamma correction of enough accuracy is against the gamma correction treatment circuit.
List of references 2 discloses a kind of method that effectively pixel value of floating point representation is converted to a byte integer value.In computer graphics was used, more convenient was to use floating point representation in calculated for pixel values is handled.Yet,,, show it is impossible if therefore the pixel data of floating point representation is not converted to integer representation because nearly all display device all is treated to integer to the pixel data (pixel value) that is used for the pixel demonstration.
Therefore, in list of references 2, the data of floating point representation are converted into integer representation.And, when being converted to integer representation, also carry out gamma correction.Yet this structure has realized the gamma correction processing of simplification, but still is included in the unresolved problem of gamma correction figure place increase afterwards.Therefore, after gamma correction, carry out under the situation of some arithmetical operation, can not realize that the precision raising and the scale of back one-level circuit reduces.
Summary of the invention
Consider the problems referred to above and proposed the present invention, therefore the objective of the invention is, a kind of gamma-correction circuit that is used to realize the coupling between input signal gamma characteristic and the display device gamma characteristic is provided, and be used for carrying out gamma correction processing or contrary gamma correction processing with enough precision, the gamma correction method of the scale of the circuit of controlling computing circuit oneself simultaneously and being provided with in the one-level thereafter, and might provide a kind of gamma correction program and equipment, use the image processing method of this gamma correction method, program and equipment, and display device.
For achieving the above object, as first aspect, the invention provides a kind of gamma correction method, this gamma correction method comprises: contrary gamma correction is handled, be used to import X bit tone input signal Tin value, and output has Y bit (the tonal signaling Tout of X＜Y) of the tone-light characteristic of the linear characteristic of being converted into non-linear tone-light characteristic; And format conversion processing, be used to import Tout, and when using constant A, when the mantissa of the exponential sum M bit signal Tm indication of N bit signal Tn indication is expressed as A^Tn * Tm with Tout, output is by one group of (N+M) bit tone output signal that forms near Tn and the Tm of Tout, wherein, the value of constant A is 2 the power of several naturally, M satisfies relational expression M≤Y-N, the minimum value of N is 2 when A equals 2, when A minimum value of N greater than 2 time is 1, and the maximal value of the N that determines according to constant A and M is to make with respect to all Tin values, and one group of Tn and Tm value all become the maximal value of one group of different value.
According to the present invention, a kind of gamma-correction circuit that is used to realize the coupling between input signal gamma characteristic and the display device gamma characteristic can be provided, and be used for carrying out gamma correction processing or contrary gamma correction processing with enough precision, the gamma correction method of the scale of the circuit of controlling computing circuit oneself simultaneously and being provided with in the one-level thereafter, and provide a kind of gamma correction program and equipment, use the image processing method of this gamma correction method, program and equipment, and display device.
Description of drawings
Figure 1A and 1B utilize digital line to show before the contrary gamma correction and signal value afterwards and the tone that distributed between graph of a relation.
Fig. 2 has shown the processing of the gamma correction method that the present invention was suitable for.
Fig. 3 has shown when there being a pair of the degenerate method of (recess) of tone that is used to during near the Tout ' of Tout to prevent.
Fig. 4 has shown the relation between M, N value and the output signal accuracy;
Fig. 5 has shown and has been used for illustrating that the input-output table of handling from the contrary gamma correction of gamma correction method obtains the input-output table of the method for suitable M, N value.
Fig. 6 has shown the structure of the liquid crystal display of first embodiment of the invention.
Fig. 7 has shown the structure of the contrary gamma-correction circuit of numeral of the liquid crystal display of first embodiment.
Fig. 8 has shown the structure of color space converting circuit of the liquid crystal display of first embodiment.
Fig. 9 has shown the structure of positive gamma-correction circuit of the liquid crystal display of first embodiment.
Figure 10 has shown the structure of the LUT that is provided with in the positive gamma-correction circuit of liquid crystal display of first embodiment.
Figure 11 has shown the structure of using the display device of the display system that is different from liquid crystal.
Figure 12 has shown the structure of the liquid crystal display of second embodiment of the invention.
Figure 13 has shown the structure of the contrary gamma-correction circuit of numeral of the liquid crystal display of second embodiment.
Figure 14 has shown the structure of color space color space converting circuit of the liquid crystal display of second embodiment.
Figure 15 has shown the structure of the liquid crystal display of third embodiment of the invention.
Figure 16 has shown the structure of the contrary gamma-correction circuit of numeral of the liquid crystal display of the 3rd embodiment.
Figure 17 has shown the structure of the brightness-contrast correction circuit of the liquid crystal display of the 3rd embodiment.
Figure 18 has shown the structure of the liquid crystal display of fourth embodiment of the invention.
Figure 19 has shown the relation between the output voltage of the input-output voltage of DAC in Source drive-signal generator and reference color tonal signal generator.
Figure 20 has shown when Tn is 3 (" 11 ") relation between the output voltage of the input-output voltage of DAC and reference color tonal signal generator in Source drive-signal generator.
Figure 21 has shown when Tn is 2 (" 11 ") relation between the output voltage of the input-output voltage of DAC and reference color tonal signal generator in Source drive-signal generator.
Embodiment
Video standard signal and television image signal such as sRGB, scRGB generally are subjected to gamma correction.The tone value that has been subjected to the picture signal of positive gamma correction has the nonlinear characteristic with respect to brightness:
(tone value)=(brightness) ^ (1/ γ)
As above illustrated, in order to carry out corresponding color rendering, need 3 * 3 matrix operations.Yet, when directly the signal that passes through positive gamma correction being carried out matrix operation, comprise that the tone of the image of a large amount of complementary colors such as cyan, magenta and yellow and medium tone will change.
Therefore, before carrying out matrix operation, must carry out contrary gamma correction, so that the linear tone value with respect to brightness is provided.
Yet because input figure place (bit number) is set to larger than the output figure place, so that the contrary gamma correction of combine digital, the change of the matrix operation circuit scale of one-level is big after the institute.
As above illustrated, the contrary required figure place increase of gamma correction caused allowing circuit scale increase, to realize the reason of corresponding color rendering.
Therefore, in the present invention, must obtain contrary gamma correction and handle tone precision afterwards, come the increase of control circuit scale simultaneously by tone value with the method representation picture signal that is different from background technology.More particularly, use based on the method for floating point representation and represent tone value.
To illustrate now the tone that utilizes this method represent the definition.In conventional method, utilize 8 bits that 8 hue table are shown 0 to 255 tone.On the other hand, this method is represented tone with the 2^Tn * Tm form of similar floating point representation.At this, Tn is the number that is called as the index in the floating point representation, and Tm is the number that is called as the mantissa in the floating point representation.
Be with the difference of common floating point representation, do not need to make position to add " 1 " than the high one digit number of highest significant position of mantissa.Under the situation of common floating point representation, determine to make position to add " 1 " than the high one digit number of highest significant position of mantissa, become impossible to use the identical value of Tn and Tm combination (2^1 * 38=2^2 * 19) expression.
Yet, adopt a kind of direct representation that adds " 1 " than the position of the high one digit number of highest significant position of mantissa that do not make in the method.Yet as possible, just highest significant position is set to " 1 ".
For example, under the situation of 2^2 * 19 because mantissa (19: decimal number=010011: highest significant position binary number) is " 0 ", and it is expressed as 2^1 * 38, makes highest significant position become " 1 ".In addition, 2^0 * 19 are represented as former state, and this is because the highest significant position of mantissa is 0, but not than 0 little index.
For example, be represented as at tone under the situation of 2^Tn * Tm, might obtain:
0＝2^0×0≤2^Tn×Tm≤2^3×63＝504
And when the field of definition of Tn is set to 0 to 3 (2 bit), when the field of definition of Tm is set to 0 to 63 (6 bit) simultaneously, represent 0 to 504 value with 8 bits.
For ascending power computing, adopt this tone method for expressing especially effective such as contrary gamma correction.And, by after contrary gamma correction, introducing tone, can reduce the circuit scale of 3 * 3 matrix operations and positive gamma correction based on floating point representation.
Use based on the method for floating point representation and carry out the advantage that tone is represented:
Now the advantage that this method tone is represented will be described.In contrary gamma correction, can be with than the high precision of common tone method for expressing, more effectively assigned hue based on the method for floating point representation.Figure 1A and 1B utilize digital line to show before the contrary gamma correction and signal value afterwards and the tone that distributed between relation.For the purpose of simplifying the description, input is taken as three (3) bits, and output is taken as six (6) bits (in the method based on floating point representation, index is that two (2) bits and mantissa are four (4) bits).Gamma value is 2.2 of general use.
When carrying out against gamma correction, input value is converted into than little value before proofreading and correct.Shown in the digital line in the intergrade, the value after the contrary gamma correction is assigned to one of tone of existing expression and floating point representation.In the digital line of next stage, indicate with perpendicular line, can use based on existing six (6) bits and represent numeral with the method representation of six (6) bit floating point representations.
Under the situation of existing expression, when advancing to darker tone one side, the digital number that can be expressed that exists between the numeral that the input value after the contrary gamma correction is given has tailed off.This true hint when and the operational precision matrix operation that will in back one-level carry out dark when tone becomes low, is easy to generate quantization error.
On the other hand, in the method, the digital number that will be expressed that exists between the numeral that the input value after the contrary gamma correction is given much at one, and is and irrelevant with the darkness of tone.This true hint, quantization error diminish and the operational precision of the matrix operation that will carry out in the one-level of back has uprised.
More particularly, when the Q value is equal to or greater than one (1) in such as contrary (output)=(input) the ^Q computing of gamma correction, uprised based on the operational precision of the method for floating point representation.
In addition, because according to the formula of the L* that is used to obtain L*a*b color display system (CIE 1976), human vision property demonstrates the linear luminance characteristic that equals brightness 1/3 power, therefore when carrying out contrary gamma correction, need improve the precision of low key tone.In other words,, therefore compare, can reduce this aberration by reducing the low key tone quantization error in the floating point representation with existing expression because quantization error directly is revealed as aberration.
As above illustrated, also can understand from the viewpoint of visual characteristic, aspect complete representation, this method more preferably because aberration has diminished.
Also might be by according to required operational precision exponential sum mantissa being reduced to two (2) bits and three (3) bits diminish circuit scale, this two (2) bit and three (3) bits obtain by deducting 1 bit respectively.
The list of references of introducing in the background technology 2 discloses the gamma correction method that a kind of pixel value with floating point representation is converted to a byte pixel value.As above illustrated, exist a kind of like this gamma correction to handle really, wherein provide input with the floating point representation value, export integer.Yet, floating point representation input data in the list of references 2 are to utilize the computing in the field of Computer Graphics to produce, and supposition is not handled the round values data of utilizing acquisitions such as ordinary numbers camera and scanner after floating point representation is changed.The invention is characterized in and be used for the integer table indicating value is converted to conversion method based on the value of the method for floating point representation.Therefore, list of references 2 can not be realized " the computing circuit precision in the back one-level improves and circuit scale reduces " as the object of the invention.
Advantage based on the method for floating point representation more than has been described.Yet, not necessarily can obtain effect of the present invention by introducing said structure.That is, need be according to the input-output table of actual contrary gamma correction, A, Tn in A^Tn * Tm method for expressing and the scope of Tm value suitably are set.Can not utilize expectation A, Tn and Tm value in the floating point representation that illustrates in the list of references 2 to obtain effect of the present invention.
Below the method that these values are set will be described.
Fig. 2 has shown the flow process of the contrary gamma correction method that the present invention was suitable for.Should be used as input signal to the X bit signal by contrary gamma correction method, and the N+M bit signal was used as output signal.
Should comprise by contrary gamma correction method: contrary gamma correction is handled, and is used to import X bit tone input signal Tin value and exports Y bit (the tonal signaling Tout of X＜Y); Format conversion processing is used to import Y bit tonal signaling Tout, and output becomes and immediate Tn of Tout and Tm when with the Tm of mantissa of the index Tn of constant A, N bit and M bit tonal signaling Tout being expressed as A^Tn * Tm; And the processing that is used to produce the output signal of forming by N bit Tn and M bit Tm.
In this contrary gamma correction method, input X bit tone input signal Tin (step S1).The tone input signal is subjected to contrary gamma correction (step S2).At this, in contrary gamma correction was handled, the Tin with non-linear tone-light characteristic was converted into linearity, is output then, as Tout.
Therefore at this,, be enough to obtain to be used for tone-light characteristic with Tin and be converted to Tout based on the linear characteristic of this standard because tone input signal Tin is sent out in accordance with for example sRGB standard.Suppose, utilize wherein conversion table such as input value Tin and output valve Tout look-up table (LUT) commonly used to be stored in structure in the storer, carry out contrary gamma correction processing.In a word, carry out contrary gamma correction and handle, so that the nonlinear characteristic of Tin is converted to linear characteristic.
Handling among the Y bit signal Tout that obtains by contrary gamma correction, Y is usually greater than X (step S3).Its reason is that as mentioned above, in contrary gamma correction is handled, if the output bit number is not more than the input bit number, then precision will reduce.
After this, conversion Tout signal (step S4, S5) in format conversion processing.In format conversion processing, the Tout value is converted into N bit signal Tn and M bit signal Tm.
At first, in the value that in format conversion processing 1, obtains to represent with formula A^Tn * Tm near the Tout ' value (step S4) of Tout.When having two during near the Tout ' value of Tout, in other words, when two values of Tout ± α can be represented as A^Tn * Tm, can be with any one value among Tout+ α and the Tout-α as Tout '.Yet, preferably definite rule of selecting used value.Its reason is, when be provided with as shown in Figure 3 relational expression Tout (x)+β=Tout ' (X), Tout ' (X)+β=Tout (X+1), Tout (X+1)+β=Tout ' (X+1), Tout ' (X+1)+during β=Tout (X+2), if Tout (X)+β=Tout ' (X) as (X+1) being used as and the corresponding value of Tout (X+1) with the corresponding value of Tout (X) and Tout (X+1)-β=Tout ', has then been produced tone degeneration (recess).Yet, when (X) being used as Tout (X)+β=Tout ' and (X+1) being used as Tout (X+1)+β=Tout ', degenerate even also never produce tone in this case with the corresponding value of Tout (X+1) with the corresponding value of Tout (X).
At this, constant A must be 2 the power of several naturally.When A is set to 2 natural number power, can use multiplier circuit and only utilize data shift to carry out the multiplication of A^Tn and Tm.If A is not 2 natural number power, then can not reduce the computing circuit scale in the contrary back one-level of gamma-correction circuit, lost the advantage that adopts floating point representation.For example, when attempting to be defined as six (6) bits in mantissa, when A is defined as carrying out under the situation of 3 bits 3^2 * 33+3^1 * 25 additions, need to carry out the multiplication of 3^2 * 33=3^1 * (3 * 33), with the coupling between the Tn value that obtains first Tn value and second.And, after the addition of the Tm of mantissa (99+25=124), need to carry out division (124/3), so that the mantissa of mantissa is set to six (6) bits once more.As mentioned above, because addition needs multiplication and division, therefore can not obtain obtainable effect when adopting floating point representation.
In addition, desirable is that the M value is equal to or less than Y-N.Its reason is, gamma correction is handled through being usually used in keeping precision, and the circuit scale of LUT can be kept or reduces by adopting contrary gamma-correction circuit of the present invention.In order to reduce the circuit scale of back one-level circuit, the figure place of mantissa must promptly require relational expression M＜Y less than the figure place of initial integer representation at least.
In addition, when A is 2, require the N value to be equal to or greater than 2, and require the N value to be equal to or greater than 1 greater than 2 the time as A.Its reason is, if the N value is taken as 1 when A is 2, then the A^Tn value only gets 1 or 2, and can be set to the M+1 bit by Tm and realize the displacement of this one (1) bit data.In addition, when A is equal to or greater than 4,, can not realize two (2) the Bit datas displacement of A^Tn even Tm is set to the M+1 bit.Therefore, it is rational adopting N=1.
At this, the N value is not enough to get maximal value.Can obtain the maximal value of N from Tout ' value.When N becomes big, must from relational expression Y 〉=M+N, deduct the M value.When the N value became big, the tone dynamic range that can represent became wideer, but because the M value has diminished, so precision has reduced.When precision reduces, can not obtain identical Tout ' value from other Tin value.When showing, this is considered to tone degeneration (recess) problem, and becomes away from theme of the present invention.Therefore, from the above description, if utilize the default constant A and the figure place M of mantissa, for all Tin values, one group of Tn and Tm value all become one group of different value, and then impossible use is than the big value of maximum N value.
From the above description, A, M and N value are restricted, and can be in format conversion processing 2 be worth according to these and obtain Tn and Tm value (step S5).At this, Tn and Tm value are set like this, make the Tm value become big as much as possible.Its reason is that when the Tm value became big, the precision of tone had improved generally.
Figure 4 shows that, tone-precision characteristic variation synoptic diagram of the tone-precision characteristic when increasing under the situation that (when M=m2＞m1) and N value increase (when N=n2＞n1) with respect to M=m1 and N=n1 in the M value, wherein transverse axis represents that tone and Z-axis represent precision (on duty big more, precision is high more).At first, with explanation M=m1 as a reference and the tone-precision characteristic during N=n1.In this case, because floating number can not be shifted in to less than 0 (working as N=0) in the tone value small range, therefore when tone value reduced, it was covered by the Tm value, thereby precision reduces.When tone becomes higher, can use the precision of Tm to greatest extent.Therefore, precision becomes constant (M figure place).Next, when the N value increases (when N=n2＞n1), floating number is shifted in wideer tone range.Therefore, except that the scope that tone only gets the small value, precision becomes and is almost constant.Yet because M value constant (M=m1), so the precision of the precision of tone value when getting higher value during with N=n1 is identical.Simultaneously, when the M value increases (as M=m2＞m1), all improve (m1-m2) figure place for all tone values, precision.Therefore, preferably, make the Tm value preferentially greater than the Tn value.
Yet, if when tone is low, require operational precision, preferably improve with respect to operational precision than low key tone, be set to a certain value according to required precision Tn value simultaneously.
At last, Tn that is obtained and Tm are output as output signal (step S6).
Can carry out these steps with single processing by only summing up these steps with a LUT.
Y 〉=M+N utilizes said structure, because so compare with existing structure, can reduce contrary gamma correction and handle required LUT scale.In addition, since Y＞M, the multiplier circuit computing scale in the time of therefore can reducing in the back one-level (handling) that contrary gamma correction is handled, to carry out multiplication corresponding to corresponding color rendering.In addition, these processing have produced no better than or have been higher than the precision of the existing structure that is used to provide the output of Y bit.That is, utilize contrary gamma correction of the present invention, picture quality never worsens.
Otherwise, under the situation of positive gamma correction, from the above description, can be by (M+N≤Y) tonal signaling Tout is expressed as the Tout=A^Tn * Tm (A is a constant) that has Tn and Tm the Y bit with linear tone-luminance accuracy, (the tonal signaling Tin of X＜Y) reduces positive gamma correction and handles required LUT scale then Tout to be converted to X bit with non-linear tone-light characteristic.
The A, the M that are used to provide effect of the present invention and the scope of N value more than have been described.Next, suitable M and the N value that the contrary gamma correction of research is handled.
In this example, for convenience of description, the A value is set to 2.Yet,, also can obtain suitable M and N value by utilizing and following identical method change value even A gets other value (2 the powers of several naturally).
The Tout value that is obtained from contrary gamma correction is handled can obtain suitable M and N value.Below with under X=6 in the key diagram 5 and the Y=10 situation, the example of the input-output characteristic that contrary gamma correction is obtained in handling.Figure 5 shows that under the condition of X=6 and Y=10, the figure place M of mantissa is set to 5 and at 6 o'clock, the output Tn of format conversion processing and the comparative result table between the Tm.By represent to utilize contrary gamma correction to handle the input-output characteristic that obtains with monotonic increasing function, so that eliminate the generation of tone reversal (phenomenon of the brightness of intrinsic signal and the counter-rotating of the relation of darkness).
At first, near the Tout value of low key tone (minimum value) in the research Tin value.Because the Tout value is converted into the value less than input value in contrary gamma correction is handled, the precision that therefore has the Tin of smaller value has reduced, as shown in Figure 1.Therefore, when Tin when to about 3 scope, changing from 0, judge that from Tout contrary gamma correction handles required precision (that is output figure place handled of contrary gamma correction: Y).In Fig. 5, because the value of Tout value is 0,1,2 and 4, and lowest difference is 1, therefore when using other input value, never gets identical value (promptly never producing tone degeneration (recess)).Therefore, as what can estimate, (Y) value can be got 10 values.
Next, M and N value are set like this, make that becoming equals one (1) or bigger by the Tout that obtains from all Tin values being carried out the poor of the adjacent tone of changing the Tm value that obtains.More particularly, research is the poor of one (1) minimum Tout with highest significant position.When Y=10, the Tout value that the Tout highest significant position becomes one (1) (promptly become and be equal to or greater than 2^9=512) for the first time equals 520 (1000001000), as shown in Figure 5, and the Tout increase, as 546 (1000100010), 571 (1000111011) and 598 (1001010110).In this case, when the M value equaled 5, the Tm value that obtains from these Tout values became 10000,10001,10001 and 10010 respectively.In this case, the Tm of Tout=546 and the Tm of Tout=571 become 10001, and identical value is continuous, thereby have produced tone degeneration (recess).Simultaneously, when the M value equaled six (6), the Tm value that obtains from these Tout values became 100000,100010,100011 and 100101 respectively, dissimilated with the corresponding Tm value of each Tout value.Therefore, when from low key tone when light tone, precision all equate, suitable M value is six (6), and must the Tn value be set to from 0 up to 4.In order under the condition of M=6, A=2 and Y=10, Y bit (10 bit) data to be expressed as A^Tn * Tm, require the Tn that 2^Tn * 63 ≈ 2^10-1 are set up.Because Tm equals 10010 when Tout equals 546 (1000100010), thus can be represented as A^Tn * Tm and equal 544 (=2^4 * 34) near 546 value, and the value that requires Tn is up to 4.
In addition, also require to satisfy relational expression Tn=2^ (2^N-1) 〉=2^ (Y-M), this is because the data shift by removing the residue Y-M bit that the higher significance bit of M bit obtains from initial Y Bit data is to represent with N.When N is found the solution, can obtain N 〉=log (1+Y-M)/log2 result.Therefore, when the M value is determined, can obtain as the N value that is equal to or greater than the smallest positive integral value of log (1+Y-M)/log2.In above example, because log (1+10-6)/log2=2.322, so the N value becomes 3 as minimum value.
In addition, increase M or N value, to obtain higher precision.For example, because the precision of light tone is mainly definite by M, therefore be enough to increase the M value.In addition, because the precision of low key tone is also relevant with the N value with M, therefore be enough to increase M or N value.
Utilize said method, can suitable M and N value be set with the value that contrary gamma correction is handled.
The preferred embodiment of the present invention based on above-mentioned principle below will be described.
First embodiment
First preferred embodiment of the present invention will be described now.Fig. 6 has shown the structure of the liquid crystal display of this first embodiment.This liquid crystal display is equipped with image processor 1 and LCD 12.This liquid crystal display is with data image signal (rgb signal for example, each color six bit) is input to image processor 1 as input signal, in image processor 1, carry out after the various operations, data image signal is outputed to LCD 12, determine the transmissivity of the pixel of LCD 12 according to data image signal, and come display image according to this data image signal.
At this, represent with the tone of each color 6 bit, come the input-output signal of presentation video processor, but figure place is not limited to six (6) bits.
At this, below will illustrate from data image signal is input to liquid crystal display, up to the processing of display image on LCD 12.Image processor 1 is exported digital signal by the supplied with digital signal executable operations is handled.Output image signal and display control signal (not shown) are sent to scanner driver 33 and Source drive and control-signals generator 34.Source drive and control-signals generator 34 be according to the voltage-light characteristic that applies of the pixel 35 of LCD 12, and from the conversion characteristic that the gamma characteristic of received image signal obtains, data image signal is carried out the DA conversion.The signal that is converted into simulating signal is applied to optionally applying the pixel 35 that the sweep trace 31 of ON voltage is connected with being scanned driver 33 by TFT 37, by being converted into brightness, is outputted as picture signal then.
Fig. 7 has shown the detailed structure of contrary gamma-correction circuit 2.Should constitute by index look-up table (LUT) 21 and the LUT of mantissa 22 by contrary gamma-correction circuit 2.Index LUT 21 receives identical digital signal as input signal with the LUT22 of mantissa.In these indexes LUT 21 and the LUT22 of mantissa, obtain value with reference to LUT according to input signal, this value is output as the output digital signal then.The output of index LUT 21 is three (3) bit values (exponential signal), and the output of the LUT of mantissa 22 is six (6) bit values (mantissa signal).Signal from index LUT 21 and the LUT of mantissa 22 outputs is sent to color space converting circuit 3 respectively.
At this, index LUT and the LUT of mantissa are formed independent structure, but the LUT of one nine (9) bit output also might only be provided, and wherein the output of six (6) bits is provided for mantissa, is provided for index and remain three (3) bits.
Fig. 8 has shown the detailed structure of color space converting circuit 3.Color space converting circuit 3 is made up of following: nine multiplier circuits are used for the multiplication of RGB mantissa signal and three constants (axy (x, y=1,2,3)); Data shift device 23 is used for according to the RGB exponential signal corresponding multiplication result being carried out data shift; Six (6) individual totalizers are used for the data shift result is added up; And signal converter 24, be used for once more addition results being divided into exponential sum mantissa.Obviously, from shown in structure, color space converting circuit 3 is carried out 3 * 3 matrix operations.
Fig. 9 has shown the structure of positive gamma-correction circuit 4.Positive gamma-correction circuit 4 is made of two-dimentional LUT 25, and wherein exponential signal and mantissa signal are defined as line number and the columns in the secondary two-dimensional arrangements respectively.Positive gamma-correction circuit 4 obtains the value according to input exponential signal and input mantissa signal with reference to this two dimension LUT 25, should be worth then as output signal (six (6) bits) and export.
Next, with the operation of liquid crystal display of this embodiment of explanation and the effect that is reached.
In the image processing circuit of Fig. 6, represent as output by adopting against gamma-correction circuit 2 based on the tone of floating number system, can under the situation that does not reduce precision, reduce the circuit scale of the color space converting circuit 3 in the one-level of back.To describe the operation of each circuit in detail along the flow process of picture signal now.
Contrary gamma-correction circuit 2 shown in Figure 7 utilizes LUT to carry out contrary gamma correction, and converts the tone method for expressing to based on floating point representation method.
Six (6) bit signals are being carried out under the contrary situation of gamma correction, if the output after proofreading and correct is set to 10 bits or bigger (2^ (9-1)=512 tone or more), then represent that with common tone (fixed point) can not keep precision, but utilize method based on floating point representation, can be set to three (3) bit indexes (only being decimal numeral 0 to 4)+six (6) mantissa bits by output, obtain to be equal to or greater than the value of 10 (2^4 * 63=1008 tone) showing with fix point method.Just because can ten bits that common representation is required reduce to nine (9) bits (=3+6), therefore compare with the circuit that the signal that is used for fix point method is represented is carried out contrary gamma correction, the physical dimension of contrary gamma-correction circuit 2 can reduce.
In the back one-level of each mlultiplying circuit, different with common matrix operation circuit, utilization shown in Figure 8 represents that based on the method for floating point representation the matrix operation circuit (color space converting circuit 3) of tone is equipped with a kind of circuit that is used for making according to exponential quantity data shift.When the index of input signal is defined as Xa, when mantissa is defined as Xb (X=R, G, B), represents from signal is input to color space converting circuit 3, the processing of handling in signal converter 24 with following formula (1).
Rout＝2^Ra×(Rb×a11)+2^Ga×(Gb×a12)+2^Ba×(Bb×a13)，
Gout＝2^Ra×(Rb×a21)+2^Ga×(Gb×a22)+2^Ba×(Bb×a23)，
Bout＝2^Ra×(Rb×a31)+2^Ga×(Gb×a32)+2^Ba×(Bb×a33)(1)
In data shift device 23,, carry out the computing (X=R, G, B) in 2^Xa the formula (1) by the Xa bit shift.Then, this computing output is output by being converted to exponential sum mantissa by signal converter 24.Because data shift device 23 is carried out exponent arithmetic, therefore can short-cut multiplication handle.Therefore, when data shift device 23 performed exponent arithmetics for example are 2 ascending power computings, as 4^Ra doubly and 8^Rb times, can obtain effect of the present invention.
At this, because the figure place of mantissa equals to import figure place, so can significantly reduce the circuit scale of multiplier circuit.In ordinary construction, 10 bits * 6 bit computings are performed, but this computing is converted into six (6) bits * six (6) bit computings, causes circuit scale to reduce about 40%.And, because the circuit scale of data shift device 23 and signal converter 24 is not too big, so circuit structure can reduce about 30% than ordinary construction.
Next, the two-dimentional LUT (two-dimentional LUT 25) that utilizes wherein exponential quantity and mantissa value to be set to the xy axle respectively carries out the positive gamma correction based on floating point representation.Figure 10 has shown the LUT structure of positive gamma-correction circuit, wherein import by three (3) bit indexes (only using 0 to 4)+six (6) mantissa bits to constitute, and output is made of six (6) bits.Because the scope of exponential quantity is 0 to 4, and the scope of mantissa value is 0 to 63, so the required bit number of LUT reaches 1920 bits (=5 * 64 * 6 bit).Yet when exponential quantity is equal to or greater than one (1), and mantissa's highest significant position is 0 o'clock, makes exponential quantity subtract one (1) and mantissa value is doubled.Therefore, not necessarily need for all combinations of exponential sum mantissa all preserve LUT (do not need the territory that dual frame surrounded among Figure 10 (exponential quantity be equal to or greater than one (1) and mantissa's highest significant position be 0 territory)).
As shown in figure 10, when index equaled " 0 ", any value in " 0 to 63 " was got by mantissa, and when index equaled " 1 to 4 ", any value in " 32 to 63 " was got by mantissa.Because under any circumstance output all is six (6) bits, so the scale of positive gamma-correction circuit 4 can become as described below.
When index equaled " 0 ", " 0 to 63 " value, i.e. 2^6 kind value can be got by mantissa.Because for each value output all is six (6) bits, so need the storer of 2^6 * 6 bits.Simultaneously, when index equaled " 1 to 4 ", " 32 to 63 " value, i.e. 2^5 kind value can be got by mantissa.Because for each value output all is six (6) bits, so need the storer of 2^5 * 6 * 4 bits.That is the storer that, all needs (2^6 * 6) bit+(2^5 * 6 * 4) bit for every kind of color.Under the situation of RGB coloured image, all need and above-mentioned as many figure place (yet, suppose that the input-output figure place of every kind of pigment of RGB is identical) for every kind of pigment.Therefore, the required figure place of LUT is expressed as follows.
(RGB)=152 * 3, (2^6 * 6+2^5 * 6 * 4) * 3 bit
Under the situation of background technology, positive gamma-correction circuit is converted to 6 bit output signals with 10 bit input signals, and therefore required figure place is 2^10 * 6 * 3=6144 * 3 bits.Therefore, can reduce circuit scale greatly in the present invention.
At this, can be by using method, the circuit scale of matrix operation circuit (adder circuit, multiplier circuit) after reducing in the one-level and positive gamma-correction circuit (LUT) based on floating point representation.
Utilize above-mentioned effect, a kind of image processing equipment and display device that reduces circuit scale by the Applied Digital gamma-correction circuit might be provided.
In above structure, contrary gamma-correction circuit is combined by inverse conversion circuit and format conversion circuit and forms.This embodiment is not limited thereto, and when contrary gamma-correction circuit and format conversion circuit during by form separately, also can obtain effect of the present invention.
In the above description, the figure place of digital tone is set to six (6), but the invention is not restricted to this, and the figure place that is set to expect when this digital tone figure place for example during eight (8) bits, also can obtain identical effect.
In addition, in the present invention, LCD is used as display, but the invention is not restricted to this.Even ought use other display device (PDP, the display that uses electroluminescent cell or liquid crystal projection apparatus etc.) as shown in figure 11, also can obtain similar effects.
Second embodiment
Below the second embodiment of the present invention will be described.
Figure 12 has shown the structure of the liquid crystal display of second embodiment.This liquid crystal display almost image processor 1 with the liquid crystal display of first embodiment is identical, but contrary gamma-correction circuit 2A, color space converting circuit 3A are different with the liquid crystal display of first embodiment with the structure of positive gamma-correction circuit 4.Except that export six (6) Bit datas with respect to two (2) bits+six (6) bits input data, positive gamma-correction circuit 4 is the same with first embodiment.
Figure 13 has shown the structure of contrary gamma-correction circuit 2A.Contrary gamma-correction circuit 2A comprises index LUT 21A and the LUT of mantissa 22.The LUT of mantissa 22 is similar to the LUT of mantissa of the liquid crystal display of first embodiment.The index LUT part that index LUT 21A is different from first embodiment is that for every kind of color, two (2) bit values are output to six (6) bit received image signals.The same with first embodiment, identical signal is imported into index LUT 21A and the LUT of mantissa 22.
In first embodiment, three (3) bit values of exporting as index LUT 21 are used directly as the index that tone is represented.Simultaneously, in this second embodiment, the output of index LUT 21A directly is not used as the index that tone is represented, and it is encoded to (compose and give) other value (for example 0,2,3,4 replace 0,1,2,3) based on output valve.Promptly, the output item number of index LUT 21A equals the number of the value that the output valve of index LUT 21A can get, and the output of index LUT 21A is assigned to by any one value that increase progressively from initial term according to equal difference not, that be equal to or greater than 0 the formed sequence of integer.
By encoding, can be set to Ra greater than the value of the maximal value " 3 " of LUT 21A output, make when two (2) bits of index LUT 21A are output as 0 Ra=0, when this was output as 1, Ra=2 was when this is output as 2, Ra=3, and when this is output as 3, Ra=4.When Ra was 1, the magnitude portion output valve changed, and made Ra become 2 (2^1 * 18=2^2 * 9).
By carrying out this coding, can realize in wideer dynamic range that tone represents, the output figure place of index LUT 21A is controlled simultaneously.Because the output figure place is controlled, therefore can reduce the circuit scale of index LUT 21A.
The exponent data that is assigned to other value is sent to color space converting circuit 3A.Figure 14 has shown the structure of color space converting circuit 3A.This circuit almost color space converting circuit 3 with the liquid crystal display of first embodiment is identical, except data shift device 23A surrogate data method shift unit 23.The corresponding deciphering of coding (making the LUT 21A output valve that is converted into other value reset to initial value) among data shift device 23A execution and the index LUT 21A.When enough hour of the scale of the circuit that is used for this decoding, also can reduce the way circuit scale.In other words, when the scale of the circuit that is used to change the index LUT 21A output valve that is assigned to other value less than when output valve being composed the circuit scale of the LUT 21A that reduces to other value, also can reduce the way circuit scale.
Utilize said structure, can reduce to be arranged in the color space converting circuit (matrix operation circuit (adder circuit and multiplier circuit)) of one-level behind the contrary gamma-correction circuit and the circuit scale of gamma-correction circuit (LUT) just.
In addition, by should this digital gamma correcting circuit, also can reduce the circuit scale of image processing equipment and display device.
The 3rd embodiment
Below the third embodiment of the present invention will be described.
In first and second embodiment, image processing circuit is set in the back one-level of digital gamma correcting circuit, is used for corresponding color rendering.In this embodiment, a kind of like this structure will be described, the image processing circuit that wherein is used for the purpose different with corresponding color rendering is set at the back one-level of contrary gamma-correction circuit.
Figure 15 has shown the structure of the liquid crystal display of the 3rd embodiment.This liquid crystal display almost liquid crystal display with first embodiment is identical, and its difference only is that contrary gamma-correction circuit 2B replaces contrary gamma-correction circuit 2, and brightness simultaneously-contrast correction circuit 5 replaces color space converting circuit 3.
Figure 16 has shown the structure of contrary gamma-correction circuit 2B.The same with the contrary gamma-correction circuit 2 of first embodiment, identical picture signal is imported into index LUT 21B and the LUT22B of mantissa.Index LUT 21B is with respect to six (6) bit input signals output, three (3) exponential signals, and the LUT 22B of mantissa is with respect to six (6) bit input signals output, six (6) mantissa bits.
The RGB picture signal that is input to image processing circuit 1 is divided into exponential sum mantissa in contrary gamma-correction circuit 2B, and is output as exponential signal and mantissa signal.
In contrary gamma-correction circuit 2B, can carry out the contrary gamma correction of gamma value 1.Be different from one (1) gamma value in utilization and carry out under the contrary situation of gamma correction, be enough to use with first and second embodiment in similar LUT, as index LUT 21B and the LUT 22B of mantissa.Simultaneously, when utilizing gamma value 1 to carry out, be enough to utilize and the above similar method, use LUT that the tone that input signal is converted to floating-point system is represented against gamma correction.
As shown in figure 16, in contrary gamma-correction circuit 2B, because the input figure place is six (6) bits, index output figure place is three (3) bits, and mantissa's output figure place is six (6) bits, therefore carry out conversion,, obtain enough precision so that allow the maximal value of exponential part under situation, to increase too many degree at above-mentioned each embodiment to be unlikely.That is, when input signal equaled tonal signaling maximal value 63, index equaled 3, and mantissa equals 63,2^3 * 63=504.When input signal was 1, index equaled 0, and mantissa equals 8, and this is because 504/63=8=2^0 * 8.
The signal of contrary gamma-correction circuit 2B output is imported into brightness-contrast correction circuit 5 then.As shown in figure 17, when the index of the signal that is input to brightness-contrast correction circuit 5 is a and mantissa when being b, carry out following the processing.
(output)=2^a * (b * c1)+c0
In data shift device 23B, utilize the bit shift of a to carry out 2a above formula operation.Then, this computing output is converted to exponential sum mantissa by signal converter 24 once more, so that output.
At this, when the c1 value is set to 1 and c0 value when being set to 0, it is equivalent to this step and does not do whatever.When the c1 value is set to 1 or when bigger, can obtain the more image of high-contrast, and be set to 1 or more hour as c1, can obtain the more image of low contrast.In addition, c0 represents compensation, and when c0 gets greater than 0 value, can increase brightness of image.In addition, by changing c0 and c1 combination, can realize contrast-gamma correction of expecting.
When input tone value hour,, carry out tone with respect to exponential sum mantissa respectively and represent that the operational precision in brightness-contrast correction circuit 5 becomes higher when representing to compare with common tone.Therefore, even behind contrary gamma correction the one-level circuit for the purpose that is different from corresponding color rendering carries out image processing, also can under the situation of the circuit scale that does not increase computing circuit (adder circuit, multiplier circuit) and positive gamma-correction circuit (LUT), realize more high-precision computing.
In addition, the application of this digital gamma correcting circuit makes it possible to reduce the circuit scale of image processing equipment and display device.
The 4th embodiment
Below the fourth embodiment of the present invention will be described.
In above-mentioned each embodiment, utilize the digital processing in the gamma-correction circuit of image processing circuit, carry out gamma correction and handle.In this embodiment, with a kind of structure of in the DA converter, carrying out positive gamma correction processing of explanation.
Figure 18 has shown the structure of the liquid crystal display of this embodiment.This liquid crystal display comprises image processor 1, selector switch 6, benchmark tone voltage generator 7 and LCD 12.
Select one group many class indexs signal that selector switch 6 is sequentially imported from color space converting circuit 3 and the mantissa signal, export this selected group then.
The a plurality of reference voltages that are used for the DAC conversion of benchmark tone voltage generator 7 outputs.
The liquid crystal display of this embodiment utilizes converter (DAC) and the benchmark tone voltage generator 7 in Source drive and the control-signals generator 34, carries out the processing (positive gamma correction) that utilizes positive gamma-correction circuit 4 to carry out in the liquid crystal display of first embodiment.
Exponential signal in the output signal of selector switch 6 is output to benchmark tone voltage generator 7, and mantissa signal is output to the DAC in Source drive and the control-signals generator 34.
Structure shown in Figure 18 is by carrying out serial conversion to selector switch 6 selected RGB parallel signals, carry out positive gamma correction, but also might adopt this structure, come to carry out processing concurrently with comprising the reference color tonal signal generator 7 that is respectively applied for every kind of color of RGB and the structure of DAC separately.
Benchmark tone voltage generator 7 output and the corresponding expectation benchmark of input exponential signal tone voltage V1 to V9, and these voltages are sent to DAC in Source drive and the control-signals generator 34.In this DAC, the output voltage when the reference color tone pitch D1 to D9 of mantissa signal is transfused to as supplied with digital signal respectively is set to benchmark tone voltage V1 to V9.That is, when the input mantissa signal was D1, V1 was output as the output of DAC, and when the input mantissa signal was D2, V2 was output.If supposition D1＜D2＜...＜D9 then when the tone between D1 and the D2 is imported into DAC, is output by two benchmark tone voltage V1 that this tone is clipped in the middle and the voltage that linear interpolation obtained of V2.Figure 19 has shown the input-output characteristic of this DAC.Dotted line is represented the input-output characteristic (input-output characteristic that is suitable for the gamma characteristic of LCD 12) expected, and solid line is represented the input-output characteristic of DAC.As shown in the figure, the input-output characteristic of DAC is similar to the characteristic of utilizing multistage to obtain from the expectation input-output characteristic.When determining display device, can determine this input-output characteristic uniquely.
At this, with a kind of benchmark tone voltage method that is used for determining benchmark tone voltage generator 7 of explanation.Benchmark tone voltage generator 7 can produce and import the as many many group benchmark tone voltage V1 to V9 of number of the value that exponential signal can get.For example, when exponential signal is two (2) bits, can produce four (=22) group voltage V1 to V9.Below with reference to Figure 20 and 20 example is described.
For example, suppose that exponential signal Tn and mantissa signal Tm as 6 output valves of selector switch among Figure 18 are respectively two (2) bits and six (6) bits (index is that two (2) bits and mantissa are six (6) bits), and tone value is represented as 2^Tn * Tm (that is, A=2).In addition, when Tn=" 11 (decimal numeral 3) " is set, suppose one group of benchmark tone voltage that benchmark tone voltage generator 7 is produced be V1M, V2M ..., V9M.Suppose with the corresponding mantissa signal value of this magnitude of voltage be D1M, D2M ..., D9M, also suppose D1M=" 000000 ", D5M=" 011111 ", D8M=" 111110 ", D9M=" 111111 " below simultaneously.In this case, become 2^Tn * D5M=2^3 * 31=248 (with reference to Figure 20) with the corresponding tone value of V5M output voltage.
At this, when Tn=" 10 (decimal numeral 2) ", organize more benchmark tone voltage V1M2, V2M2 ..., V9M2 is set up as follows.For example because the pairing tone value of V1M2 is 2^Tn * D1M=2^2 * 0=0, so can obtain V1M2=V1M (because 2^3 * 0=0).In addition, because the pairing tone value of V8M2 is 2^Tn * D8M=2^2 * 62=248, so can obtain V8M2=V5M (because 2^3 * 31=248) (with reference to Figure 21).As mentioned above, based on the V1M to V9M that obtains when Tn=" 11 ", the V1 to V9 during for other value as Tn is set.In other words, at first obtain the reference color tonal signal when Tn gets maximal value,, obtain to get reference color tonal signal when being different from peaked value as Tn then according to this reference color tonal signal.
In above example, nine reference color tone pitches (D1 to D9) of mantissa signal have been considered.Yet the reference color tone pitch number of mantissa signal never is limited to this.
In Figure 19, the minimum value that requires D1 to get mantissa signal, and the maximal value that requires D9 to get mantissa signal are so that change all mantissa signals according to the characteristic of positive gamma correction.In addition, as linear interpolation method, it is considered to utilize resistance string to obtain the part resistive voltage.
From the above description, obviously be the output (benchmark tone voltage) that utilizes reference color tonal signal generator 7, determine the voltage after the conversion among the DAC.
At this, will illustrate from data image signal being input to image processor 1, up to the processing of display image on LCD 12.In DAC, according to the conversion characteristic that applies voltage-light characteristic and obtain of the pixel 35 of LCD 12, to carrying out the DA conversion from the signal of image processor 1 output from the gamma characteristic of received image signal.
Reference color tonal signal generator 7 produces and is used for the benchmark tone voltage V1 to V9 that DAC carries out conversion, then it is outputed to DAC.At this, utilize conversion characteristic and exponential signal that benchmark tone voltage V1 to V9 is set.Then, the signal that is converted into simulating signal in DAC is applied to pixel 35 by Source drive and control-signals generator 34 and TFT 37.This simulating signal is converted into luminance signal, is used to produce image then.
More particularly, as shown in figure 18, the exponential signal of selector switch 6 outputs is sent to reference color tonal signal generator 7.Reference color tonal signal generator 7 receives this exponential signal, produces 2^ (exponential signal) is provided the reference voltage of times input signal tone, then this reference voltage is outputed to DAC.Because having produced, reference color tonal signal generator 7 provides 2^ (exponential signal) the reference color tonal signal of times input signal tone, though therefore mantissa signal is imported into DAC, but the output simulating signal becomes 2^ (exponential signal) * (mantissa signal), is output then.
Utilize said structure, can not carry out utilizing reference color tonal signal generator 7 and DAC to realize positive gamma correction under the synthetic situation of numeral the signal that is divided into exponential sum mantissa.And, proved when the precision of DAC is maintained six (6) bits, by utilization index modulation reference color tonal signal, can carry out conversion more accurately.
Utilize said structure, the same with the situation of above-mentioned each embodiment, can carry out high-precision computing by tonal signaling being divided into exponential sum mantissa.
In addition, by should this digital gamma correcting circuit, can reduce the circuit scale of image processing equipment and display device.
Above-mentioned each embodiment is the example of the preferred embodiment of the present invention, and the present invention never is limited to this.For example, liquid crystal display mainly has been described, but the present invention never is limited to this in each embodiment, the present invention also is suitable for using the display device of PDP and organic electroluminescent in addition.
In addition, image processor 1 is formed circuit in above-mentioned each embodiment, but also might computing machine be used as image processor 1 by software processes.In this case, contrary gamma corrector and positive gamma corrector are formed on and are used to carry out in the arithmetic facility (central processing unit (CPU) etc.) of contrary gamma correction program and positive gamma correction program, so that carry out the processing that is similar among each embodiment.Contrary gamma correction and positive gamma correction can utilize software processes to realize naturally, and any correction in these corrections can utilize software processes to realize.In addition, also might utilize the software processes that comprises the processing (color space transformation processing, brightness-contrast correction processing etc.) that between contrary gamma correction and positive gamma correction, to carry out, carry out these corrections.
In addition, also might make up structure within the bounds of possibility about each embodiment.As mentioned above, the present invention allows various changes.
Previous embodiment describes and is used to make those skilled in the art can implement and use the present invention.And the various changes of these embodiment will be conspicuous for those skilled in the art, and can will be applied to other embodiment in the rule and the specific examples of this definition under the situation without the present invention's instruction.Therefore, the invention is not restricted to embodiment described here, but should meet the restriction the widest defined scope of claim and equivalence requirement.
Further, the inventor is intended that, even require also to keep desired all equivalents of the present invention in the pendend lite modification right.
Claims (23)
1, a kind of gamma correction method comprises:
Input has the X bit tone input signal Tin value of non-linear tone-light characteristic, and output has the Y bit of the tone-light characteristic of the linear characteristic of being converted into, and (the tonal signaling Tout of X＜Y) realizes thus against gamma correction; And
Import described Tout, and when described Tout being expressed as A^Tn * Tm with the mantissa of the exponential sum M bit signal Tm indication of constant A, N bit signal Tn indication, output realizes format conversion thus by one group of Tn of the most approaching described Tout and (N+M) bit tone output signal that Tm forms;
Wherein A is 2 the power of several naturally,
M satisfies relational expression M≤Y-N,
When A equaled 2, the minimum value of N was 2, when A greater than 2 the time, the minimum value of N is 1, and
The maximal value of the N that determines according to A and M is to make with respect to all Tin values, and one group of Tn and Tm value all become the maximal value of one group of different value.
2, gamma correction method according to claim 1, wherein, Tn has a predetermined value that is encoded as by the element that is equal to or greater than the sequence that 0 N integer form that increases progressively from initial value according to equal difference not.
3, gamma correction method according to claim 1, wherein, A=2, M=Y-Z ', and N is set to be equal to or greater than the smallest positive integral value of 1+logZ '/log2, when the contrary gamma correction table that is used to import described Tin value and export described Tout value is determined in advance, be set under the condition of Tz than the Tout value after the low order being reduced the Z position, in the Tz value that obtains with respect to all Tout values, only extract the highest significant position value and be those Tz values of 1, and make the maximal value in the Z value of different value of Tz value of all extractions be set to Z '.
4, a kind of gamma correction method comprises:
Input comprises (N+M) bit signal of N bit signal Tn and M bit signal Tm, and as input signal, and output is expressed as Tout=A with constant A, N bit index Tn and M mantissa bit Tm
TnY bit tonal signaling Tout * Tm, that have linear tone-light characteristic realizes format conversion thus; And
Input Tout value, and output have by the X bit of the tone-light characteristic of non-linear conversion (the tonal signaling Tin of X＜Y) realizes the gamma correction processing thus,
Wherein A is 2 the power of several naturally,
M satisfies relational expression M≤Y-N,
When A equaled 2, the minimum value of described N was 2, when A greater than 2 the time, the minimum value of described N is 1, and
The maximal value of the N that determines according to A and M is to make with respect to all Tin values, and one group of Tn and Tm value all become the maximal value of one group of different value.
5, gamma correction method according to claim 1, wherein, described contrary gamma correction and described format conversion are carried out in a utilization only look-up table.
6, a kind of image processing method comprises:
After gamma correction according to claim 1, carry out color space transformation and handle.
7, a kind of image processing method comprises:
After gamma correction method according to claim 1, carry out brightness-contrast correction and handle.
8, a kind of control computer that is used for is carried out the gamma correction program that gamma correction is handled, and described gamma correction is handled and comprised:
Input has the X bit tone input signal Tin value of non-linear tone-light characteristic, and output has the Y bit of the tone-light characteristic of the linear characteristic of being converted into, and (the tonal signaling Tout of X＜Y) realizes thus against gamma correction; And
Input Tout, and when Tout being expressed as A^Tn * Tm with the mantissa of the exponential sum M bit signal Tm indication of constant A, N bit signal Tn indication, output realizes format conversion thus by one group of (N+M) bit tone output signal that forms near Tn and the Tm of Tout;
Wherein A is 2 the power of several naturally,
M satisfies relational expression M≤Y-N,
When A equaled 2, the minimum value of N was 2, when A greater than 2 the time, the minimum value of N is 1, and
The maximal value of the N that determines according to A and M is to make with respect to all Tin values, and one group of Tn and Tm value all become the maximal value of one group of different value.
9, gamma correction program according to claim 8, wherein Tn gets a predetermined value that is encoded as by the element that is equal to or greater than the sequence that 0 N integer form that increases progressively from initial value according to equal difference not.
10, gamma correction program according to claim 8, wherein, the described computing machine of described gamma correction programmed control is carried out following setting the: A=2, M=Y-Z ', and N is for being equal to or greater than the smallest positive integral value of 1+logZ '/log2, when the contrary gamma correction table that is used to import described Tin value and export described Tout value is determined in advance, be set under the condition of Tz than the Tout value after the low order being reduced the Z position, in the Tz value that obtains with respect to all Tout values, only extract the highest significant position value and be those Tz values of 1, and make the maximal value in the Z value of different value of Tz value of all extractions be set to Z '.
11, a kind of control computer that is used for is carried out the following gamma correction program of handling, and comprising:
Format conversion processing is used to import (N+M) bit signal that comprises N bit signal Tn and M bit signal Tm, and as input signal, and output is expressed as Tout=A with the N bit signal of constant A, described Tn indication and the M mantissa bit of described Tm indication
TnY bit tonal signaling Tout * Tm, that have linear tone-light characteristic; And
Gamma correction is handled, and is used to import described Tout value, and export have by the X bit of the tone-light characteristic of non-linear conversion (the tonal signaling Tin of X＜Y),
The value of wherein said constant A is 2 the power of several naturally,
Described M satisfies relational expression M≤Y-N,
When A equaled 2, the minimum value of described N was 2, when A greater than 2 the time, the minimum value of described N is 1, and
The maximal value of the described N that determines according to described constant A and described M is to make with respect to all Tin values, and one group of Tn and Tm value all become the maximal value of one group of different value.
12, a kind of gamma-correction circuit comprises:
Be used to import X bit tone input signal Tin value, and output has the Y bit of the tone-light characteristic of the linear characteristic of being converted into, and (the tonal signaling Tout of X＜Y) realizes the circuit of contrary gamma correction thus with non-linear tone-light characteristic; And
Be used to import described Tout, and ought described Tout be expressed as A with the mantissa that constant A, the exponential sum M bit signal Tm of N bit signal Tn indication indicate
TnDuring * Tm, output realizes the circuit of format conversion thus by one group of Tn of the most approaching described Tout and (N+M) bit tone output signal that Tm forms;
Wherein A is 2 the power of several naturally,
M satisfies relational expression M≤Y-N,
When A equaled 2, the minimum value of N was 2, when A greater than 2 the time, the minimum value of N is 1, and
The maximal value of the N that determines according to A and M is to make with respect to all Tin values, and one group of Tn and Tm value all become the maximal value of one group of different value.
13, gamma-correction circuit according to claim 12, it further comprises and is used for described Tn is encoded to device by the predetermined value of the element that is equal to or greater than the sequence that 0 N integer form that increases progressively from initial term according to equal difference not.
14, gamma-correction circuit according to claim 12, wherein, A=2, M=Y-Z ', and N is set to be equal to or greater than the smallest positive integral value of 1+logZ '/log2, when the contrary gamma correction table that is used to import described Tin value and export described Tout value is determined in advance, be set under the condition of Tz than the Tout value after the low order being reduced the Z position, in the Tz value that obtains with respect to all Tout values, only extract the highest significant position value and be those Tz values of 1, and make the maximal value in the Z value of different value of Tz value of all extractions be set to Z '.
15, gamma-correction circuit according to claim 12, wherein, the output of described tone is imported in adder circuit, multiplier circuit and the lut circuits any one circuit at least.
16, gamma-correction circuit according to claim 15, wherein, described tone output is imported into and is used for carrying out the circuit that color space transformation is handled according to the colorimetric properties of display device.
17, gamma-correction circuit according to claim 15, wherein, described tone output is imported into the circuit that is used for carrying out according to the brightness-contrast-response characteristic of display device brightness-contrast conversion process.
18, a kind of gamma-correction circuit comprises:
Format conversion circuit is used to import (N+M) bit signal that comprises N bit signal Tn and M bit signal Tm, and as input signal, and output is expressed as Tout=A with the M mantissa bit that the described Tm of N bit exponential sum of constant A, described Tn indication indicates
TnY bit tonal signaling Tout * Tm, that have linear tone-light characteristic; And
Gamma-correction circuit is used to import the Tout value, and output have by the X bit of the tone-light characteristic of non-linear conversion (the tonal signaling Tin of X＜Y),
The value of wherein said constant A is 2 the power of several naturally,
Described M satisfies relational expression M≤Y-N,
When A equaled 2, the minimum value of described N was 2, when A greater than 2 the time, the minimum value of described N is 1, and
The maximal value of the described N that determines according to described constant A and described M is to make with respect to all Tin values, and one group of Tn and Tm value all become the maximal value of one group of different value.
19, gamma-correction circuit according to claim 12, wherein, a utilization only look-up table realizes described contrary gamma-correction circuit and described format conversion circuit.
20, a kind of image processing equipment that comprises gamma-correction circuit according to claim 12.
21, a kind ofly be used to export aanalogvoltage, comprise to determine the image processing equipment of display device tone:
Digital to analog converter is used for the M bit input signal execution digital-to-analog conversion to the input signal that comprises N bit signal and M bit signal; And
Pedestal generator is used for exporting the reference voltage that is used for described digital to analog converter digital-to-analog conversion,
Wherein, provide the index Tn of described N bit signal indication and the Tm of mantissa of described M bit signal indication, so that utilize constant A that the Y bit tonal signaling Tout with linear tone-light characteristic is expressed as Tout=A
Tn* Tm,
The value of described A is 2 the power of several naturally, and
Described pedestal generator produces described reference voltage according to described N bit signal.
22, a kind of display device comprises:
Display device; And
Image processing equipment according to claim 20.
23, display device according to claim 22, wherein, described display device is a LCD.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2004286502 | 2004-09-30 | ||
| JP2004286502 | 2004-09-30 | ||
| JP2005161554 | 2005-06-01 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1755791A true CN1755791A (en) | 2006-04-05 |
| CN100472602C CN100472602C (en) | 2009-03-25 |
Family
ID=36688961
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2005101041936A Expired - Fee Related CN100472602C (en) | 2004-09-30 | 2005-09-29 | Gamma correction method and circuit, image processing apparatus, and display apparatus |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100472602C (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102761749A (en) * | 2012-07-25 | 2012-10-31 | 四川虹微技术有限公司 | Device for generating gamma correction data table of plasma TV video image |
| CN102761748A (en) * | 2012-07-25 | 2012-10-31 | 四川虹微技术有限公司 | Method for generating gamma correction data table of plasma TV video image |
| CN101821774B (en) * | 2007-09-21 | 2013-03-06 | 全球Oled科技有限责任公司 | Preferential tone scale for electronic displays |
| WO2016107268A1 (en) * | 2014-12-31 | 2016-07-07 | 小米科技有限责任公司 | Method and device for color adjustment |
| CN117372296A (en) * | 2023-10-19 | 2024-01-09 | 上海成运医疗器械股份有限公司 | Economical gamma correction method and device |
-
2005
- 2005-09-29 CN CNB2005101041936A patent/CN100472602C/en not_active Expired - Fee Related
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101821774B (en) * | 2007-09-21 | 2013-03-06 | 全球Oled科技有限责任公司 | Preferential tone scale for electronic displays |
| CN102761749A (en) * | 2012-07-25 | 2012-10-31 | 四川虹微技术有限公司 | Device for generating gamma correction data table of plasma TV video image |
| CN102761748A (en) * | 2012-07-25 | 2012-10-31 | 四川虹微技术有限公司 | Method for generating gamma correction data table of plasma TV video image |
| CN102761749B (en) * | 2012-07-25 | 2014-04-23 | 四川虹微技术有限公司 | Device for generating gamma correction data table of plasma TV video image |
| CN102761748B (en) * | 2012-07-25 | 2014-04-23 | 四川虹微技术有限公司 | Method for generating gamma correction data table of plasma TV video image |
| WO2016107268A1 (en) * | 2014-12-31 | 2016-07-07 | 小米科技有限责任公司 | Method and device for color adjustment |
| US10269129B2 (en) | 2014-12-31 | 2019-04-23 | Xiaomi Inc. | Color adjustment method and device |
| CN117372296A (en) * | 2023-10-19 | 2024-01-09 | 上海成运医疗器械股份有限公司 | Economical gamma correction method and device |
| CN117372296B (en) * | 2023-10-19 | 2024-04-05 | 上海成运医疗器械股份有限公司 | Economical gamma correction method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN100472602C (en) | 2009-03-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100347735C (en) | Method for driving displaying device, displaying device and related computer programm | |
| US7663678B2 (en) | Gamma correction, image processing method and program, gamma correction circuit, image processing apparatus, and display apparatus | |
| CN1145064C (en) | Circuit and method for driving electrooptic device, electrooptic device and electronic equipment made by using the same | |
| CN1208973C (en) | Color signal correcting circuit, device, method and program and color display | |
| CN100346635C (en) | Gradation attributes control according to characteristics of image | |
| CN1077294C (en) | Liquid crystal display apparatus and method of driving the same, and power supply circuit for liquid crystal display apparatus | |
| CN1882103A (en) | Systems and methods for implementing improved gamut mapping algorithms | |
| CN1661666A (en) | Method and apparatus for converting color spaces and multi-color display apparatus using the color space conversion apparatus | |
| CN1664905A (en) | Apparatus and method of converting image signal for four-color display device, and display device including the same | |
| CN101069227A (en) | Multi-line addressing methods and apparatus | |
| CN1838729A (en) | Gamma correction device, image conversion apparatus using the same, and display device | |
| CN1494710A (en) | Display | |
| CN1664912A (en) | Color correction circuit and image display device equipped with the same | |
| CN1625764A (en) | Liquid crystal display and method of modifying gray signals for the same | |
| TW201112166A (en) | Image signal processing system and method of processing images | |
| CN1755791A (en) | Gamma correction method and circuit and image processing equipment and display device | |
| CN1941869A (en) | Display image correcting device, image display device, and display image correcting method | |
| CN101047772A (en) | Color conversion device and image display apparatus having the same | |
| CN1748241A (en) | Color space correction circuit in display device | |
| CN1928982A (en) | Display apparatus containing controller driver with correcting circuit and method of driving display panel | |
| TW201312531A (en) | Multi-primary color LCD and color signal conversion device and method thereof | |
| WO2007125630A1 (en) | Image display device, method for driving image display device, driving program, and computer readable recording medium | |
| CN1419781A (en) | Selected data compression for digital pictorial information | |
| CN1277785A (en) | Gamma correction circuit and gamma correction method | |
| CN113808550B (en) | Device applicable to brightness enhancement in display module |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20090325 Termination date: 20150929 |
|
| EXPY | Termination of patent right or utility model |