Method and apparatus for writing information on an optical recording medium

w’bptt nww inn bj> tjw’ib ipjim ηυ»υ

Method and apparatus for writing information on an optical recordingmedium

Koninklijke Philips Electronics N.V. C.112389 WO 98/25267 1 PCT/3B97/01489

Method and apparatus for writing optical recording media.

The invention relates to an optical , recording apparatus for writinginformation on an optical recording medium by a radiation beam, and in particular to amethod for setting an optimum write power level of the radiation beam, comprising a firststep of writing a series of test patterns on the recording medium, each pattern with a 5 different value of the write power level, a second step of reading the patterns to formcorresponding read signals, and a third step of deriving a value of a read parameter fromeach read signal, and a fourth step of curve-fitting the values of the read parameter to afunction defining a relation between the write power level and the read parameter. Theinvention further relates to an apparatus for writing information on an optical recording 10 medium, comprising a radiation source for emitting a radiation beam having a controllablewrite power level, a control unit for writing a series of test patterns, each pattern with adifferent value of the write power level, a read unit for reading the patterns and formingcorresponding read signals, a first processor for deriving a value of a read parameter fromeach read signal, a second processor for curve-fitting the values of the read parameter to a 15 function defining a relation between the write parameter and the read parameter. A processor may be of an analog or digital type and includes programmable processors and fixed-program processors. The processor may comprisememory means beside an arithmetic unit. A recording method according to the preamble is known from the United 20 States patent no. 5 185 733. In this method a series of test patterns is written on the medium,each subsequent test pattern being written with an increased write power level. After readingthe written patterns, a curve fitting algorithm fits a second order polynomial to theamplitudes of the read signal of each pattern as a function of the write power level. Thewrite power corresponding to the maximum of the polynomial is selected as the optimum 25 write power for subsequently recording information of the medium. The problem of the noise character of the read signal is mitigated the by the curve-fitting. A disadvantage of this known method is that the optimum power level as determined by the method does not sufficiently take into account variations in properties of recording apparatuses and variations in properties of recording media. Hence, use of this optimum write power may cause WO 98/25267 2 PCT/EB97/01489 unreliable storage of information on the recording medium.

It is an object of the invention to provide a reliable method for setting theoptimum write power in dependence on read signals from test patterns written on a mediumand being less affected by noise. 5 This object is achieved when the method according to the invention is characterized in that it comprises a fifth step of determining a derivative of the function andnormalising the derivative by multiplying it by a factor equal to the value of the write powerover the value of the read parameter, and a sixth step of selecting the optimum write powerlevel in dependence on the normalised derivative of the function. The value of the normalised 10 derivative as a function of the write power turns out to be little affected by variations ofparameters of the recording apparatus and the recording media. The variations of theapparatus may relate to the distribution of the radiation energy over the cross-section of theradiation beam used for recording, the power calibration of the radiation source. Thevariations of the media may relate to changes in write sensitivity or reflection over the 15 surface of a medium or between media. If the optimum write power level is selected independence on the normalised derivative, the selected level is suitable for reliable recordingon a large variety of recording media by different recording apparatuses. The level may beselected by taking the power level corresponding to a preset value of the normalisedderivative. 20 In alternative embodiment of the method according to the invention, the method is characterised in that it comprises a fifth step of selecting the optimum write powerlevel in dependence on a predetermined value of a derivative of the function normalised bymultiplication by a factor equal to the value of the write power over the value of the readparameter. 25 To achieve a stable curve fitting procedure, the function to be fitted is preferably defined in terms of orthogonal polynomials. Then the value of a coefficient for apolynomial determined by the procedure does not depend on the values of the coefficientsdetermined for the other- polynomials.

The polynomials are preferably Legendre polynomials, which require 30 relatively little computing power in the curve fitting procedure because of their mathematicalsimplicity.

The read parameter is preferably an amplitude of the read signal obtained from the medium, because this amplitude can easily be derived from a radiation detection system intercepting radiation coming from the medium. Moreover, the amplitude is a WO 98/25267 3 PCT/IB97/01489 satisfactory parameter for determining the quality of the recorded signals.

The selection process for the optimum write power may be refined byreading a parameter value from the medium and using this value in selecting the value of thewrite parameter. The value of the parameter may be characteristic for the specific medium 5 on which it is recorded. The parameter allows an adaptation of the selection process to thecharacteristics of the medium being scanned. The parameter may be an initial value of orrange for the optimum write power level to reduce the range of power levels with which towrite the test patterns. The parameter may also be a preset value of the normalisedderivative, making it possible to select an optimum write power in dependence on the 10 properties of the medium. In contrast, a preset value fixed by the apparatus does not allowsuch a choice.

Before making calculations on signal values, the analog signals arepreferably converted to digital signals by an analog-to-digital convertor. The derivative maybedetermined by taking difference values of the fitted function. The values of the function 15 must then he represented by a relatively large number of significant digits, and thecalculations must also be performed over this large number of digits in order not to beaffected by the additional noise introduced by this method of determining the derivative. Thederivative of the function is preferably determined analytically, because such a determinationintroduces hardly any additional noise. The number of significant digits can then be reduced, 20 which reduces the required computing power of the second processor and allows a reducedquality of the analog-to-digital convertors.

In general, the write power levels for the test patterns are chosen in arange around the expected value of the optimum write power level. However, the optimumwrite power level is often close to the maximum power the radiation source-can generate. 25 The range will then be limited by the maximum laser power. This problem can be avoidedby selecting a power level from the derivative of the function and subsequently determiningthe optimum power level by multiplying the power level by a constant larger than one. Thetest patterns can then be written in a range further removed from the maximum laser power.This has the additional advantage that the measurements are made in a power range where 30 the function and its derivative are removed from their saturation values, which allows a higher accuracy of the determination of the function and its derivative. A further aspect of the invention relates to the apparatus for writing information on the optical recording medium. This apparatus is according to the invention characterized in that the second processor is operatively connected for deriving a derivative WO 98/25267 4 PCT/IB97/01489 of the function, normalising the derivative by multiplying it by a factor equal to the value ofthe write power over the value of the read parameter, and for selecting a value of the writeparameter in dependence on the normalised derivative.

The objects, features and advantages of the invention will be apparentfrom the following more particular description of preferred embodiments of the invention, asillustrated in the accompanying drawings.

Figure 1 is a diagram of an optical recording apparatus according to the invention,

Figure 2 illustrates read signals from two test patterns,

Figure 3 is a graph showing the measured modulation as a function ofwrite power and its derivative,

Figure 4 is a plan view of a recording medium, and

Figure 5 is a plan view of a pattern of marks in the medium.

Figure 1 shows an apparatus and an optical recording medium 1 accordingto the invention. Medium 1 has a transparent substrate 2 and a recording layer 3 arranged onit. The recording layer comprises a material suitable for writing information by means of aradiation beam. The recording may be of e.g. the magneto optical type, the phase-changetype, the dye type or any other suitable material. Information may be recorded in the form ofoptically detectable regions, also called marks, on recording layer 3. The apparatuscomprises a radiation source 4, e.g. a semiconductor laser, for emitting a radiation beam 5.The radiation beam is converged on recording layer 3 via a beam splitter 6, an objective lens7 and substrate 2. The medium may also be air-incident, where the radiation beam is directlyincident on recording layer 3 without passing through a substrate. Radiation reflected frommedium 1 is converged by objective lens 7 and, after passing through beam splitter 6, tailson a detection system 8, which converts the incident radiation in electric detector signals.

The detector signals are input to a circuit 9. The circuit derives several signals from the detector signals, such as a read signal SR representing the information being read from medium 1. Radiation source, beam splitter 6, objective lens 7, detection system 8 and circuit 9 form together a read unit 10’. The read signal from circuit 9 is processed in a first processor 10 in order to derive signals representing a read parameter from the read signal WO 98/25267 5 PCT/IB97/01489 and necessary for controlling the laser power level. The derived signals are fed in a secondprocessor 11, which processes a series of values of the read parameter and based thereonderives a value for an optimum write power control signal. The write power control signal isconnected to a control unit 12. An information signal 13, representing the information to bewritten on medium 1, is also fed into control unit 12. The output of control unit 12 isconnected to radiation source 4. A mark on recording layer 3 may be written by a singleradiation pulse, the power of which is determined by the optimum write power level asdetermined by processor 11. A mark may also be written by a series of radiation pulses ofequal or different lengths and a power determined by the optimum write power level.

The actual radiation power emitted by radiation source 4 may be measured by anot-shown detector arranged in an otherwise not-used side lobe of the radiation beam or inradiation reflected off an element in the optical path of the radiation beam. The signal of thedetector may be connected directly to processor 11. Alternatively, the signal may beconnected to control unit 12, where it may be combined with the peak amplitude of the readsignal, which is a measure for the radiation power received at recording layer 3, andsubsequently fed into processor 11.

Before writing information on medium 1 the apparatus sets its write powerto the optimum value by performing the following procedure. First the apparatus writes a 'series of test patterns on medium 1. The test patterns should be selected so as to give adesired read signal. If the read parameter to be derived from the read signal is the maximummodulation of the read signal, the test pattern should comprise marks sufficiently long toachieve a maximum modulation of the read signal. When the information is coded accordingto the so-called EFM modulation, the test patterns preferably comprises the long I„ marks.The test patterns are recorded each with a different write power. Subsequent patterns may bewritten with a step-wise increased write power under the control of processor 11. Thepatterns may be written anywhere on the medium. They may also be written in speciallyprovided test areas on the medium.

Figure 2 shows the read signals 18 and 19 obtained from two patternswritten at two different write power levels. The patterns comprise a short mark, a long markand a short mark, as shown by the signal parts 15, 16 and 17, respectively in both readsignal 18 and read signal 19. An actual pattern may comprise a few hundred marks ofdifferent or equal lengths.

Processor 10 derives from the read signal SR a read parameter used for finding the optimum write power. A possible read parameter is the ratio of the lowest WO 98/25267 6 PCT/E897/01489 amplitude of the signal parts in a read signal, indicated by ‘a’ in Figure 2, and the maximumvalue of the read signal ‘b’. A preferred read parameter is the normalised modulation beingthe ratio of the maximum peak-to-peak value of a read signal ‘c’ and the maximum value ‘b’of the read signal.

After reading the test patterns on medium 1, processor 11 has available aseries of pairs of values for the modulation of a pattern and the write power belonging to thatpattern. The write powers may be taken from the value of the write power control signalduring recording the test patterns, or from a measurement of the radiation power. Figure 3shows schematically the result of reading; the crosses are measured values of the modulationm as a function of the write power P. Processor 11 fits a curve through the measuredmodulation values in order to obtain an analytic expression for the variation of themodulation as a function of the write power. The curve is indicated in Figure 3 by a dashedcurve. The fitting may be done by the well-known least-squares fitting algorithm. ~~ As a next step, processor 11 calculates analytically a normalised derivative ‘g’ with respect to the write power of the modulation. The normalised derivativeg(P) is equal to the function (dm/dP)P/m. The function g derived from the fitted modulationm in Figure 3 is shown by the drawn curve.

The processor derives an intermediate write power P, from the normalisedderivative by taking the value of the write power P belonging to a preset value go, asindicated by the dashed lines in Figure 3. The value of go may be a value set by themanufacturer of the recording apparatus and stored in a memory of the apparatus, or it maybe a value stored on the medium to be written and read prior to or during the procedure toset the optimum write power. As a next step, the value of the intermediate power Pj ismultiplied by a constant h larger than one, resulting in the optimum write power level Po.

The values of the preset value go and the multiplication constant h aredetermined by the manufacturer of the medium or by the user during initialisation of themedium. The value of g0 is set within a range from 0.2 to 5.0. For values higher than 5.0the normalised derivative looses it predicting value, because the proximity of an asymptotecauses the values of P related to go to lie closely together on the write power axis. For valuesof go lower than 0.2 the normalised derivative has a small slope, through which small errorsin the value of the derivative result in a large spread of the values of Pj associated with go.Experiments on rewritable recording media having a CD format gave values of go within arange from 0.2 to 2.0 and on higher density media in a range from 2.0 to 4.0. Themultiplication constant h is set preferably within a range from 1.00 to 1.35. The optimum WO 98/25267 7 PCT/IB97/01489 write power, Po, equal to (h Pj), is in general set to a value near the write power where themodulation m starts saturating. In a preferred method of setting Go and h, the optimum writepower of a specific medium is determined by finding the write power giving the lowest jitterof the read signal for information written on the medium. The information is preferablyrandom information. Next, the normalised derivative dm/dP (P/m) is determined from awritten series of test patterns as described above. A value for P, is selected such that theassociated value of g0 lies within the above range, and where the normalised derivative isneither too fiat nor too steep. The associated values of h, equal to Po/Pj, and g0 can now beused for all media of this type and for all recording apparatuses.

The value of the normalised derivative turns out to be hardly affected byvariations of parameters of the recording apparatus and the recording media. If the optimumwrite power level is selected in dependence on the normalised derivative, the selected level issuitable for reliable recording on a large variety of recording media by different recordingapparatuses. The level may be selected by taking the power level corresponding to a presetvalue of the normalised derivative. The advantages of the use of the normalised derivativemay also be achieved when no curve-fitting is performed. In that case the derivative may bedetermined from the read parameter versus write power level data by e.g. calculatingdifferences between the measured values. However, omitting the step of curve-fitting willincrease the noise in the values of the derivative, thereby making it impossible for somemedia to use the derivative for setting the optimum write power.

The values of go and h may be stored in the apparatus, thereby providingmedium-independent values of the parameters. Preferably, the value of go is stored in themedium, allowing media-dependent values. Figure 4 shows optically readable recordingmedium 1, provided with a track "30. The track may be spiral and in the form of for examplean embossed groove or ridge. The area of the medium is divided in an information recordingarea 31 for writing user information and a control area 32 for storing information relevantfor writing, reading and erasing information on the medium and in general not intended forrecording user information. For some types of medium the information in the control isembossed. Control area 32 is marked by dashed track 32 in the Figure. Informationrecording area 31 is of a type which is subject to change in an optically detectable propertywhen exposed to radiation above a specific write power level. The value of go may be storedas a pattern of control information in control area 32 of the medium. When the control areais embossed, the manufacturer of the medium must record the value. Alternatively, the usermay record the value in the medium during for instance initialisation of the medium, WO 98/25267 8 PCT/3B97/01489 allowing the recording of a disc-specific value. The value of h may also be recorded like thevalue of go. Figure 5 shows a strongly enlarged portion of track 33 comprising a pattern ofmarks 34.

The curve to be fitted through the pairs of values (m,P) may be one ormore polynomials, which are preferably orthogonal. The curve may then be written as: (1) »I(P) = £a^(P) where f (P) is the derivative of function f with respect to the parameter P. The normalisedderivative is given in the analytic form

(2T g(F)

The value of P; can be found from the equation (3) g(Pt) - g0.

Depending on the choice of the curve to be fitted, the value of Pj can be found in the form ofan analytic expression or as the result of a numerical successive root approximation methodsuch as the regula falsi or Newton’s method. The use of an analytic expression, whenpossible, has the advantage that it always provides the correct root, whereas a successiveapproximation may divert to an undesired root. When an analytic form of equation (3) can befound, the normalised derivative g need not be determined anymore, but the preset value gomay be inserted directly into equation (3) for determining the associated value of Ph A suitable set of orthogonal polynomials f: are the Legendre polynomials.The three lowest order Legendre polynomials are given by:

V WO 98/25267 9 PCT/E397/01489 (4)

fo(P) = 1fi(P) = P

Since these polynomials are defined on the interval -1 <P< +1, the write power values to befitted should be scaled according to 2P-(P +p .) \ nm mm/ (5)

<img img-format="tif" img-content="drawing" file="IL125635AD00021.tif" id="idf0001" />

The scaled write power levels P, must now be used in the formulae of equation (4). Thevalue of P; found from equation (3) must be scaled back to the range P^, P^. 10 When using a digital processor, the input values of m and P must be converted from an analog value to a digital value by an analog-to-digital convertor. Thenumber of bits of the digital output values can be made to correspond to the noise in themeasured values. If, for example, the noise in the values of a parameter is 1 % of themaximum value of the parameter, the convertor should be at least 8 bits deep, thereby 15 introducing an additional 1/28 = 1/256 quantization noise.

If for reason of costs processor 11, in which the above calculations are made, is a small processor, the calculations are preferable performed in an integer format.The values of m and P should therefore be converted from real to integer values. Themultiplication constant for this conversion should be large enough not to introduce additional 20 noise and small enough not to require too require much computing power. A good guide is tochoose die constant such that the noise present in the value of m or P, as determined in theinteger representation of the value, is slightly larger than the value corresponding to the leastsignificant bit in the integer representation. The noise in the value includes the above-mentioned quantisation noise. If, for example, the noise in the values of m is 0.5% of the 25 maximum value of m, then a multiplication factor of about 1000 divided by the maximum value of the parameter is reasonable.

Instead of using a series of polynomials for fitting the values of m and P, WO 98/25267 10 FCT/IB97/01489 a single function may also be used. The function must have an asymptote for large values ofP, a zero value for a non-fixed value of P&gt;0 and a non-fixed value of the derivative of thefunction for the value of P where the function is zero. A suitable function is 5 (6)

The normalised derivative can be given in an analytic form. Equation (4) reduces to aquadratic equation, allowing to find the root without successive approximation.

Other suitable functions are the arc tangent and the hyperbolic tangent: 10

(7) f(P) = fl0 arctan (axP-aJ (8) f(P) - a0 tanh 15 The values of the arc tangent and the hyperbolic tangent may be stored in a lookup table tospeed up the calculations.