CROSS REFERENCE TO RELATED APPLICATIONS

This application is a continuation of U.S. patent application Ser. No. 16/867,359, filed May 5, 2020, the entirety of which is hereby incorporated herein by reference for all purposes.

BACKGROUND

Over the past decade, much of the world's data has moved into the cloud. To meet the increasing demand, cloud providers rely on a variety of data-storage technologies. These include non-volatile memory (NVM), flash, hard disk drives (HDDs), magnetic tape, and optical discs. These storage technologies differ from each other in terms of cost, latency, throughput, storage density, failure rate, and media lifetime.

SUMMARY

This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter. Furthermore, the claimed subject matter is not limited to implementations that solve any or all disadvantages noted in any part of this disclosure.

A promising technology for storing data is to encode the data as localized birefringent voxels in a dielectric storage medium. Such data may be stored at high densities, and the storage medium may have a long lifetime compared to magnetic and other storage media. However, reading the localized birefringence to decode the data involves performing a plurality of optical measurements at different polarization states. Each additional measurement increases the time needed for reading the medium. Thus, aspects of the technology disclosed herein help to reduce the time utilized for reading data stored as localized birefringence.

One aspect provides a system for reading birefringent data. The system comprises one or more light sources, a first polarization state generator positioned to generate first polarized light from light of a first wavelength band output by the one or more light sources, a second polarization state generator positioned to generate second polarized light from light of a second wavelength band output by the one or more light sources, an image sensor configured to acquire an image of a sample region through which the first polarized light and the second polarized light passes, a polarization state analyzer disposed optically between the sample region and the image sensor, a first bandpass filter disposed optically between the polarization state analyzer and the image sensor, the first bandpass filter configured to pass light of the first wavelength band, and a second bandpass filter disposed optically between the polarization state analyzer and the image sensor, the second bandpass filter configured to pass light of the second wavelength band. In some examples, three or more polarization state generators can be used to generate three or more different polarization states using light of three or more wavelength bands, and three or more corresponding bandpass filters can be used between the polarization state analyzer and the image sensor. With this aspect, plural wavelength-multiplexed measurements of a voxel can be made in a temporally overlapping manner, thereby reducing an amount of time utilized for reading birefringent data compared to the use of sequential measurements without wavelength multiplexing.

Another aspect provides, on a computing device, a method comprising acquiring measurement data for a birefringent voxel by directing light comprising one or more predetermined polarization states through the birefringent voxel and receiving the light at an image sensor, based upon the measurement data, determining two points on a surface of a Poincaré sphere corresponding to two possible birefringent states of the birefringent voxel, each state comprising a set of birefringence values including an azimuth angle and a retardance, applying a constraint to determine an azimuth angle and a retardance; and outputting determined birefringence values comprising the azimuth angle and the retardance determined. With this aspect, by applying the constraint to determine the azimuth angle and retardance, a birefringence value for the voxel may be determined with fewer measurements than where the constraint is not applied.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 schematically illustrates the reading of birefringent voxels in a storage medium.

FIG. 2 shows a schematic depiction of a storage medium comprising data encoded as birefringent voxels.

FIG. 3 shows a Poincaré sphere that represents polarization states as locations on the sphere.

FIG. 4 shows a flow diagram depicting an example method for determining birefringence values based upon a maximum value determined for a likelihood function.

FIG. 5 shows examples of polarization states for measuring birefringence values.

FIG. 6 shows example solutions for birefringence values based upon two measurements and based upon one measurement.

FIG. 7 shows a flow diagram depicting an example method for acquiring birefringence measurements using wavelength multiplexing.

FIG. 8 shows a block diagram of an example system for reading birefringent data utilizing wavelength multiplexing.

FIG. 9 shows an example system for wavelength multiplexing light of different polarization states using light from a same image source.

FIGS. 10A-10B show a flow diagram depicting an example method for measuring a birefringent voxel via wavelength multiplexing.

FIG. 11 shows a flow diagram depicting an example method for performing background correction for birefringence values.

FIG. 12 shows a block diagram of an example computing system.

DETAILED DESCRIPTION

As mentioned above, one promising technology for cloud data storage involves the use of high-power, short-pulse laser irradiance to optically write data into a solid dielectric substrate, such as glass. The irradiance induces at its focus a localized birefringence that can be later read back using polarization imaging. The term ‘voxel’ is used herein to refer to any discrete volume of a substrate where an individual data value (i.e., symbol) may be stored. The data stored in a voxel may take various forms. In principle, any of the Muller-matrix coefficients of a substrate lattice can be manipulated to encode data. In examples using silica glass substrates, the lattice perturbation from focused, polarized irradiance takes the form of a non-native birefringence localized at the focus. Accordingly, each voxel of the substrate can be modeled as a very small waveplate having a retardance magnitude and an azimuth angle. These model parameters may be manipulated independently to write a desired symbol to a given voxel. Here, the polarization angle of the beam determines the azimuth angle of a voxel, while various other factors (pulse amplitude, duration, energy, number, and/or spacing between pulses) determine the retardance of the voxel.

By dividing the continuous space of achievable azimuth angles and/or retardance magnitudes into discrete intervals, multi-bit data values can be encoded into each voxel by writing the birefringence of that voxel to be within one of the discrete intervals. Further, in some examples, plural parallel layers of voxel structures may be written to the same substrate by focusing the laser irradiance to specified depths below the irradiated surface of the substrate. These features, individually or in combination, may allow a high volume of data to be written into a single medium. In some examples, the storage medium comprises a solid, plate-like configuration. In other examples, the storage medium comprises a thin layer formed on another substrate. In further examples, the storage medium may have any other suitable configuration, such as a prism or cylinder.

A birefringent voxel behaves anisotropically when light passes through it, as different light polarization states travel at different speeds through the sample. As light travels through or reflects from a birefringent voxel, its polarization state is altered in a specific way that depends on the azimuth angle and retardance of the sample. As such, information regarding the azimuth angle and retardance of a voxel can be obtained by measuring the polarization state of polarized light that has interacted with a birefringent voxel.

Some methods of measuring a polarization state (and therefore the angle and retardance of a voxel) involve performing a series of measurements using different input or output (or both) polarizations. FIG. 1 schematically illustrates the reading of birefringent data stored in a storage medium 100. Light from a light source 102 passes through a polarization state generator (PSG) 104, which outputs polarized light with a polarization angle determined by the PSG. The light source may include an LED, laser or other light source.

After passing through the storage medium 100, the light from the light source 102 and PSG 104 passes through one or more voxels of a storage medium 105, through a polarization state analyzer (PSA) 106 and then to detector 108. The settings of PSA 106 and PSG 104 define a polarization state k for an intensity m_kof a measurement, as described in more detail below. The detector 108 may include, for example, a CMOS image sensor (e.g. high-resolution/high frame-rate sensor) or other suitable photodetector array that can image an entirety of a focal plane positioned within the storage medium 105, and thereby image a plurality of arrayed voxels in a same image. In other examples, a point detector or small detector array, for example, a photodiode, phototransistor, or SPAD (single photon avalanche diode) may be used to build up an image point-by-point). Although FIG. 1 shows transmission of light rays through the storage medium and on to the image sensor, the light rays may, in other examples, reach the image sensor by reflection from the storage medium.

In examples in which data is to be read from a plurality of layers of storage medium 105, variable-focus optics 110 may be used to adjust a focal plane of the detector 108, such that voxels at the focal plane are read while others are out-of-focus. In other examples, varying the focus can be achieved by moving the sample. FIG. 2 schematically shows an example storage medium 200 comprising multiple layers of voxels, two of which are illustrated at 202 and 204.

To describe the measurement of a voxel that is made, a commonly used construction is the Poincaré sphere, illustrated at 300 in FIG. 3. The Poincaré sphere 300 is used to represent the polarization states of light by mapping the last three components of the 4-D Stokes vector onto a 3-D cartesian coordinate system. For fully polarized light, a particular measurement state is described by a particular position on the surface of the sphere representing a particular polarization state of light. Partially polarized light is represented by points inside from the surface of the Poincaré sphere. The north pole represents RCP—right circularly polarized light. The south pole represents LCP—left circularly polarized light. The states on the equator are linear polarizations with angle defined by the azimuthal angle on the sphere. A general point on the sphere is elliptically polarized with ellipticity determined by the angle between the state and the pole, and azimuth defined by the angle around the sphere. An example polarization state is shown at 302 on the sphere. The black dashed curved line is a circle of constant latitude, and thus of constant retardance. The term “swing” refers to the angle 304 of the measurement state relative to the pole of the sphere. The horizontal angle is the azimuth 306. As the azimuth angle and retardance of the voxel leads to the output state of polarization of probe light, the azimuth angle and retardance of the voxel also can be represented as a point on the surface of the sphere.

Current methods for determining birefringence values for a voxel of a storage medium involve three or more (typically four) measurements of the voxel taken at different probe light polarization states. Multiple measurements are used because the determination involves four degrees of freedom—the voxel retardance, the voxel azimuth angle, a measurement scale and a measurement offset. Four-measurement methods determine all four of these values for each read process. Shribak and Oldenbourg (M. Shribak and R. Oldenbourg, “Techniques for Fast and Sensitive Measurements of Two-Dimensional Birefringence Distributions” in Applied Optics, Volume 42, Issue 16, 2003, published by Optical Society of America, (www.osapublishing.org/ao/abstract.cfm?uri=ao-42-16-3009; see also doi.org/10.1364/AO.42.003009) describe a three-measurement method that assumes that the offset is zero, and solves for the other three parameters via the three measurements. However, the use of three or four measurements can impact a speed at which data can be retrieved from storage, due to the number of measurements made.

Accordingly, example measurement processes are disclosed herein that may be used to determine azimuth angle and retardance values for a birefringent voxel using fewer than three measurements. Briefly, the disclosed methods utilize constraints on retardance values based upon prior knowledge of the retardance to allow retardance and azimuth angle values to be determined with two measurements, or even one measurement. The disclosed methods also may be used with three measurements, and offer the advantages over prior three-measurement methods that the offset does not need to be assumed to be zero as long as it is known prior to measurement, and that relative retardance and angle may be determined. As scale and offset do not vary significantly spatially or temporally in the context of reading data encoded as birefringence in a storage medium, the scale and offset may be determined one time or periodically using a four (or more) measurement technique, and then the determined scale and offset values may be used for subsequent determinations of retardance and angle using fewer than four measurements. Further, as described below, scale and offset, when not known, may be determined via numerical optimization techniques.

FIG. 4 shows a flow diagram illustrating an example method 400 for determining birefringence values. Method 400 is an example of a technique that can be used to determine scale and offset value for use for subsequent determinations using fewer measurements, and offers the advantage over current methods that arbitrary probe light polarization states may be used to make the determination, rather than predetermined states that may be hard to implement via physical instruments. Method 400 first comprises, at 402, acquiring four or more of measurements of a voxel at different polarization states, wherein each measurement comprises an observed intensity m_kat measurement state k defined by a measurement polarization state with swing χ_kand angle θ_k, Method 400 further comprises, at 404, determining a likelihood function for the measurements of the voxel, wherein the likelihood function represents, for each birefringence value set of a plurality of birefringence value sets, a likelihood of the measurements of the voxel being produced by the birefringence value set. The likelihood is determined as a function of angle and retardance using an assumption of a noise model and a model for the data. Any suitable data model and noise model may be used, including Gaussian, Poisson, and a combination of Gaussian and Poisson. As one example, a likelihood of measuring intensity m_kgiven measurement state k using a Gaussian noise model (indicated at 406) is given by equation (1).

P
⁡
(

m
k

|

I
k

)

=

(

2
⁢
π
⁢

σ
2

)

-

1
2

⁢
exp
⁢

(

-

(

I
k

-

m
k

)

2

2
⁢

σ
2

)

(
1
)

In equation (1), I_k=a(1−cos χ_kcos δ+sin χ_ksin δ sin(2ϕ−2θ))+b, where I_kis the expected measurement taken for polarization state k, δ is the sample retardance, ϕ is the sample angle, a is the scale parameter, b is the offset, and σ is the noise. The term “birefringence value set” indicates a set of values for {a, b, δ, ϕ}. Using this expression, a likelihood function to determine a likelihood of a set of measurements can be expressed as {m_k}_k=1^N, as L=Π_k=1^NP(m_k|I_k). L is computed for a plurality of birefringence value sets, and a maximum value is determined for the likelihood function, as indicated at 408. The maximum value of the likelihood function then is used, at 410, to determine a most likely set of birefringence values for the set of measurements of the voxel based upon the set that produced the maximum determined value. This may be expressed mathematically as equation (2).

{a,b,δ,ϕ}=argmax(L(a,b,δ,ϕ)) (2)

The scale and offset determined may be used in determinations using three or fewer measurements. It will be understood that the term maximum determined value, and similar terms used herein, are not meant to indicate an actual global maximum for a likelihood function, but rather to represent a maximum observed value for all sets of parameters utilized. It further will be understood that, in other examples, any other suitable method may be used to determine the birefringence values including scale and offset. For example, in accordance with Bayes' theorem, it is also possible to include prior knowledge about angle or retardance distributions and instead of calculating the likelihood of the parameters, one can calculate the posterior. In this case one could find the values of the parameters that maximized the posterior (often this is called MAP—maximum a posteriori). In such examples, parameter values may be determined, for example, by finding the expectation of the posterior, or by using other statistical measures.

As mentioned above, an advantage offered by the use of a maximum determined likelihood method to determine birefringence values for a voxel is that the polarization states used for the measurements may be arbitrary, rather than predetermined. Nevertheless, in some examples, some configurations of polarization states may provide for better use of available signal than other configurations of states. Two examples of such configurations are as follows. Referring again to the Poincaré sphere of FIG. 3, a first example of a set of polarization states for determining birefringence values for voxel comprises a circularly polarized input state (as set by the PSG) with one-handedness, and elliptically polarized output states (as set by the PSA) with the opposite handedness to the input state and equal ellipticity to each other, but having varying azimuths that are spaced equally around half or all of a circle at an equal latitude on the Poincaré sphere, such as at {0, 45, 90, 135} degrees or {0, 22.5, 45, 67.5} degrees. A second example set of polarization states comprises an output state circularly polarized with one handedness and input states that are elliptically polarized with the opposite handedness to the output state and equal ellipticity to each other but having varying azimuths, wherein the azimuths are equally spaced around half or all of a circle on the surface of the Poincaré sphere. FIG. 5 depicts examples of three such polarization state configurations, each configuration denoted by points with a different symbol from the other configurations. In FIG. 5, the view of the Poincaré sphere is along the polar axis of the sphere, the dashed circle corresponds to a selected latitude comprising the polarization states, and the outer border corresponds to the equator of the sphere.

When measurements are configured in this way, the azimuth angle of the voxel can be determined as follows:

ϕ
=

1
2

⁢
arctan
⁢

(

∑

i
=
1

N

m
k

⁢
sin
⁢
2
⁢

θ
k

∑

i
=
1

N

m
k

⁢
cos
⁢
2
⁢

θ
k

)

,

N
>
3

(
3
)

Here ϕ is the measured angle, the index k indicates the measurement number with m_kbeing the intensity of the k^thmeasurement, and 2θ_kbeing the angle of the k^thmeasurement state on the Poincaré sphere. To find the retardance, one suitable method comprises numerically optimizing a likelihood of the data given the measurements, as described above with regard to FIG. 4.

As mentioned above, prior knowledge of the scale and offset parameters may be used to determine the azimuth angle and retardance of a voxel using a reduced number of measurements. For example, the azimuth angle for a voxel may be determined using three-measurements plus a known offset value by using equation (4).

ϕ
=

1
2

⁢
arctan
⁢

(

∑

i
=
1

N

m
k

⁢
sin
⁢
2
⁢

θ
k

∑

i
=
1

N

m
k

⁢
cos
⁢
2
⁢

θ
k

)

,

N
=
3

(
4
)

The retardance may be found, for example, by numerical maximization of a likelihood function, as described above. Further, even where scale and offset are not initially known, numerical optimization techniques may be used to determine the scale and offset. In such examples, the scale and offset may be determined where information about the expected distribution of angles is known (as may be the case for previously written birefringent voxels). With prior knowledge of the expected distribution of angles, the scale and offset may be adjusted computationally until the expected distribution is achieved to a suitable approximation.

Where some prior information on the retardance is known, but the retardance value itself is not known, a two-measurement method may be used to determine birefringence values for a voxel. In some examples, the two-measurement method also utilizes prior knowledge of scale and offset, while in other examples, the scale and offset are determined by numerical optimization, as mentioned above.

When using two measurements, even with knowledge of the offset and scale, there are still two possible solutions for the retardance and angle of the sample. This is because each measurement restricts the solutions to lie on a 2D plane in the 3D space of the Poincaré sphere. Therefore, the two measurements define a line in 3D. FIG. 6 illustrates examples of two-measurement determinations for a hypothetical voxel on Poincaré sphere 602. In this example, the line 604 defined by the two measurements for the voxel intersects the surface of the sphere at two points 606 and 608.

To determine which of the two points corresponds to the actual birefringence values of the voxel that was measured, a constraint is applied that selects the point with the lower retardance as corresponding to the state of the voxel, as it can be shown mathematically that the other solution is at least as large as the effective retardance of the measurement states. For example, given the two measurements in Poincaré sphere 602, point 606 would be selected since it represents a lower retardance value compared to point 608. This prior information-based constraint may be used in any system where the effective measurement retardance is known to be different (either larger or smaller) than the measurement retardance. This constraint narrows down the solution to one point on the sphere, which allows the sample retardance and azimuth to be determined.

A more detailed mathematical description of an example two-measurement methods is as follows. It will be noted that the method may be performed either by using values for scale and offset that are initially determined using a greater number of measurements (e.g. four), and then applied to the two-measurement determinations, or by determining scale and offset values using numerical optimization methods. Where values for scale and offset are determined initially, any suitable method to determine these parameters may be used, including the maximum likelihood example described above with regard to FIG. 4.

After acquiring measurements, the example two-measurement method first involves solving the following set of simultaneous equations (5, 6, 7) for the retardance δ:

m₁=a(1−cos χ₁cos δ+sin χ₁sin δ sin(2ϕ−2θ₁))+b (5)

m₂=a(1−cos χ₂cos δ+sin χ₂sin δ sin(2ϕ−2θ₂))+b (6)

(sin δ sin(2ϕ−2θ₁))²+(sin δ sin(2ϕ−2θ₂))²+cos²δ=1 (7)

As there are two possible solutions for δ, the smaller value is selected, as described above. Next, the azimuth angle can be computed using equation (8),

ϕ
=

1
2

⁢
arctan
⁢

(

sin
⁡
(

2
⁢
ϕ

-

2
⁢

θ
1

)

sin
⁡
(

2
⁢
ϕ

-

2
⁢

θ
2

)

)

+

θ
2

-

θ
1

⁢

when
⁢

mod
⁢
 

(

θ
2

-

θ
1

,
 

π
2

)

=

π
4

.

(
8
)

Where known values of the scale and offset were used in this determination, then the determination is complete at this stage. On the other hand, where assumed scale and offset values were used, then prior information may be used to adjust a and b. For example, where the distribution of angles will be uniform (as may be the case with birefringent data, where the writing process for the data is known), then the distribution of angles may be computed, and equations (5)-(7) and (8) may be repeatedly determined while adjusting a and b until the measured angular histogram is sufficiently similar to the expected one.

Where the retardance of a voxel is known (as the case may be where the birefringence writing properties are known), birefringence values may be determined via a single measurement, again using known or assumed values for the scale and offset parameters. Where assumed values are used, the values can be numerically optimized. Based upon the known or assumed scale and offset values, the single measurement 612 defines a two-dimensional plane in three-dimensional space, with reference to the Poincaré sphere 610 of FIG. 6. This two-dimensional plane intersects the surface of the sphere along a circle 614, where the circle represents a continuous range of possible sample angles and retardance values. Where the sample retardance (represented by angle 616) is known accurately, then the angle can be determined to within one of two values (represented by points 620 and 622), namely, where the retardance plane 618 intersects the measurement circle 614. In comparison, the above-described two-measurement example utilized less-detailed prior information, namely, that the retardance is less than the swing. Where it is known that the angle is in some range that spans only half of the total available angles, for example between 0 and 90 degrees, or between 45 and 135 degrees (or any other 90 degree range) then these two points can be restricted down to one, which determines the retardance and angle of the specimen. Mathematically, the computation to perform the disclosed example one-measurement determination is expressed by equations (9) and (10).

m
1

=

a
⁡
(

1
-

cos
⁢

χ
1

⁢
cos
⁢

δ

+

sin
⁢

χ
1

⁢
sin
⁢
δsin
⁢

(

2
⁢
ϕ

-

2
⁢

θ
1

)

)

+
b

(
9
)

ϕ
=

1
2

⁢
arcsin
⁢

(

1
-

(

m
1

-
b

)

a

-

cos
⁢

χ
1

⁢
cos
⁢
δ

sin
⁢

χ
1

⁢
sin
⁢
δ

)

+

θ
1

(
10
)

FIG. 7 shows a flow diagram depicting an example method 700 for determining birefringence values for a voxel using two or fewer measurements. As described above, the method of FIG. 7 may utilize scale and offset parameters determined via the method of FIG. 4, or may initially assume and then numerically optimize these parameters. Method 700 comprises, at 702, acquiring measurement data for a birefringent voxel by directing probe light comprising one or more predetermined polarization states through the birefringent voxel and receiving the light at the image sensor. In some examples, as indicated at 704, the measurement data may comprise measurement data acquired using light of a first polarization state, and measurement data acquired using light of a second, different polarization state. In other examples, the measurement data may comprise data from a single measurement, as indicated at 706.

Continuing, method 700 comprises, at 708, based upon the measurement data, determining two points on a surface of a Poincaré sphere corresponding to two possible birefringent states of the birefringent voxel each state comprising a set of birefringence values, and applying a constraint to determine birefringence values for the voxel. For example, where measurements are taken at two polarization states, two azimuth angle solution exist to equations (5)-(7) above. In this instance, method 700 comprises, at 710, applying a constraint specifying that the set of birefringence values having the lower retardance value is the correct set. By selecting the point on the Poincaré sphere that represents the lower retardance, the azimuth angle may be solved using equation 8 above, thereby providing a determined set of birefringence values for the voxel.

Where a single measurement is used, method 700 comprises, at 712, determining a circle on the surface of the Poincaré sphere based upon the measurement, the circle comprising two locations that intersect a plane representing a known retardance of the voxel. Then, at 714, the point representing the birefringence values of the voxel may be selected based upon the azimuth angle of the point being within an expected range of angles, and the birefringence values may be determined using equations (9) and (10) above. Method 700 further optionally comprises, at 716, numerically optimizing a likelihood function based upon the determined azimuth and retardance values to determine scale and offset values in instances where these values are not known initially. Method 700 further comprises, at 718, outputting the birefringence values of the voxel.

The above-described examples may help to reduce the time and computing resources consumed when reading birefringent voxels in a storage medium compared to methods that use four or more measurements. Other processes also may alternatively or additionally be used to provide for the efficient reading of birefringent voxels. For example, wavelength multiplexing may be used to reduce a number of individual images acquired during a reading process.

FIG. 8 shows a schematic depiction of an example system 800 for reading birefringent storage medium. System 800 utilizes wavelength multiplexing in which N different wavelength bands of light having different polarization states are multiplexed to acquire a number N of measurements in a temporally overlapping manner. System 800 comprises N light sources, illustrated as first light source 802, second light source 804, and N^thlight source 806, each configured to output light of a different wavelength band (e.g. red, green and blue). Each light source directs light into a corresponding PSG, shown respectively for light sources 802, 804 and 806 as PSG 808, PSG 810, and PSG 812, to allow a different polarization state to be set for each wavelength band. In other examples, a system to perform two temporally overlapping measurements may have two light sources and corresponding PSGs.

Light from each PSG is directed toward a storage medium 814 that is placed in a sample region of the system for reading the storage medium. The term “sample region” is used herein to represent a location at which a storage medium is placed for reading. In the depicted example, N−1 beam combiners (illustrated as beam combiner 1 816 and beam combiner N−1 818) are used to combine the light from each PSG into a beam for probing the sample medium.

In the depicted embodiment, an optic in the form of a condenser lens 820 directs light through the storage medium, and an objective lens 817 focuses the light onto a detector in the form of an image sensor 822 that images an entirety of a data layer within the storage medium 814 per image frame. An achromatic PSA 824 is positioned between the storage medium and the image sensor 822. The image sensor 822 comprises a plurality of integrated wavelength-selective bandpass filters 825A-C, such that light of different wavelength bands passes through different filters and onto different areas of a pixel of the image sensor 822. In the example depicted in FIG. 8, image sensor 822 comprises a first bandpass filter 825A, a second bandpass filter 825B, and a third bandpass filter 825C. In this manner, an intensity of each wavelength band of light (each of which has a different polarization setting) may be measured in a same image frame.

In some examples, one or more physical masks may be used for pupil engineering to help enhance a quality of a signal used to read birefringence-encoded data in a storage medium. Example masks that may be used are illustrated schematically as intensity mask 825 and phase mask 826. The engineering of the pupil may depend upon a layout (e.g. an x, y, z spatial distribution) of voxels in a storage medium. For example, a ring-shaped intensity mask added to the pupil of a lens generates a Bessel beam rather than the conventional Gaussian beam. Phase masks can also be applied to engineer the polarization field at sample plane, to optimize the pupil profile for the intended type of birefringence distribution of the sample. Thus, if a specific measuring optical probe is desirable, by engineering the shape of the input light, the signals that can be used as input for these methods can be tailored with the purpose of enhancing the quality of the measurements.

In some examples, adjustable focus optics may be moved to selectively focus voxels at particular layers within a volume of the storage medium 814, and thereby allow the reading of different layers. In other examples, the storage medium may be moved to focus on voxels in different layers. While the depicted image sensor includes integrated bandpass filters, in other examples bandpass filters may be included elsewhere in a system. For example, a system may utilize plural PSGs and wavelength multiplexing in combination with a rotating bandpass filter (e.g. a color wheel) to allow the sequential acquisition of images of different polarization states.

As mentioned above, in some examples a lesser number of light sources may be used to produce a greater number of wavelength bands having different polarization states. FIG. 9 shows an example light source configuration in which light from a single light source 902 is split via a dichroic beam splitter 904 into two beams 906, 908 of different wavelength bands. Beams 906 and 908 are directed through respective PSGs 912, 914 using any suitable optics (such as mirrors 916, 918 in the depicted example), and PSGs 912, 914 set different polarization states for beams 906 and 908. After passing through the PSGs, beams 906, 908 are combined with dichroic beam combiner 920 for probing a storage medium. In other examples, light from a suitable light source may be split into three or more different wavelength bands.

FIGS. 10A-10B show a flow diagram depicting an example method 1000 for acquiring plural temporally overlapping measurements via wavelength multiplexing. First, referring to FIG. 10A, method 1000 comprises, at 1002, generating first polarized light of a first wavelength band, the first polarized light comprising a first polarization state, and generating second polarized light of a second wavelength band different than the first wavelength band, the second polarized light comprising a second polarization state that is different from the first polarization state. In some examples, light of the first wavelength band is output via a first light source and light of the second wavelength band is output by a second light source, as indicated at 1004. In other examples, light may be output from a lesser number of light sources, and then split into a greater number of beams of different wavelength bands, as indicated at 1006. Further, it will be understood that additional wavelength bands beyond two may be used. As such, method 1000 may comprise, at 1008, outputting light of three or more different wavelength bands having different polarization states to generate third polarized light and potentially additional other polarized light beams.

Continuing, method 1000 comprises, at 1010, passing the first polarized light and the second polarized light through a voxel of a storage medium, thereby changing the first polarization state of the first polarized light to a first modified polarization state, and changing the second polarization state of the second polarized light to a second modified polarization state. Passing the first polarized light and the second polarized light through the voxel of the storage medium may comprise, at 1012, combining the first polarized light and the second polarized light via a beam combiner before passing the first and second polarized light through the storage medium. Further, process 1010 also may comprise, at 1014, combining third polarized light, and any additional wavelength bands of polarized light, with the first and second polarized light before passing the light through the storage medium. In some examples, one or more masks may be used to implement an engineered pupil. As such, method 1000 may comprise, at 1016, passing the polarized light through an intensity mask before passing the polarized light through the voxel of the storage medium. Alternatively or additionally, method 1000 may comprise, at 1018, passing the polarized light through a phase mask before passing the polarized light through a voxel of the storage medium.

Next referring to FIG. 10B, method 1000 comprises, at 1022, passing the first polarized light, the second polarized light and any additional wavelength bands of polarized light through a polarization state analyzer after passing the polarized light through the voxel of the storage medium, wherein the analyzer attenuates an intensity of the light based upon a polarization state of the light compared to a state of the analyzer. Then, at 1024, method 1000 comprises passing the first polarized light through a first bandpass filter onto the image sensor, and passing the second polarized light through a second bandpass filter and onto the image sensor, wherein the first bandpass filter selectively passes the first wavelength band and the second bandpass filter selectively passes the second wavelength band. Further, as indicated at 1026, additional wavelength bands of polarized light having different polarization states may be passed through corresponding additional bandpass filters. In this manner, intensities of the first wavelength band, the second wavelength band, and any additional wavelength bands, are separately imaged.

In some examples, the first bandpass filter, the second bandpass filter, and any additional bandpass filters are integrated with the image sensor as spatially separate filters integrated with a pixel of the image sensor. In such examples, method 1000 comprises, at 1028, passing the first polarized light onto a first region of the image sensor, passing the second polarized light onto a second region of the image sensor, and passing any additional wavelength bands of polarized light onto the image sensor. In this manner, images for each polarization state can be separately acquired in a same image frame. In other examples, the different bandpass filters may be applied in a time-multiplexed manner, such as by a color wheel. In such examples, different image frames are acquired for each wavelength band. After acquiring the measurements using the different wavelength bands for different polarization settings, method 1000 comprises, at 1030, determining birefringence values for the voxel based upon the first polarized light received at the image sensor via the first bandpass filter and the second polarized light received at the image sensor via the second bandpass filter, plus any additional wavelength-multiplexed polarized light. The birefringence values may be determined using the example methods disclosed above, or in any other suitable manner.

With any of the above-described methods, the system and medium that probe light passes through when probing a voxel can impart additional rotation and/or retardance on the polarized light used to probe voxel. For example, reading a layer of voxels in a data storage medium comprising a three-dimensional array of voxels may result in a state of polarized probe light being rotated by passing through other out-of-focus layers of voxels in a read process. System imperfections also may lead to background noise. As such, it is often of interest to measure the birefringent properties of a sample such as a data storage medium relative to imperfections in the system and storage medium (e.g. the presence of other voxels in the optical path that are not being read). This is referred to as background subtraction.

Current methods involve the removal of background signal by capturing a set of images {b_k} in addition to the measurement set {m_k} and computing the angle and retardance of the background. The {b_k} are captured using the same instrument with the same set of polarization states as the {m_k}. The {b_k}, are captured simply by removing the actual sample from the field of view. It is also possible to estimate {b_k} from {m_k} by for example performing a blurring operation. The blur can be achieved by a simple low-pass filter. However, such methods may not correct adequately for both system imperfections and sample imperfections such as other voxels.

As such, examples are disclosed that relate to estimating {b_k} from {m_k} based on a two-step process in which two low-passed versions of the measured intensities are calculated and combined using a multiplicative constant to form an enhanced background image. In this manner, both a local and a global background intensity are estimated, which may represent a more accurate estimation for the background intensities at the large and small scales, as the large-scale background may compensate for the imperfect system while the small-scale background may compensate for the three-dimensional nature of the storage medium. The filter parameters and multiplicative constant may be derived, for example, by performing a minimization of an error between a known specimen's retardance and calculated retardance distributions. Such a background correction method may be more efficient than separately acquiring a background image for the system (e.g. without a data storage medium present in the sample region), as fewer physical measurement processes are performed, thereby saving time and resources utilized for a separate physical background measurement. As a more specific example, for a sample with known data encoded into the glass and for which a set of image frames have been acquired, initial filter parameters and multiplicative constant are set, and the encoded data is decoded. This is repeated with updated filter parameters and multiplicative constant. The parameters and multiplicative constant that give a determined lowest error in decoding are then selected. The process of updating the filter parameters can be a brute search in the parameter space or a gradient method in various examples.

FIG. 11 shows a flow diagram depicting an example method 1100 for correcting birefringence values using a plurality of low-pass filters. Method 1100 comprises, at 1102, acquiring an intensity image of a voxel of a storage medium. To measure the birefringence of the voxel, multiple images of the voxel are acquired at different polarization states, as indicated at 1104. In some examples, the image may be of a plurality of voxels arrayed in an image plane within the storage medium, as indicated at 1106. The image of the voxel comprises higher frequency image information arising from a birefringence state of the voxel and lower frequency image information arising from one or more birefringent regions of the storage medium other than the voxel.

At 1108, method 1100 comprises applying a first low-pass filter to the image of the voxel to obtain a first background image, and applying a second low-pass filter to the image of the voxel to obtain a second background image. As described above, the first and second low-pass filters have different cut-off frequencies, such that the first background image may represent more local background characteristics, such as birefringence noise arising from other voxels in the storage medium, while the second background image may have a lower cut-of frequency than the first low-pass filter and represent more global background characteristics. The effect of the more localized low-pass filter may be to remove high-frequency data representing more detailed features of a voxel that is in focus, thereby leaving less detailed features from other voxels that are out-of-focus during the read process. As indicated at 1110, the first and second low-pass filters are applied to each measurement image acquired for the voxel to form a background image for each measurement image.

Continuing, method 1100 includes, at 1112, determining an enhanced background image from the first background image and the second background image. The enhanced background image may be determined in any suitable manner. In some example, the enhanced background image may be determined by combining the first background image and the second background image using a multiplicative constant, as indicated at 1114. A more specific example utilizes the following equation:

b_k=a(LowPass₁(m_k)−LowPass₂(m_k))+LowPass₂(m_k), (11)

where measured intensities=m_k, local background intensities are represented by LowPass₁(m_k), global background intensities are represented by LowPass₂(m_k), a is a multiplicative constant used as a scale factor, and b_kis the enhanced background image.

Method 1100 further includes, at 1116, determining birefringence values for the enhanced background image and, at 118, birefringence values for the image of the voxel. Each set of birefringence values comprises a retardance value and an azimuth value angle, and may be determined using the examples described above, or in any other suitable manner.

Method 1100 further comprises, at 1120, correcting the birefringence values for the image of the voxel based on the birefringence values for the enhanced background image. Correcting the birefringence values for the image may comprise, for example, determining a relative rotation and relative retardance of the birefringence values for the image of the voxel compared to the birefringence values for the enhanced background image, as indicated at 1122. In one example, measurement retardance δ_mand angle θ_mare from {m_k}, and the background retardance δ_band angle θ_bare determined from {b_k}, for example, using any of the methods described above, or other suitable method. The relative angle θ_rand retardance δ_rare determined by the following equations.

δ
r

⁢
sin
⁢
2
⁢

θ
r

=

δ
m

⁢
sin
⁢
2
⁢

θ
m

-

δ
b

⁢
sin
⁢
2
⁢

θ
b

=
U

(
12
)

δ
r

⁢
cos
⁢
2
⁢

θ
r

=

δ
m

⁢
sin
⁢
2
⁢

θ
m

-

δ
b

⁢
sin
⁢
2
⁢

θ
b

=
V

(
13
)

δ
r

=

U
2

+

V
2

(
14
)

θ
r

=

1
2

⁢

arctan
⁡
(

U
V

)

(
15
)

In some embodiments, the methods and processes described herein may be tied to a computing system of one or more computing devices. In particular, such methods and processes may be implemented as a computer-application program or service, an application-programming interface (API), a library, and/or other computer-program product.

FIG. 12 schematically shows a non-limiting embodiment of a computing system 1200 that can enact one or more of the methods and processes described above. Computing system 1200 is shown in simplified form. Computing system 1200 may take the form of one or more personal computers, server computers, tablet computers, home-entertainment computers, network computing devices, gaming devices, mobile computing devices, mobile communication devices (e.g., smart phone), and/or other computing devices.

Computing system 1200 includes a logic subsystem 1202 and a storage subsystem 1204. Computing system 1200 may optionally include a display subsystem 1206, input subsystem 1208, communication subsystem 1210, and/or other components not shown in FIG. 12.

Logic subsystem 1202 includes one or more physical devices configured to execute instructions. For example, the logic subsystem may be configured to execute instructions that are part of one or more applications, services, programs, routines, libraries, objects, components, data structures, or other logical constructs. Such instructions may be implemented to perform a task, implement a data type, transform the state of one or more components, achieve a technical effect, or otherwise arrive at a desired result.

The logic subsystem may include one or more processors configured to execute software instructions. Additionally or alternatively, the logic subsystem may include one or more hardware or firmware logic subsystems configured to execute hardware or firmware instructions. Processors of the logic subsystem may be single-core or multi-core, and the instructions executed thereon may be configured for sequential, parallel, and/or distributed processing. Individual components of the logic subsystem optionally may be distributed among two or more separate devices, which may be remotely located and/or configured for coordinated processing. Aspects of the logic subsystem may be virtualized and executed by remotely accessible, networked computing devices configured in a cloud-computing configuration.

Storage subsystem 1204 includes one or more physical devices configured to hold instructions executable by the logic subsystem to implement the methods and processes described herein. When such methods and processes are implemented, the state of storage subsystem 1204 may be transformed—e.g., to hold different data.

Storage subsystem 1204 may include removable and/or built-in devices. Storage subsystem 1204 may include optical memory (e.g., CD, DVD, HD-DVD, Blu-Ray Disc, etc.), semiconductor memory (e.g., RAM, EPROM, EEPROM, etc.), and/or magnetic memory (e.g., hard-disk drive, floppy-disk drive, tape drive, MRAM, etc.), among others. Storage subsystem 1204 may include volatile, nonvolatile, dynamic, static, read/write, read-only, random-access, sequential-access, location-addressable, file-addressable, and/or content-addressable devices.

It will be appreciated that storage subsystem 1204 includes one or more physical devices. However, aspects of the instructions described herein alternatively may be propagated by a communication medium (e.g., an electromagnetic signal, an optical signal, etc.) that is not held by a physical device for a finite duration.

Aspects of logic subsystem 1202 and storage subsystem 1204 may be integrated together into one or more hardware-logic components. Such hardware-logic components may include field-programmable gate arrays (FPGAs), program- and application-specific integrated circuits (PASIC/ASICs), program- and application-specific standard products (PSSP/ASSPs), system-on-a-chip (SOC), and complex programmable logic devices (CPLDs), for example.

The term “program” may be used to describe an aspect of computing system 1200 implemented to perform a particular function. In some cases a program may be instantiated via logic subsystem 1202 executing instructions held by storage subsystem 1204. It will be understood that different programs may be instantiated from the same application, service, code block, object, library, routine, API, function, etc. Likewise, the same program may be instantiated by different applications, services, code blocks, objects, routines, APIs, functions, etc. The term “program” may encompass individual or groups of executable files, data files, libraries, drivers, scripts, database records, etc.

It will be appreciated that a “service”, as used herein, is an application program executable across multiple user sessions. A service may be available to one or more system components, programs, and/or other services. In some implementations, a service may run on one or more server-computing devices.

When included, display subsystem 1206 may be used to present a visual representation of data held by storage subsystem 1204. This visual representation may take the form of a graphical user interface (GUI). As the herein described methods and processes change the data held by the storage subsystem, and thus transform the state of the storage subsystem, the state of display subsystem 1206 may likewise be transformed to visually represent changes in the underlying data. Display subsystem 1206 may include one or more display devices utilizing virtually any type of technology. Such display devices may be combined with logic subsystem 1202 and/or storage subsystem 1204 in a shared enclosure, or such display devices may be peripheral display devices.

When included, input subsystem 1208 may comprise or interface with one or more user-input devices such as a keyboard, mouse, touch screen, or game controller. In some embodiments, the input subsystem may comprise or interface with selected natural user input (NUI) componentry. Such componentry may be integrated or peripheral, and the transduction and/or processing of input actions may be handled on- or off-board. Example NUI componentry may include a microphone for speech and/or voice recognition; an infrared, color, stereoscopic, and/or depth camera for machine vision and/or gesture recognition; a head tracker, eye tracker, accelerometer, and/or gyroscope for motion detection and/or intent recognition; as well as electric-field sensing componentry for assessing brain activity.

When included, communication subsystem 1210 may be configured to communicatively couple computing system 1200 with one or more other computing devices. Communication subsystem 1210 may include wired and/or wireless communication devices compatible with one or more different communication protocols. As non-limiting examples, the communication subsystem may be configured for communication via a wireless telephone network, or a wired or wireless local- or wide-area network. In some embodiments, the communication subsystem may allow computing system 1200 to send and/or receive messages to and/or from other devices via a network such as the Internet.

Another example provides a system for reading birefringent data, the system comprising one or more light sources, a first polarization state generator positioned to generate first polarized light from light of a first wavelength band output by the one or more light sources, a second polarization state generator positioned to generate second polarized light from light of a second wavelength band output by the one or light sources, an image sensor configured to acquire an image of a sample region through which the first polarized light and the second polarized light passes, a polarization state analyzer disposed optically between the sample region and the image sensor, a first bandpass filter disposed optically between the polarization state analyzer and the image sensor, the first bandpass filter configured to pass light of the first wavelength band, and a second bandpass filter disposed optically between the polarization state analyzer and the image sensor, the second bandpass filter configured to pass light of the second wavelength band. In some such examples, the system further comprises a third polarization state generator positioned to generate third polarized light from light of a third wavelength band output by the one or more light sources, the third wavelength band being different than the first wavelength band and the second wavelength band, and a third bandpass filter disposed optically between the polarization state analyzer and the image sensor, the third bandpass filter configured to pass light of the third wavelength band. In some such examples, the system further comprises a beam combiner configured to combine the first polarized light and the second polarized light optically upstream of the sample region. In some such examples, the system the first bandpass filter and the second bandpass filter may be integrated with a pixel of the image sensor. In some such examples, the one or more light sources comprise a first light source configured to output light of the first wavelength band, and a second light source configured to output light of the second wavelength band. In some such examples, the system further comprises a computing system comprising instructions executable to determine birefringence values for a voxel of a storage medium placed in the sample region based upon light received at the image sensor via the first bandpass filter and light received at the image sensor via the second bandpass filter. In some such examples, the system further comprises one or more of an intensity mask and a phase mask positioned optically upstream of the sample region.

Another example provides a method for performing birefringence measurements, the method comprising generating first polarized light from light of a first wavelength band output by one or more light sources, the first polarized light comprising a first polarization state, generating second polarized light from light of a second wavelength band output by the one or light sources, the second wavelength band being different than the first wavelength band, and the second polarized light comprising a second polarization state that is different from the first polarization state, passing the first polarized light and the second polarized light through a voxel of a storage medium, thereby changing the first polarization state to a first modified polarization state and changing the second polarization state to a second modified polarization state, passing the first polarized light and the second polarized light through a polarization state analyzer after passing the first polarized light and the second polarized light through the voxel of the storage medium, passing the first polarized light through a first bandpass filter and onto a first region of an image sensor, and passing the second polarized light through a second bandpass filter and onto a second region of the image sensor. In some such examples, the method further comprises generating third polarized light from light of a third wavelength band output by the one or more light sources, the third wavelength band being different than the first wavelength band and the second wavelength band, and the third polarized light comprising a third polarization state that is different than the first polarization state and the second polarization state, and passing the third polarized light through a third bandpass filter disposed optically between the polarization state analyzer and the image sensor, the third bandpass filter configured to pass light of the third wavelength band. In some such examples, the method further comprises combining the first polarized light and the second polarized light via a beam combiner before passing the first polarized light and the second polarized light through the voxel of the storage medium. In some such examples, the method further comprises outputting light of the first wavelength band via a first light source, and outputting light of the second wavelength band via the second light source. In some such examples, the method further comprising determining birefringence values for the voxel of the storage medium based upon the first polarized light received at the image sensor via the first bandpass filter and the second polarized light received at the image sensor via the second bandpass filter. In some such examples, the method further comprises passing the first polarized light and the second polarized light through one or more of an intensity mask and a phase mask before the voxel of the storage medium.

Another example provides, on a computing device, a method of determining birefringence values for a birefringent voxel of a storage medium, the method comprising acquiring measurement data for the birefringent voxel by directing light comprising one or more predetermined polarization states through the birefringent voxel and receiving the light at an image sensor, based upon the measurement data, determining two points on a surface of a Poincaré sphere corresponding to two possible birefringent states of the birefringent voxel, each state comprising a set of birefringence values including an azimuth angle and a retardance, apply a constraint to determine an azimuth angle and a retardance, and output determined birefringence values comprising the azimuth angle and the retardance determined. In some such examples, the measurement data comprises measurement data from a first measurement using light of a first polarization state, and measurement data from a second measurement using light of a second polarization state. In some such examples, two points on the Poincaré sphere are determined based upon the measurement data from the first measurement and the measurement data from the second measurement, and wherein applying the constraint comprises selecting from the two points on the Poincaré sphere a point representing a set of birefringence values having a lower retardance value. In some such examples, the method further comprises acquiring the first measurement and the second measurement using light of a first wavelength band having the first polarization state multiplexed with light of a second wavelength having the second polarization state. In some such examples, the measurement data comprises measurement data from a single measurement. In some such examples, determining the two points on the Poincaré sphere comprises determining a circle on the Poincaré sphere based upon the measurement, wherein the two points comprise locations at which the circle intersects a plane representing a known retardance of the voxel, and wherein applying the constraint comprises selecting from the two points a point representing a set of birefringence values having an azimuth angle within a range of expected angles. In some such examples, the method further comprises determining an offset parameter and a scale parameter via a numerical optimization after determining an azimuth angle.

Another example provides, on a computing device, a method for reading data stored as birefringence values in a storage medium, the method comprising acquiring a measurement comprising an intensity m_kimage of a voxel of the storage medium using light of a polarization state k, determining a likelihood function for the image of the voxel, the likelihood function representing a likelihood, for each birefringence value set of a plurality of possible birefringence value sets, of an intensity of the voxel being produced by the birefringence value set, the likelihood function being based upon a selected data model and a selected noise model; determine a maximum value of the likelihood function, and determine a most likely set of birefringence values for the voxel based upon a birefringence value set that produce the maximum value of the likelihood function. In some such examples, the noise model comprises a Gaussian noise model. In some such examples, the likelihood of measuring an intensity m_kof an image of a voxel for a measurement state k is given by:

P
⁡
(

m
k

|

I
k

)

=

(

2
⁢
π
⁢

σ
2

)

-

1
2

⁢

exp
⁡
(

-

(

I
k

-

m
k

)

2

2
⁢

σ
2

)

,

where polarization state k comprises a swing χ_kand an angle θ_k,

where I_k=a(1−cos χ_kcos δ+sin χ_ksin δ sin(2ϕ−2θ))+b and is an expected measurement taken for the polarization state k,

where δ is a retardance of the voxel of the storage medium,

where ϕ is an azimuth angle of the voxel of the storage medium,

where a is a scale parameter,

where b is an offset, and

where σ is a representation of noise.

In some such examples, the likelihood function, for a set of measurements {m_k}_k=1^N, is represented by L and is given by L=Π_k=1^NP(m_k|I_k), and wherein the method comprises computing L for a plurality of values of each of a, b, δ and ϕ, and determining the most likely set of birefringence values as {a, b, δ, ϕ}=argmax(L(a, b, δ, ϕ)).

It will be understood that the configurations and/or approaches described herein are exemplary in nature, and that these specific embodiments or examples are not to be considered in a limiting sense, because numerous variations are possible. The specific routines or methods described herein may represent one or more of any number of processing strategies. As such, various acts illustrated and/or described may be performed in the sequence illustrated and/or described, in other sequences, in parallel, or omitted. Likewise, the order of the above-described processes may be changed.

The subject matter of the present disclosure includes all novel and non-obvious combinations and sub-combinations of the various processes, systems and configurations, and other features, functions, acts, and/or properties disclosed herein, as well as any and all equivalents thereof.