1: \documentstyle[preprint,aps,graphicx]{revtex}
2: \tightenlines
3: \begin{document}
4: \newcommand{\bref}[1]{eq.~(\ref{#1})}
5: \newcommand{\be}{\begin{equation}}
6: \newcommand{\en}{\end{equation}}
7: \newcommand{\bs}{$\backslash$}
8: \newcommand{\us}{$\_$}
9:
10: \title{Diffusers for holographic stereography}
11: \author{Lars Egil Helseth${}^{1}$, Ingar Singstad${}^2$}
12: \address{${}^1$ University of Oslo, Department of Physics, N-0316 Oslo, Norway }
13: \address{${}^2$ University of Bergen, Department of Physics, N-5007 Bergen, Norway }
14:
15: \maketitle
16: \begin{abstract}
17: Holographic diffusers have long been recognized as versatile
18: components with a broad number of applications. In this work we discuss
19: holographic diffusers for projection of laser light from a Liquid Crystal Display(LCD) onto a
20: holographic recording medium. In holographic stereography, projection of the
21: information from a LCD onto a holographic recording medium has traditionally
22: been done by a lens or a ground glass. It is suggested that the holographic
23: diffusers can replace these elements and improve image quality and light
24: economy.
25:
26: \end{abstract}
27:
28: \narrowtext
29:
30: \newpage
31:
32:
33: \section{Introduction}
34: Storage of data by means of holography has long been an active field of
35: research. Usually, the
36: storage must be done in coherent light. However, by the
37: application of holographic stereography, this requirement can be relaxed.
38: A holographic stereogram is created by optical multiplexing of a finite number of
39: viewpoints onto different regions of a high resolution recording medium
40: (see ref.\cite{Okada,Klug} and references therein).
41: The perspectives(or viewpoints) are usually displayed on a Liquid Crystal
42: Display(LCD), and then projected onto a small area of the recording medium.
43: Traditionally the multiplexing have been done in one of two ways\cite{Okada,Klug};
44: 1)A cylindrical lens is placed behind a LCD, focusing the laser light into a small slit. This kind of
45: holographic stereogram is usually refered to as a multiplex stereogram. The cylindrical lens conserve the image plane
46: characteristics in the vertical direction, and gives rise to a rainbow colored 3D
47: image\cite{Okada}. Unfortunately, the intensity distribution at the focal plane
48: of the cylindrical lens usually contains large variations due to diffraction.
49: Unless the recording media has large dynamic range, these intensity variations
50: gives rise to saturation and loss of information\cite{Klug}.
51: 2)The alternative is to produce a socalled DeBitetto stereogram\cite{Klug,DeBitetto}.
52: In this case an ordinary diffuser is placed right behind the LCD, and diffuses the light
53: in both the horizontal and vertical direction.
54: Since the slit width is of order 1mm, most of the light from the diffuser is
55: scattered outside the slit, giving very long exposure times. To reduce the
56: exposure times, a ground glass with finite diffusing angle is applied.
57: Unfortunately, this results in an uneven illuminated 3D image,
58: especially when the ground glass is close to the holographic film.
59: Note also that a ground glass will diffuse the light in the vertical direction as well, thus destroying the image plane characteristics. The DeBitetto stereogram is
60: therefore not suitable for production of white light viewable transmission
61: stereograms on thin emulsions. However, the image may be transferred to the
62: plane of a second filmplate. Such image plane stereograms can be displayed in
63: white light with very little color blur\cite{Klug}.
64:
65: Ideally, the intensity distribution at the emulsion plane should be as rectangular as possible,
66: it should contain equal contributions from all parts of the LCD(to obtain an
67: even illuminated 3D image), and all the incoming light should be utilized.
68: Furthermore, if the stereogram is recorded on a thin
69: silver halide emulsion, the image plane characteristics must be preserved, to
70: suppress cross-talk images\cite{Klug}. None of the traditional techniques
71: fulfill all these requirements. Therefore, alternative solutions must be found.
72:
73: Pseudorandom diffusers seems to be a mature solution for both
74: volume data storage and holographic stereograms\cite{Nakayama,Yamaguchi,Gao},
75: due to their ability to distribute the light at the focal plane, without destroying
76: the image plane characteristics.
77: Pseudorandom diffusers were introduced in holographic stereography by Klug $et$
78: $al$\cite{Klug}, and independently by Yamaguchi $et$ $al$\cite{Yamaguchi}, and
79: were shown to increase the image quality of the holographic stereograms.
80:
81: In this paper we propose two different holographic diffusers. The first type, which
82: we will simply call a 'holographic diffuser', can be considered as an alternative to the classical
83: diffuser in a DeBitetto stereogram. It is capable of directing more light
84: into the slit in the holoprinter, and perhaps also improve the image quality,
85: by providing a more even illuminated image than that resulting from a ground
86: glass. The second type, which we will call a 'multiple-point holographic diffuser', can
87: be considered as an alternative to the cylindrical lens in a multiplex
88: stereograms. We think it is capable of improving the image quality of multiplex
89: stereograms, without significant reduction of the intensity incident on the
90: recording medium. To confirm the quality of the holographic diffusers described above,
91: we recorded a few holographic stereograms with the setup shown in fig. \ref{f6},
92: which will be described later.
93: This system will be referred to as the holoprinter.
94: We would like to point out that the diffusers should be designed according to standard
95: stereographic projection relationships, to ensure that the final
96: images are free of anamorphic distortions(see e.g. \cite{Okada,Klug,DeBitetto}
97: and references therein).
98:
99:
100: \section{Holographic diffuser}
101: As mentioned in the previous section, application of a ground glass results
102: in loss of light and uneven illumination of the different parts of the 3D image.
103: A solution to this problem may be holographic diffusers.
104: Holographic diffusers have been developed for many display
105: purposes\cite{Pawluczyk}. They are
106: superior to classical diffusers in that they can redirect and control up to
107: $100\%$ of the incoming light. A holographic diffuser suitable for holographic
108: stereography can be created as follows:
109: 1)First a strip diffuser is created by covering an isotropic diffuser with a
110: slit.
111: 2)Next a hologram is recorded with this strip diffuser as the object. A hologram recorded in this way will simply
112: be called a $holographic$ $diffuser$.
113:
114: To use this holographic diffuser in the production of a holographic
115: stereogram, it must be placed in contact with the LCD screen. As pointed out
116: in the introduction, the purpose of the LCD is to provide a perspective picture
117: which can be projected onto a small area of the holographic film.
118: By illuminating the LCD(+holographic diffuser) with a proper reconstruction
119: beam, the real image of the strip diffuser is projected
120: into the slit of the holoprinter. We believe that the real image of the
121: slit should match that of the holoprinter since this ensures
122: that the whole perspective picture is passed through the slit in the
123: holoprinter, and that maximum amount of light is utilized. If the slit width of
124: the holographic difuser is too small, we probably get picket fence
125: artifacts(a detailed description of this problem is given in ref.\cite{Klug}). If it is too big, some of the perspective
126: picture is cut off, and there is also some loss of light.
127:
128: To produce a holographic diffuser using the principles described
129: above, we covered a diffuser(opal diffuser) with a 0.5mm slit. This strip diffuser was placed
130: slightly away from focus of a cylindrical lens, see fig. \ref{f1}. The lens was
131: introduced to increase the amount of light passing through the strip diffuser, without introducing too
132: large speckles. Since the lens was placed in front of the diffuser, its quality
133: was not important. A collimated reference beam was illuminating the film at an angle of
134: $\sim 50^{\circ}$, and the distance between the diffuser and the holographic
135: films was approximately 30cm. With this setup we recorded a hologram with a 20mW HeNe laser
136: on BB640 thin silver halide emulsions from
137: HRT technologies. The size of the exposed film was 12.7$\times$10cm, and we used
138: a collimated reference beam which was larger than the film(and therefore
139: considered uniform). The film was processed by a procedure similar to that
140: suggested by Belendez et.al.\cite{Belendez}, resulting in Silver Halide
141: Sensitized Gelatin(SHSG). In this way we were able to produce a holographic
142: diffuser with little unwanted scattering and $40\%$ diffraction
143: efficiency. During reconstruction, the real image will form a rectangular
144: strip of 0.5mm width. We found that the effective(actual) size of the diffuser was
145: 10$\times$10cm(the holographic film was partially screened by the holder).
146: Since the distance between the original strip diffuser and the holographic film
147: was 30cm, this results in a convergence angle of approximately
148: $\sim 9.5^{\circ}$ during reconstruction.
149:
150:
151: \section{Multiple-point holographic diffuser}
152: In many applications we are interested in conserving the image plane
153: characteristics in the vertical direction, to be able to display the (thin
154: emulsion) stereogram in white light\cite{Okada}. An optical element which does this
155: can also be created by application of holographic principles.
156: The experimental procedure is illustrated in fig. \ref{f2}.
157: We recorded a DeBitetto stereogram of a rectangular 'point' diffuser.
158: The 'point' was a Edmund Opal diffuser of size 1mm$\times$1.2mm. 120 element
159: holograms were recorded on 12.7$\times$10cm BB520 with a 50 mW 532 nm
160: $Nd:YVO_{4}$ laser. The BB520 film was translated by 0.9mm between each
161: exposure, equal to the width of the the slit covering the filmplate.
162: The distance between the 'point' diffuser and the holographic film
163: was approximately 30cm. The effective size of the diffuser was found to be
164: 10.8$\times$9cm, which results in a convergence angle of
165: approximately $\sim 8.5^{\circ}$ during reconstruction.
166: A reference beam illuminated the film from above with an angle $\sim 50^{\circ}$.
167: This reference beam could be considered uniform, since it was much larger than the
168: rectangular slit in the holoprinter. The BB520-film was processed by a procedure similar to that
169: suggested by Belendez et.al.\cite{Belendez}, resulting in Silver Halide
170: Sensitized Gelatin(SHSG). In this way we were able to produce a
171: diffuser with little unwanted scattering and $30\%$ diffraction
172: efficiency. When the real image of this multiple element
173: hologram was reconstructed, the diffuse points were reconstructed side by side,
174: thus creating a vertical line, see fig. \ref{f2}. We will name this kind of
175: hologram a $multiple-point$ $holographic$ $diffuser$, since it acts differently
176: from the holographic diffuser discussed in the previous section, and is supposed
177: to be used during production of a multiplex stereogram.
178: By illuminating the LCD(+ multiple-point holographic diffuser), the perspective
179: information on the LCD is projected onto a line of size $\sim 1.2mm$.
180: The line looks continous, since the width of each point(1mm) is slightly
181: larger than the distance translated between each exposure(0.9mm).
182: This gives a small overlap between reconstructed 'points'. Unfortunately, we
183: have not conducted any analysis to see how this overlap influences the image
184: quality. More research is needed to find out how this as well as the size of the slit
185: influences the final image.
186:
187:
188: \section{Measurements of the intensity distributions}
189: We have measured the intensity distributions from the two holographic diffusers
190: described above. The experimental setup used is shown in fig. \ref{f3}. A detector from Pasco
191: Scientific(model 8020) was covered with a small vertical slit of size $10\mu m$.
192: The detector can be moved(longitudinally) to the proper 'focal' plane of the diffuser
193: by a micrometer screw. A second micrometer screw was used to move the
194: detector(transversally) across the intensity distribution in steps of $10\mu m$.
195: The inaccuracy of the translating mechanism is about 5$\mu m$.
196: The same collimating lens was used during recording and reconstruction, and the
197: collimated reference beam was larger than the exposed
198: film(12.7$\times$10cm), and therefore considered uniform. The detector
199: was placed at the point were the real image was supposed to be located.
200: In both cases this distance was 30cm(as indicated in the descriptions above).
201: By translating the detector along the optical axis, we
202: measured the highest possible peak intensity 30cm from the diffusers
203: within an accuracy of $\pm 1cm$.
204:
205: Fig. \ref{f4} shows the relative intensity distribution of
206: the holographic diffuser. The width of the distribution is 0.5mm, which is the same as the
207: width of the strip diffuser used to record the holographic diffuser. Some light
208: is distributed outside the original boundaries, which may be caused by diffraction
209: from the edge, unwanted scattering, or misalignment during reconstruction.
210: The intensity variations within the passband is about 0.3. A major contributor
211: to the rapid intensity variations is speckle noise
212: introduced by the ground glass, but the emulsion will also contribute. Note also that the intensity distribution is
213: not completely symmetric. We believe this is due to a slight misaligment
214: of the slit during recording.
215: In fig. \ref{f5}, the intensity distribution at the 'focal' point of the
216: multiple-point diffuser is shown. Note that the
217: measured intensity distribution from the multiple-point diffuser is nearly
218: rectangular. Again, some light is distributed outside the
219: boundaries of the original 'point' diffuser(width 1.2mm).
220: It is seen that the intensity variations within the passband is about 0.5.
221: Furthermore, the variations are much more rapid than in fig. \ref{f4}.
222: This is believed to be due to the fact that we used different
223: emulsions(BB520 and BB640), although this has not been completely
224: clarified.
225:
226: \section{Setup for recording holographic stereograms}
227: To confirm the quality of the holographic diffusers described above, we
228: recorded a few holographic stereograms with the setup shown in fig. \ref{f6}.
229: The laser is a 20mW 632nm HeNe laser or a 50mW 532nm $Nd:YVO_{4}$-laser.
230: The reference and object beams are separated by a beamsplitter, and then
231: transported to different levels. The object beam is cleaned and expanded by a spatial filter, after which it
232: illuminates the LCD. The LCD is placed in close contact with the
233: diffuser, which directs the light into the slit of the printer.
234: The reference beam is also cleaned and expanded, and then directed
235: towards the slit in the printer. Each perspective picture on the LCD is
236: exposed onto a strip of the holographic film, one by one. In total, a
237: holographic stereogram is generated. We applied both the holographic and
238: multiple-point diffusers in front of the LCD, with good results.
239: We hope to publish a more detailed evaluation of the holographic stereograms
240: produced with this holoprinter in a separate paper.
241:
242:
243: In conclusion, we have described two different diffusers which may be of
244: interest in holographic stereography. We have performed some simple
245: measurements to get a better picture of the performance of these diffusers.
246: However, many questions are still not answered, and should be the subject of
247: future investigations. In particular it would be interesting to see
248: the influence of different numerical apertures, the reference beam angle, the
249: polarization of the incident beam, and the size of the slit. Also the influence
250: of overlap in the multiple-point diffuser should be studied in more detail.
251:
252:
253: This work was done while L. E. Helseth was at the University of Bergen.
254:
255: \begin{references}
256:
257: \bibitem{Okada} K. Okada, T.Honda, J. Tsujiuchi,
258: {\em Opt. Commun.} 36, 17-21 (1981).
259:
260: \bibitem{Klug} M. A. Klug, M. W. Halle, M. Lucente, W. J. Plesniak,
261: {\em SPIE} 1914, 15-23 (1993).
262:
263: \bibitem{DeBitetto} D. J. DeBitetto,
264: {\em Appl.Opt}, 8 1740-1741 (1969).
265:
266: \bibitem{Nakayama} Y. Nakayama, M. Kato,
267: "Diffuser with Pseudorandom Phase Sequence",
268: {\em J. Opt. Soc. Am. A.} 69 , 1367-1372 (1979).
269:
270: \bibitem{Yamaguchi} M. Yamaguchi, H. Endoh, T. Honda, N. Ohyama ,
271: {\em Opt. Lett.} 19, 135-137 (1994).
272:
273: \bibitem{Gao} Q. Gao, R. Kostuk,
274: {\em Appl. Opt.} 36, 4853-4861 (1997).
275:
276: \bibitem{Pawluczyk} R. Pawluczyk,
277: {\em SPIE} Vol.2042, 156-169 (1993).
278:
279: \bibitem{Belendez} A. Belendez, C. Neipp, I. Pascual,
280: {\em Opt.Commun.} 155 241-244 (1998).
281:
282: \end{references}
283:
284: \newpage
285: \begin{figure}
286: \caption{ Recording and reconstruction of the holographic diffuser described in
287: the text. An ordinary diffuser is covered by a small slit, and is
288: illuminated from behind. The diffused light interfere with the reference beam at the
289: filmplane. When reconstructed, the real image is a strip of the same size as
290: the original strip diffuser.
291: \label{f1}}
292: \end{figure}
293:
294: \begin{figure}
295: \caption{ Recording and reconstruction of the multippel-point holographic
296: diffuser described in the text. The light from a point diffuser is applied as
297: object beam. Between each exposure the film is transported a distance $s$.
298: \label{f2}}
299: \end{figure}
300:
301: \begin{figure}
302: \caption{ Setup for measuring the intensity distributions. A detector is
303: covered by a $10\mu m$ slit. During measurement, the detector is translated
304: in steps of $10\mu m$.
305: \label{f3}}
306: \end{figure}
307:
308: \begin{figure}
309: \caption{Intensity distribution from the holographic diffuser. The diffuser
310: is recorded on BB640, and reconstructed with a 20mW $632nm$ HeNe laser.
311: \label{f4}}
312: \end{figure}
313:
314: \begin{figure}
315: \caption{Intensity distribution from the multippel-point holographic diffuser.
316: It was recorded on BB520, and reconstructed with a 50mW $Nd:YVO_{4}$ laser.
317: \label{f5}}
318: \end{figure}
319:
320: \begin{figure}
321: \caption{Setup for recording a holographic stereogram with a holographic
322: diffuser or multiple-point holographic diffuser. The dashed line represents
323: the beampath on the second level. The reference beam illuminate the filmplate
324: from above.
325: \label{f6}}
326: \end{figure}
327:
328: \newpage
329: \centerline{\includegraphics[width=14cm]{figart61.ps}}
330: \vspace{2cm}
331: \centerline{Figure~\ref{f1}}
332:
333:
334: \newpage
335: \centerline{\includegraphics[width=14cm]{figart62.ps}}
336: \vspace{2cm}
337: \centerline{Figure~\ref{f2}}
338:
339:
340: \newpage
341: \centerline{\includegraphics[width=14cm]{figart63.ps}}
342: \vspace{2cm}
343: \centerline{Figure~\ref{f3}}
344:
345:
346: \newpage
347: \centerline{\includegraphics[width=14cm]{figart64.eps}}
348: \vspace{2cm}
349: \centerline{Figure~\ref{f4}}
350:
351:
352: \newpage
353: \centerline{\includegraphics[width=14cm]{figart65.eps}}
354: \vspace{2cm}
355: \centerline{Figure~\ref{f5}}
356:
357:
358: \newpage
359: \centerline{\includegraphics[width=14cm]{figart66.ps}}
360: \vspace{2cm}
361: \centerline{Figure~\ref{f6}}
362:
363:
364:
365:
366:
367: \end{document}
368: