Heads Up and Near Eye Display!

Transcription

1 Heads Up and Near Eye Display! What is a virtual image? At its most basic, a virtual image is an image that is projected into space. Typical devices that produce virtual images include corrective eye ware, binoculars, magnifying glasses, microscopes, telescopes, rifle scopes, near-eye displays, and heads-up displays. Each of these devices forms a near-field image that is viewed with the eye and adapted to the distance based upon the design of the instrument. In creating a virtual image, luminance and color measurement are key factors. Both of these are impacted by the fixed physical entrance size of the pupil. There are four distinct physical categories for image measurement. They are: Virtual Reality which involves no ambient illumination and is near-eye display Augmented Reality consisting of partial ambient illumination and near-eye display Digital Eyeglass requiring full ambient illumination and near-eye display Automotive/Avionics requiring full ambient illumination and heads-up display Using Standard Luminance Standard luminance products can generally be used for measuring HUD displays, but near-eye displays need special consideration due to the compacted optics. One of those considerations is the effect of ambient illumination, which complicates both design and measurement. Illuminant A is the basic standard used by all manufacturers for calibration source characterization. The benefits of Illuminant A are that the physics of the source is well understood, the source is very stable when properly powered, inter-laboratory comparisons show a high degree of correlation, and spectra is continuous with no inflection points. However, the source has very low blue content, is not representative of any modern display, and has very low color temperature relative to contemporary lighting/viewing standards.

2 2.5 NIST Amplitude Uncertainty for illum.a compared with High Brightness White LED 2 2 k uncertaint k uncertainty high brightness white LED Wavelength This creates uncertainty around using Illuminant A as a calibration source. Most devices are referenced to Illuminant A, but measurement uncertainty on typical non-illuminant A sources is relatively high and inter-instrument agreement can suffer. Therefore, most electronic displays are not based upon Illuminant A. HUD and near-eye displays have varied temporal characteristics due to image refresh. Sync timing is critical for short exposure times at high luminance. Exit pupil uniformity is another issue that affects luminance accuracy. These uncertainties compound to degrade inter-instrument agreement. Augmented Reality and the Transparency Issue Virtual Reality (VR) is a digital environment that shuts out the real world. Augmented Reality (AR) is digital content on top of the real world. The dynamic range of ambient illumination dictates the dynamic range of the display element. The difference between a traditional heads-up display and an augmented reality display is that the HUD is generally used to provide collateral information to the scene rather than to augment the scene with another scene.

3 This eye box example above is a simple magnifier where the real object is placed in front of the focal plane of the lens. Ray trace shows that the virtual image is formed behind the real object. A second ray trace must be performed to see how the observer will see the image. How does measuring a virtual image differ from a real image? The virtual image on the above eyebox example shows the extent is limited by the exit pupil. This means that proper measurement of luminance requires three non-negotiable properties: 1. The extent of the illumination from the exit pupil must exceed the entrance pupil of the measurement instrument. 2. The field of view of the measurement instrument must be less than the extent of the eye box or exit pupil. 3. The illumination must be uniform at the entrance pupil of the measurement instrument. This is a key, degrading element in the comparison of data from different instruments. Viewing the above virtual image demonstrates a second ray trace performed to show the viewing conditions. The virtual image is now defined as a real object in the ray trace. The optimal eye position is arbitrarily defined to be the plane represented by the paraxial chief ray (shown in red) and

4 the paraxial marginal ray (shown in green). Observation from this point allows for viewing the complete field. The eye box extent plane represents the limiting distance that an observer can view the entire image field. So where is the problem? In theory, there shouldn t be a problem. Luminance is conserved and thus it is the exactly the same at any position in the optical system. However, the problem arises from the physical configuration of the measurement device. The configuration of the pupil and the eye box requires a strict understanding of the optical characteristics of the system. The illumination in the eyebox must be completely uniform, a non-negotiable requirement. If the physical extent of the virtual image is smaller in near-eye situations than in the heads-up display, the inter-instrument agreement becomes a problem. There are several other challenges in measurement in the HUD/near-eye marketplace. Many customers do not understand the two fundamental requirements: the extent of the illumination from the exit pupil must exceed the entrance pupil of the measurement instrument, and the field of view of the measurement instrument must be less than the extent of the eye box or exit pupil. Luminance is conserved everywhere in an optical system and is only modified by transmission of the optics. This problem was highlighted in the recent eye ware discussions Eyewear Display Measurement Method: Entrance Pupil Size Dependence in Measurement Equipment Kosei Oshima, et al. Incorrect use or design of luminance devices can lead to changes in fixed aperture size and large changes in luminance. The basic data between two instrument types have fairly large differences even at small aperture size. Display non-uniformity exacerbates these differences. Fixed Aperture Solution In the real world, a fixed aperture attachment allows for measurement in the eye box. The telecentric design eliminates focus, but it is absolutely important to understand the field of view. This fixed aperture system has a Field of Vision of 18 degrees. In the real world, there is no standard configuration for eye ware. Physical jigs must be constructed to manage the ear hooks and simulate

5 ! a nose bridge. Limited space makes measurement difficult, but it is possible using a spectroradiometer if the aforementioned three non-negotiable conditions are met. A standard spectroradiometer in position to measure the near eye device can provide comparisons of luminance measurements between the fixed aperture and the standard device. This can be a factor of as much as three times lower. In a test example, the entrance pupil of a standard instrument was not filled, so positioning the spectroradiometer farther away yielded much better correlation. However, it was impossible to get both configurations to agree until the Field of Vision of both systems were equal. Using a fiber optic probe to fit inside head box can be created if the fiber adapter attaches to the standard lens. The fiber probe is then mounted in the fixture. This makes the fixed physical entrance pupil easy to locate. The average pupil dimension of the human eye at moderate luminance is 3-5mm. This is extremely useful in psycho-physical studies where retinal illuminance is easily calculated. One technique employs a concept known as telecentricity. There is correlation of fixed aperture measurements with a standard luminance meter. This is not simple because of the non-uniformity of the displays coupled with the optical systems. The exit pupil of the standard device must be over-filled for an accurate spot measurement using a standard photometer/spectroradiometer. When using the fixed aperture system, it is therefore very important to under-fill the area of measurement. The benefit of a fixed aperture size solution is that it minimizes the problem of measurement geometry. By definition, this solution is very tolerant of imaging position. The aperture can be defined to be the same as a human pupil dimension. Insertion loss due to stop lens is less than 3%.

6 120 Relative change in luminance with distance using fixed aperture design Releative luminance Series Distance in mm from source Photo Research s Fixed Aperture Attachment The Photo Research Fixed Aperture Attachment provides very small changes within fairly large motion of the measurement device. Rapid change in measured luminance occurs when the measured source no longer fills the entrance pupil of the lens. There is virtually no change in calibration between aperture sizes of 2-6mm. Referencing the above diagram, a telecentric stop located at the back focal point of L1 defines changes in image size with the position of the object. The second lens focuses the stop to a fixed

7 Heads Up and Near Eye Display point. L1 is positioned at the front focal point of L2, thus making the exit pupil of L1 telecentric. The relationship between the positions of L1 and L2 provide for a high degree of pupil mixing. The field of view of this system is adjustable by putting a physical stop at the front surface of L1. This design can be embodied as an attachment to the PRISM-75 lens through use of the MS-75 lens as L2 in the optical. 2-Dimensional Measurement What about 2-dimensional measurement? 2-dimensional measurement is important for defect detection (i.e. MURA) and for other criteria such as MTF, image motion smear, fixed pattern noise, distortion, and more. Fixed pupil in the eye box is VERY challenging but it may not be necessary for these spatial measurements. The basic entrance pupil requirements must be met by being completely and uniformly filled. This is often impossible due to the optical limitations imposed by the eye box geometry and the function of display uniformity. If absolute radiometry/colorimetry is not important, then this is less problematic. In general, if the full field can be captured, the luminance will measure low due to entrance pupil issues, but this is correctable by using a standard radiometer with the correct geometry. Measurement of MTF is also possible, but the MTF image capture system must be characterized and used to correct for the imager characteristics. This can lead to significant noise at higher frequencies. Measurements of image smear require very precise control of exposure time. The nature of the electronic shutter of the sensor may complicate the understanding of the data. So correlation with fixed aperture systems will be very difficult. Lack of knowledge about the eye-box parameters inhibits the development of a generic solution for all situations. The imager characteristics have a large impact on the resulting data, and the imager must be completely characterized before use. Conclusion For heads-up displays (used in Avionics/Automotive), use of a standard spectroradiometer is advised with the provision that care is taken to ensure that the entrance pupil of the instrument is properly filled. For near-eye displays, the fixed aperture method should be used. A standard spectroradiometer can also be used in a standoff position to ensure the entrance pupil is filled. Keep in mind correlation between methods will be compromised by display non-uniformity. Repeatability measurements should be made, with no fewer than seven, to ensure that there is not an issue with temporal syncing. Reliability measurements should also be made to confirm that the measurement process is working. For more information regarding colorimeter and spectroradiometer solutions offered by PHOTO RESEARCH, visit us at or call us at