2353427 LIGHT DETECTOR FOR A FILM SCANNER This invention relates to the
scanning of cinematographic film to produce electrical signals corresponding to the images stored on the film. In particular, the invention relates to a light detector for use in a film scanner or telecine.
Telecine or film scanning equipment used to produce such signals from cinematographic film have been known for many years, and are described for example in "TV and video Engineers Reference Book" Chapter 39 Butterworth and Heinemann ISBN 0- 7506-1021-2. The invention relates particularly to the type of telecine or film scanner that uses a scanning light source. Examples of commercially available telecines of this type are made by Cintel International Limited and sold under the names Ursa or CReality telecines.
The key parts of a raster scan telecine are described in simple terms with reference to Figure 1. A cathode ray tube (1) produces a raster scan that is imaged onto the film (2) by an imaging lens (3). The light passing through the film is modulated by the colour and density of the film at each location or pixel scanned, this light being analysed into it's red, green and blue components using dichroic mirrors (4) and then converted into electrical signals by avalanche photo diodes (5). The three electrical colour signals are then passed through electronic processing circuits (6), converted into a television signal format and provided as output signals which are typically then recorded on video tape equipment.
The basic raster scan telecine described is well known to those skilled in the art.
It is realised that it would be beneficial to collect more of the light that is scattered by scratches on the surface of, or other deformities in, the film. Such scratches cause scattering of the light, which may then be lost from the optical system and cause a reduction in the signal received by the photo sensors, if this light could be effectively collected then the visibility of the scratches would be much reduced. WO-A-83/02869 and US-A-4481414 both describe various methods of improving the collection of such scattered light by imaging and or by reflective means. The purpose of the systems disclosed in these documents is to try and reduce the visibility of scratches on film scanned in a raster scan telecine by collecting as much light scattered by scratches as possible.
A second approach is disclosed in GB-A-1409153 which discloses a system having an additional detector for detecting light scattered by scratches. The output of this additional detector is used to determine whether the signal from the main image sensor should be substituted with another signal. The substitution is for the purpose of reducing the visibility of the signal in the resultant video signal.
A third approach is disclosed in GB-A-2323495A which also discloses improved collection of the scattered light by imaging and reflective means and the use of additional separate sensors displaced from the main sensor to collect the scattered light and add this in suitable proportion to the main signal. The additional sensors are stated to provide an indication of the level of scattering, and so do not necessarily need to collect all scattered light. The signal from the additional sensors is used to compensate the signal from the main detector.
We have appreciated problems with the known systems for reducing visibility of scratches in telecines. In relation to the first approach noted above of attempting to capture the scattered light, it has been appreciated that there is a limit to the amount of this scattered light that can be collected in a practical optical system. This limit is effected by the LaGrange optical invariant that can be referenced at page 2-8 of the "Handbook of optics" published by the McGraw-Hill Book Company ISBN 0- 07-047710-8, a principal of optics known to the skilled person. In the C-Reality telecine described above the optical invariant states that the product of the maximum scanned film dimension and the numerical aperture of the rays passing through the film will in (a theoretically perfect optical system) be equal to the product of the maximum active sensor dimension and the numerical aperture of the rays arriving at the sensor. The dimensions of the film and sensor together with the numerical aperture of the imaging lens are chosen such that substantially all of the imaged light rays fall on the sensor within an acceptable numeric aperture. Typically, the optical collection system is arranged so that light falling on the sensor is an image of an imaging lens pupil and that image substantially fills the photosensor active dimension.
However when the light passing through the film is scattered by scratches, it acquires a much larger effective numerical aperture and in consequence the image at the sensor will be much larger and will fall outside the active dimension of the sensor.
one solution to this difficulty is to change the magnification of the optical collection system to fit the scattered light onto the active sensor dimension; however, due to the aforementioned optical invariant this will result in a much increased numerical aperture that falls beyond the range of numerical aperture which can be effectively viewed by the sensor. It should be noted that this limitation is fixed by the film dimension, the sensor dimension, and the numerical aperture at the film (for scattered light this is at the maximum value of 1-0), and cannot be improved beyond this limit by any of the imaging or reflective means described in the aforementioned references. It is also known that the effective numerical aperture of the photo sensor can, in some instances, be increased by the use of a high refractive index substance fitted between the active surface of the photo sensor and the optical system.
The second and third approaches are reliant on the assumption that light is scattered uniformly, which is not always the case, and that the separate detector produces a signal representative of the scattered light.
An alternative known solution is to use a larger sensor. However, we have appreciated, in the example of the C-Reality telecine that for reasons of best performance the sensor should be an avalanche photo diode of 10 mm diameter.
In view of the foregoing difficulties, we have appreciated that a better solution is needed.
Accordingly, in a broad aspect there is provided apparatus for producing an electrical signal representative of an image on film, comprising a light detector for detecting light and producing an electrical signal; means for directing a spot of light to illuminate film; and means for directing light transmitted by film to the light detector; wherein the light detector comprises at least first and second separate light sensitive regions.
The invention provides the advantage that light modulated by film can be received in two or more separate light sensitive regions allowing a larger than usual detector to be constructed with the signals produced by those regions to be manipulated differently. This has the advantage of not worsening the signal to noise ratio and allowing the apparatus to be configured to detect and reduce the effect of light scatter. Preferably, in a first preferred aspect, the second light sensitive region is arranged around the first light sensitive region. The advantage of arranging the second region around the first is that the two regions are in similar environments and so any external influence on the detector, such as drift caused by heat, will be the same on both regions. The second region can be used to detect light arriving at the detector which has been scattered by film and is better representative of the scattered light than the known devices described above. The second light sensitive region preferably surrounds the first light sensitive region, thereby ensuring that light from the film arriving at the area of the detector fall on either the first or second region.
Preferably the first light sensitive region is circular and the second light sensitive region is annular. This has the advantage in that an optical system can be arranged to produce a circular patch of light on the detector. In an embodiment, the means for directing a spot of light comprises an imaging lens, and the means for directing light transmitted by film comprises a collecting lens arranged to produce an image of the imaging lens on the first light sensitive region. This arrangement ensures that, in the absence of scattering, light received at the detector is imaged onto the first region from which an output signal is produced. In the event that the film scatters light, due to scratches on other defects, this light is collected on the second light sensitive region around the first. The signal from the second region can then be used to correct the signal f rom the f irst to reduce the visibility of defects.
The first and second light sensitive regions described above are portions of a single sensor, but could each be a separate sensor. Preferably, however, the first and second light sensitive regions are portions of a single sensor, each having a separate preamplifier. There are preferably four light sensitive regions. The diameter of the detector is chosen to be 20 mm allowing a four times greater area than the usual 10 mm detector producing a four times greater signal, but without a four fold increase in noise.
In a second aspect the detector is preferably circular comprising four light sensitive quadrants. This aspect allows a detector to be constructed which is larger than usual without worsening the signal to noise ratio to allow for improved reduction of scratch visibility.
An embodiment of the invention will now be described with reference to the figures, in which:
Figure 1: shows a schematic view of a film scanner arrangement; Figure 2: shows the optics and detector of an embodiment of the invention which may be used in the film scanner of Figure 1; Figure 3: shows a schematic view of a signal processor and detector embodying the invention; Figure 4: is a schematic view of the limitations of an optical system because of the physical geometry described by Lagrange invariant; and Figure 5: is a schematic view of the limitations of a reflective collection system.
The embodiment of the invention is a set of optics and a light detector Particularly for use in a film scanner or telecine. A film scanner embodying the invention described is a CRT telecine. However, the invention can be applied to other film scanners in which a spot of light is used to scan film.
The embodiment described comprises a detector with first and second light sensitive regions. The invention is applicable to a detector with a plurality of regions and may have more than two. These regions could themselves be separate sensors, but are preferably regions of a single sensor. The embodiment provides the advantage that a larger than usual sensor can be constructed to allow reduction of scratch visibility without worsening the signal to noise ratio.
In a variation of the embodiment embodying a second aspect, the detector has four light sensitive regions which are quadrants of a circular detector. This variation allows a large detector to be constructed to collect scattered light without worsening the signal to noise ratio. Unlike the first variation, however, the signals from the different regions do not differentiate between scattered and unscattered light.
The film scanner embodiment comprises the basic film scanner previously described as shown in Figure 1. A cathode ray tube (1) produces a raster scan that is imaged onto the film (2) by an imaging lens (3). The light passing through the film is modulated by the colour and density of the film at each location or pixel scanned, this light being analysed into its red, green and blue components using dichroic mirrors (4) and then converted into electrical signals by detectors in the form of avalanche photo diodes (5). The three electrical colour signals are then passed through electronic processing circuits (6), converted into a television signal format and provided as output signals which are typically then recorded on video tape equipment.
The embodiment as shown in Figure 2 uses photo detectors to detect the scanned light beam passing through the film, in which the photo detectors comprise a coaxial structure with a central sensitive area or region (14) used to detect primarily the imaged light passing through the film, and an outer annular sensitive region (15) used primarily to detect the light scattered by scratches as indicated by the dashed lines in Figure 2. The signa 1 from the outer region is then summed with the signal from the inner region in proportions suitable to render the scratches least visible. Using a single photo detector that is partitioned in this way rather than two or more sensors results in less wasted space between the sensors hence better efficiency of light collection. A further benefit is that the two regions will be much better matched than would two separate sensors with regard to operating temperature, gain dark current, etc. The two regions could, however, be separate sensors arranged close together.
The signal from the inner region will be substantially imaged light (I.a) with a small amount of the diffracted light (I.b) From the outer region the signal will be a small amount of image light O.c and predominantly diffracted light O.d.
Since the desired signal is the total of all imaged and diffracted light, it would seem that a simple sum of the two signals would suffice. However, the diffracted light will be spread in all directions and a portion will not be received by either sensor, to make up for this lacking portion it is desirable to increase the gain of the signal from the outer region before adding. Since the distribution of the scattered light is dependent on the nature of the film scratch, it is also preferable that the amount of increase in gain be adjustable by the equipment operator, as shown in the signal processor shown in Figure 3.
The sensor shown in Figure 3 is a partitioned detector as previously described and comprises an APD with a barrier gap between the silicon of the two separate regions. A separate connection to each region provides an output to two separate channels of the signal processor. In an embodiment having a detector with four separate active regions, four separate channels are provided. The channel for the first region 14 comprises an amplifier 22 resistor 23 and fixed gain resistor 24. The channel for the second region 15 comprises an amplifier 20, a resistor 21 and a variable gain resistor 25. The gain for the second region is thus variable with respect to the first, and the signals summed at out output channel comprising an amplifier 26 and resistor 27.
In a variation of the embodiment of the invention the active area of the photo detectors are divided into several parts such as four quadrants. This arrangement makes it possible to use a photo sensor of larger total active dimension but where each part has a relatively small active area. An advantage of having a small active area for each part is that the electrical capacitance of each part is smaller and therefore the following amplifier can provide higher frequency response and lower noise.
The larger overall area of the photo detector will permit increased output signal levels with further improvement in signal to noise ratio, and will also collect more of the scattered light so reducing the visibility of scratches.
By way of example the photo detector could be of 20 mm diameter which corresponds to four times the area of the presently used photo detector and divided into four sectors, for example quadrants. Each sector would have the same capacitance as at present and would operate into its own preamplifier so that the frequency response, and the multiplication of amplifier noise by the ratio of feedback resistance to input capacitance, will be unchanged. The photo detector gains could be adjusted to give the same output signal as before, so that when the four amplifier signals are summed there will be four times the signal (m+ m+m+m=4m), and the amplifier noise will only increase by two times V(n 2+ n 2+n 2+n 2) =2n, (noise is root sum of squares of individual noise contributions). The technology required to produce such segmented photo sensors and avalanche photo diodes is known (for example the Bi-cell and Quadrant photo diodes available from Advanced Photonix Inc. of Camarillo, California, USA).
As discussed, an advantage of the invention is to allow construction of a larger than usual sensor without worsening the signal to noise ratio. The advantage of a larger detector can be seen by considering Figure 4. The lagrange invariant states that at any plane in the optical path, the product of, the height of a first ray from the optical axis, the angle of a second ray which passes through the optical axis in t hat plane, and the refractive index of the material, will be a fixed a value.
In the above case U2 is fixed by the Numerical Aperture (N.A.) of the imaging lens, h2 is fixed by the film dimension, and U3 obviously cannot exceed JJ/2 rads, but in practice the sensor response will be limited at an even smaller angle. The invariant requires h3 X U3 = h2 X U2. So the minimum image dimension at the sensor will be h3 = h2 X U2 or typically 40mm (diameter) x 0.157 rads 11/2 1.57 (giving -4mm diameter but in practice is more like 8mm due to limited angle of acceptance from the sensor).
The actual size of the image at the sensor is chosen (by adjusting the lens magnifications) to be closer to the sensor dimension of 1Omm so that the edges of the film do not present too steep an angle on the sensor.
The Lagrange Invariant also applies to reflective systems. Consider a reflective cone as in Figure 5, used to collect light from a source where the exit diameter is smaller than the input diameter. Rays of large angle will be reflected back to the input and will not arrive at the exit port. The same relationship as for lenses applies where the product of the input aperture size and the ray angle equals that of the exit aperture and the exit ray angle, with the exit ray angle limited to 11/2, and angles which would exceed that being redirected back to the input.
It should be realised that whereas the above description is applied to the example of a C-Reality telecine using avalanche photo diodes the invention can be applied equally well to other telecines or film scanners and to other forms of photo detector.