MOTION PICTURE SUBTITLE METHOD AND APPARATUS
BACKGROUND OF THE INVENTION Field of the Invention
This invention relates to the formation of supplemental visual displays, in particular subtitles, on a motion picture.
Description of the Related Art
There are three conventional techniques for superimposing subtitles on motion pictures. In the first, a non- subtitled motion picture print film is coated with a protective layer of wax in all areas of the motion picture image except for the areas which define the desired subtitle characters. The print emulsion is then removed from the unwaxed subtitle character areas by an acid treatment, followed by removal of the wax coating from the rest of the print. .When projected, the acid treated areas appear as subtitle characters.
While subtitles produced by this method are generally acceptable in quality, once the subtitles have been estab- lished they cannot be removed. Thus, separate prints are required for each different subtitle language, and for a print without subtitles.
An alternate to the wax/acid technique involves forming the subtitle characters by a laser beam which remove the emulsion from the character areas. This process gener-
ally produces a greater character resolution, but it still requires separate prints for different subtitle languages. The third technique is to include the subtitles in the original film exposure so that they are formed as integral parts of the film images. This is generally more time consuming, and again requires an entirely new print for each different subtitle language.
SUMMARY OF THE INVENTION The present invention allows the subtitles for a single motion picture print to be changed from one projection of the print to the next, or to be left off entirely, thereby greatly reducing the cost and inventory that would otherwise be required for multiple prints. The invention provides this ability without making any physical change to the film print itself, and thus eliminates the danger of damaging the film by the subtitling process.
The invention involves using a supplemental projector in the motion picture theater to project subtitles, or any other desired supplemental visual display, onto a motion picture projection from a primary projector that runs the film print. The projection of subtitles is synchronized with the motion picture projection so that the correct subtitles (or other supplemental visual displays) are superim- posed on the motion picture at the right times..
The subtitles can be made available for projection in several ways, such as storing them on a compact disk or other digital storage medium, or providing them electronically for either storage or real time display. A prefer- red technique for synchronizing the projection of subtitles from a storage medium with the projection of the film from a primary projector is to provide the film with a time code that identifies the particular portion of the film cur-
rently being projected, and to tag the subtitles in the storage medium with their corresponding time code units. The time code is read from the film as it is projected, and is used to access the corresponding subtitles for superim- posed projection onto the film display.
The storage medium data frames can have equal lengths, regardless of the number of character spaces required by any particular subtitle, or the frame lengths can be tailored to the number of character spaces required by their respective subtitles. Equal length frames, when the subtitle frame is an integral number of time code frames in length, allow for a very rapid accessing of any desired frame in the storage medium, since the locations of all frames are known. For example, the subtitle frame could be 6 time code frames long, allowing for second time resolution for the subtitles, but reducing the data space required for subtitles by a factor of 6 compared to having the full subtitle repeated for every time code frame. This approach is particularly useful when the data storage fram- es also include respective portions of the motion picture audio track.
Unequal frame lengths allow for a more efficient utilization of the medium's storage capacity, but also takes longer to access; desired frames can be reached by fast forwarding through the medium, rather than making a single jump as with equal length frames. However, the accessing process can be significantly speeded up by first scanning the storage medium to establish a file, either in the medium itself or in a computer which controls access to the medium, correlating time code units with the locations within the storage medium of the corresponding data frames. The file is then used to make a very rapid jump to the proper location within the storage medium that corresponds
to the current time code input.
The supplemental projector for the subtitles or other supplemental visual display can be implemented in numerous way, including liquid crystal light valves (LCLVs) and video, slide and film projectors. Video projectors are satisfactory for smaller projection screens, while LCLVs are preferred for larger screens.
These and other features and advantages of the invention will be apparent to those skilled in the art from the following detailed description, taken together with the accompanying drawings .
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a block diagram illustrating the principal components of the invention;
FIG. 2 is a block diagram illustrating in generic form the manner in which the supplemental display can be generated;
FIGs. 3, 4 and 5 are block diagrams illustrating dif- ferent ways to produce the supplemental display;
FIG. 6 is a block diagram illustrating the provision of supplemental display data to an LCLV;
FIG. 7 is a block diagram illustrating the integration of the subtitling system with a playback system for the motion picture audio track;
FIG. 8 is an enlarged plan view illustrating a motion picture film which bears a synchronizing time code;
FIG. 9 is a combined plan view and block diagram illustrating how the time code is read from the film in syn- chronism with the projection of the motion picture image;
FIG. 10 is an illustrative representation of a series of digital subtitle data storage frames with different frame lengths;
FIG. 11 is a flow diagram of a process for accessing a storage medium with data frames such as those illustrated in FIG. 10, including the formation of a file correlating time codes with frame locations; and FIGs. 12 and 13 are illustrative representations of equal length data storage frames respectively for a supplemental visual display by itself, and a supplemental visual display combined with a motion picture audio track.
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 illustrates a primary application of the invention: the projection of subtitles onto a motion picture display. A primary projector 2 projects a motion picture onto a display surface 4 such as a conventional movie screen. Subtitles are projected from a supplemental projector 6 onto a desired area 8 of the movie screen designated for subtitles. The subtitle information is provided to the supplemental projector from a controller 10 that stores the subtitles, preferably in a digital format, in a storage medium 12. In response to a SYNC signal over line 14 from the projector, the controller 10 provides whatever subtitle is currently desired to the supplemental projector 6 for projection onto the screen 4. The SYNC signal maintains synchronism between the primary and supplemental pro- jectors so that the appropriate subtitles are superimposed on the motion picture display at the correct times.
Since the subtitles are generated separately from the film print, a non-subtitle print can be projected from the primary projector 2 and any desired subtitles added via the supplemental projector 6. The subtitles can be changed for the same print simply by replacing the storage medium 12 with a new storage medium having a different set of subtitles. Thus, any number of different subtitle sets can be
used with the same film print, or no subtitles at all, without having to modify the print or provide a different print for each different subtitle set.
FIG. 2 is a block diagram of a more generic presenta- tion of the invention. Supplemental visual displays other than subtitles, such as special effects or advertising, can also be superimposed onto the motion picture display. Controller 10 can be located either at the site of the primary and supplemental projectors, or at a more remote location. The storage medium 12 for the supplemental visual display can be a conventional mechanism such as a compact disc (CD) or a digital tape. With this type of storage medium, the controller 10 is implemented as a computer which accesses the medium in response to a SYNC signal on the input SYNC line 14, and provides the synchronized supplemental display information to the supplemental projector 6 over an output display information line 16. The input SYNC line 14 and output information line 16 can either be hard wired as shown, or implemented in other ways such as on electronic delivery system or RF or optical transmission channels.
The SYNC signal on the input SYNC line 16 can have different formats, depending upon how the system is set up. Many film prints are provided with a time code distributed along the length of the film, with each time code unit con- sisting of a digital data string that uniquely identifies its location along the film. Presently available time codes include the DTS (Digital Theater Systems, Inc.) format of 24-bit digital words, each word beginning with a synchronizing series of bits, and the SMPTE (Society of Motion Picture and Television Engineers) standard. If the film being projected has a time code, the time code is read at the primary projection site and provided as the input SYNC signal to the controller 10. The storage medium 12
would include a sector corresponding to each time code location for which a subtitle is desired, with each sector frame including the corresponding time code and subtitle (or other visual display) information. When a particular time code unit is input to the controller as a SYNC signal, the corresponding memory storage frame is accessed and its subtitle information transmitted to the supplemental projector 6 for superimposed display on the motion picture. The storage medium 12 can include frames for all time code units, with blank frames for the time code units which do not require a subtitle, or only frames for time code units with corresponding subtitles.
If the film does not have a time code, an equivalent film location signal can be provided by using a film frame counter (tachometer) to count the film frames as the film is projected. This is not as desirable as a time code, however, because of the common practice of editing film prints after distribution. If portions of a film print have been edited out and removed, for example to eliminate damaged portions of the print or to reduce the length of the movie, the remaining time code units will still maintain synchronization with the subtitle information. With a tachometer, on the other hand, removing a portion of the film after the subtitle storage medium has been set up for the full film length will result in a loss of synchronism when the first edit is encountered, and increasing synchronism losses for each subsequent edit.
The SYNC input could also be provided as serial data from another computer, as in a show controller configura- tion. In this case the SYNC signal can be generated at a remote location, without obtaining it directly from the film being projected.
The controller 10 will normally include a SYNC input
card 18 which converts the SYNC input signal to a computer recognizable format that allows it to access the corresponding frame in the storage medium 12, and an OUTPUT card 20 which converts the subtitle or other supplemental visual display information read from the storage medium to a format suitable for driving the supplemental projector 6. For example, the supplemental projector could be implemented as a video projector, in which case the supplemental display information on line 16 would be formatted as a video sig- nal . For a film or slide projector, the supplemental display information would be provided in a film frame format. For an LCLV, serial text would be provided over line 16 to activate the desire pixels of the light valve's pixel array. Under screen text displays are another possible sup- plemental projection technique.
It is also possible for the storage medium 12 to be provided at a central location and accessed electronically via the Internet. In that case the SYNC input signal could be provided to the controller electronically from the local film projection site, and the output supplemental display information transmitted back electronically to the supplemental projector.
The use of a video projector for the supplemental projector is generally acceptable for smaller size movie screens. However, video projectors are designed for images with true motion. When used for still images such as subtitles, the image can be distorted. While the degree of distortion is not great for smaller screens, it can become distracting for larger screens. One of their advantages, however, is that they can accept electronic image delivery. Slide projectors produce a brighter still image with better contrast than video projectors. However, they require a film and do not accept electronic deliveries.
LCLVs are specifically designed for still images; their grids are controlled directly to produce a still image. LCLVs avoid an analog stage used by video projectors, which produces distortions in the projected image. They produce a better still image and are less expensive than video projectors, and are also accessible electronically.
The storage medium 12 can either be loaded with the supplemental visual display and corresponding SYNC data prior to delivery to the motion picture theater, or a blank storage medium can be provided and loaded electronically on-site prior to use. The information can either be left on the storage medium after the film run has been completed, or it could be erased and the storage medium reloaded with subtitles for the next film to be played. FIG. 3 is a block diagram of a system that incorporates a motion picture digital sound system provided by DTS. This system is disclosed in U.S. Patent Nos. 5,386,- 255 and 5,751,398, the contents of which are incorporated herein. The DTS system provides the film print with an unique time code that is used to synchronize a playback of the motion picture audio track, which is stored on a CD or other digital medium separate from the film itself. The digital audio track for FIG. 3 is encoded on a CD 12' and organized into information frames, each of which includes a portion of the audio track and a time code unit which identifies the location on the film for which the audio frame is intended. The CD 12' is played in a CD player 10' which receives the time code (TC) signal from the film as it is being projected, and outputs a digital audio signal for playback in the theater in synchronism with the film projection.
The system illustrated in FIG. 3 assumes that the data frames in the CD which include the audio data also include
digital data comprising any subtitle or other supplemental visual display which is to be projected at the same time. If the disc includes only audio data, the subtitle data could be placed together with the corresponding time code information on a separate disc and played back in synchronism with the audio disc.
The subtitle data is output from the CD player via a communications port 22, preferably implemented as a serial card already in most computers that provides for communi- cation with another computer or other digital device capable of receiving serial digital data.
The serial subtitle data is transmitted over line 24 to the input communications port 26 of another computer 28. Here the serial data is converted to a video signal by a conventional video card 30, and transmitted over a video line 32 to a video projector 34 which projects the subtitles onto the screen. The video projector 34 includes an internal LCLV, but the process of converting the input video signal to an LCLV control signal introduces distor- tions into the subtitles.
FIG. 4 illustrates a variation of the FIG. 3 system which eliminates the need for an intermediate computer. In this case the CD player 10' includes an output video card 36, such- as an SVGA or DGA card for IBM-compatible comput- ers, which converts the subtitle file read from the CD 12' directly to a video signal on output video line 38. This allows the CD player 10' to provide a direct video input to the video projector 34. Other types of projectors could also be used, with an appropriate adjustment of the output card 36.
FIG. 5 illustrates a system in which the subtitle or other supplemental visual display information is stored on a CD 12" that does not include audio information and is run
in a computer controller 10". A time code or other synchronizing signal is applied to an input card 40 that converts it to a format readable by the computer. In the illustrated system an LCLV 42 is used for the supplemental projector, and an LCLV control card 44 in the computer converts the subtitle data read from the CD to an LCLV control format. The LCLV image is projected through an anamorphic lens 46 that reduces the image height to the desired subtitle area of the screen, and thus increases the subtitle brightness.
FIG. 6 illustrates the manner in which the LCLV 42 is addressed. The matrix of pixels used to form the subtitle characters is stored in an RGB (red/green/blue) data RAM 48. Horizontal and vertical SYNC signals are provided from input lines 50 and 52, respectively, along with a clock signal on clock line 154, to both the LCLV and the RGB data RAM. As each pixel in the LCLV is addressed, the data RAM 48 provides the corresponding illumination signal to control the brightness for that pixel. In this manner the LCLV pixels which collectively form the characters of the desired subtitle emit a light output, while the other pixels remain dark. When projected onto the screen, the emitting pixels collectively produce the characters of the desired subtitle. A system that can be used to produce both an audio output in the theater and subtitles or other supplemental visual display superimposed on the motion picture, in response to a digital time code on the film print, is illustrated in FIG. 7. The motion picture film 56 that includes the time code travels from a playout reel 58, past a time code reading head 60 and the primary projector 62, to a takeup reel 64. The projector 62 is positioned at a known distance from the time code reader 60, so that the time
required for the film to travel from the time code reader to the projector with a normal film advancement speed is precisely known.
The time code information from reading head 60 is transmitted to a data transfer controller 66, such as a conventional IBM® compatible personal computer, that receives the time code data via an input port. The computer 66 manages all of the system's necessary data transfers.
The sound and subtitles for the movie are stored in a digital storage device 68. This may be a CD ROM which is driven by a suitable drive such as a Toshiba XM3300 CD ROM drive, via an SCSI host adapter. Other types of digital data stores could also be used, such as a hard disc drive, a magneto-optic drive (a laser read/write system with a performance similar to a conventional hard disc drive) , or even a relatively slow access device such as a digital audio tape (DAT) . The primary requirement for the digital data store in this embodiment are (1) its access time must be less than the film travel time from the time code read- ing head 60 to the primary projector 64, thus assuring that a jump can be made to an out-of-sequence time code within the data store before the corresponding portion of the film has reached the projector, and (2) the data store's average data rate capacity must be greater than the average data rate at which information is read out of a FIFO memory (described below) .
The data transfer controller 66 writes the audio data that has been read out of data store 68 into an audio FIFO memory 70, and the subtitle that has been read out of data store 68 into a subtitle FIFO memory 72; both FIFOs are preferably implemented within the microprocessor' s conventional RAM memory.
A typical application for a FIFO memory is to receive
an input data in bursts, but to read out the data at a constant rate. During periods when no data is received, the data already stored in the memory is progressively read out through its output, with each data byte progressively drop- ping down through the successive memory cells (represented by bars 74) . In this manner the memory "empties" out through its bottom, with the upper memory cells becoming vacant as the readout progresses. The next data burst "fills" the memory, while data continues to be read out at the same rate from its bottom. The flow of information through the memory from "top" to "bottom" is indicated by arrows 76.
The minimum capacity size required for the FIFO memories 70 and 72 is equal to the maximum data access time for the data store 68; the maximum effective FIFO memory capacity for any particular readout rate is equivalent to the film travel time from the time code reading head 60 to the aperture of the primary projector 62. Any additional FIFO memory capacity will not be utilized. Data is read out from the bottom of the audio and subtitle FIFO memories 70 and 72 by a read controller 78 that transfers data out at a constant rate. To operate the system, data transfer controller 66 simply turns on the FIFO memory r-ead controller 78 via a control line 80 as the first picture frame arrive at the aperture of projector 64. The theater projector is powered from an AC mains 81, and the readout of data from the FIFO memories 70 and 72 is phase-locked to the AC mains signal. This allows the time code signals to be used simply to assure that the proper audio and subtitle data is supplied to the input of the FIFO memories 70 and 72, without also having to use it for a positive synchronization of the audio and subtitle playback to the film projection. By synchronizing the reading
out of data from the FIFO memories to the AC mains, which also operates the synchronized motor used in the projector, synchronization between the readout of the digital audio and subtitle signals and the projection of the film is as- sured.
The digital audio signal read out from the audio FIFO memory 70 is transmitted by the read controller 78 to digital-to-analog converters (DACs) for the various theater speakers, illustrated as DACs 1-4 for a four-speaker the- ater; a single DAC with a multiple output could also be used. The signals are decoded in the usual fashion and amplified by amplifiers A1-A4 for playback in the theater speakers S1-S4.
The subtitle signal read out from the subtitle FIFO memory 72 is transmitted by the read controller 78 to the supplemental projector 82. The particular format in which the subtitle data is stored and read out to the supplemental projector will depend upon the type of projector used. Some possible variations to this system are indicated in dashed lines. Instead of supplying an input synchronization signal directly from the time code read head 60, the synchronization signal could be provided as serial data 84, possibly from a remote location which also controls the film pro-jection instead of the on-site digital data store 68; the audio and subtitle data could be provided via an electronic delivery service 86. Furthermore, the audio information could be eliminated entirely, and the system used only to synchronize the film and subtitle projections.
FIG. 8 shows a piece of 35 mm release motion picture film 88 with a digital time code in the DTS format. A series of sprocket holes 90 lie between the edge 92 of the film and the normal optical sound track area 94. The picture frames are printed by a picture printing head in an
area 98 that is spaced inward from the sound track area 94. FIG. 9 illustrates the manner in which the print film 88 is read during theater projection. Assume that the film is moving past the projector in the direction of arrow 100. It first passes the digital time code reading head in housing 102, which reads the time code with a colored beam 104 that strikes detector 106 on the opposite side of the film. The film then advances to projection lamp 108. A beam 110 from lamp 108 projects the picture frames onto the theater screen 112.
The picture frames are illuminated by the projection lamp 108 a predetermined period of time after their respective time code units have been read, as determined by the spacing between time code lamp housing 102 and projector 108, and the film speed. This allows time for processing the time code signal, checking its validity, and accessing the proper digital audio and subtitle data. Processing of the time code signal and the production of sound and subtitles therefrom is synchronized with the illumination of the picture frames so that the frames are displayed on the screen at the same time the subtitles and sound derived from their respective digital time codes are respectively superimposed on the film and played back in the theater.
The- two principal approaches to storing the subtitle data in the digital store are (1) adjusting the data frame size so that the number of character spaces included in the frame is just enough to accommodate the subtitle data (plus control data at the beginning and end of each frame) , and (2) using a constant frame size that is at least large enough to accommodate the maximum amount of data for any single subtitle, with unused portions of the frame remaining blank. For example, an LCLV with a typical 800x600 pixel arrays would require a maximum frame capacity of
8,640,000 bits for the maximum level of subtitle information (6 data bits/character space x 3 colors (R,G,B) x 800 600, reduced by any unused portion of the LCLV resulting from restricting the subtitles to only a fraction of the full screen area) . With typically 6 bits per byte or character space, and each character space representing a letter or number in the subtitle or a control character such as carriage return, back space, start of transmission, load text into buffer, print content of buffer now, etc., a considerable savings in subtitle storage space can be achieved by limiting the size of each from to the number of character spaces required for its particular subtitle. This approach is illustrated in FIG. 10, in which three data frames 114, 116, 118 (not to scale) are illustrated. Each data frame begins and ends with control characters 120, including the time code sent for that frame, while the subtitle or text portion 122 of each frame is shown as varying in length; frame 114 stores the shortest subtitle, frame 116 the longest, and frame 118 a subtitle of interme- diate length.
While economical in terms of storage space, the variable size data frame approach makes any individual frame more difficult and time consuming to access, since the exact location of any particular frame is not predictable. In general, to access a frame such as frame 118 would require a fast forwarding (indicated by wide arrow 124) through the storage medium until the desired frame is located, following by a slower readout (indicated by narrow arrow 126) of the frame itself. This limitation can be alleviated by creating a file, in either the storage medium itself or in its computer controller, which correlates the individual time code units with the locations within the storage medium of their re-
spective data frames. FIG. 11 illustrates this process. The storage medium is first scanned to identify the location of each frame within the medium (block 128). Based upon this scanning, a file (illustrated by matrix 130) is created in either the disc or other storage medium 132, or in the controller computer 134, to correlate each time code unit with the location of its corresponding data frame (step 136) . When time code data is thereafter input into the computer from the film (step 138), an almost instantan- eous access jump can be made to the corresponding frame location within the data store (step 140) .
The access process without first creating this file is illustrated in dashed lines. The time code data is input into the system (step 142), initiating a frame-by-frame fast forward through the storage medium until the desired frame is located (step 144) .
The constant data frame size approach is illustrated in FIG. 12. Each frame 146, 148, 150 includes the same number of character spaces. Character spaces required to store a particular subtitle or other supplemental visual display are occupied, with the remaining spaces left blank. This is illustrated in the figure by data frame 146 having a small number of occupied character spaces (indicated by hatching-) , frame 148 having a large number of occupied character spaces, and frame 150 having an intermediate number. Since the length of each frame is known in advance, rapid jump accessing to a desired frame corresponding to a given time code input can be done without prescanning the storage medium to establish a preliminary time code/frame location file.
For either the variable frame size of FIG. 10 or the constant frame size of FIG. 12, the data space required for the supplemental visual display can be significantly re-
duced by having each display frame extend for a duration of multiple time code frames. For example, in a known DTS system the film runs at 30 film frames per second and 24 time code frames per second. If there is only one visual display frame for each 6 time code frames, the visual display will still have a time resolution of second. While a higher resolution is necessary for good audio reproduction, second is sufficient for subtitles, and reduces the data space required for the subtitles by a factor of 6. Subtitle data is presented only once each 6 time code frames, and the same subtitle display is continued until a change appears in the subtitle data.
The subtitle data store is readily integrated with a film's sound track store, as illustrated in FIG. 13. Each data frame includes beginning and ending control character segments 152, a subtitle text segment 154 and an audio data segment 156; one or more control characters could also be inserted between the subtitle and audio segments to note the end of the subtitle data and the beginning of the audio data. The audio data in the DTS system is stored in equal size frames. Since the data storage capacity devoted to the audio track is considerably greater than that needed for subtitle data, there is little advantage to be gained in terms -of maximizing overall storage capacity by adjust- ing the size of the subtitle text space to match the size of each individual subtitle. In FIG. 13, which exaggerates the amount of storage capacity required for the subtitle text relative to the audio data, the portion of the subtitle segments that might be occupied by particular subtitles are again shown by hatching. Only a relatively small portion of the subtitle segment 154 is occupied for the first frame, with a larger occupied portion for the second frame and no subtitle at all in the third frame.
While particular embodiments of the invention have been shown and described, numerous variations and alternate embodiments will be apparent to those skilled in the art. It is accordingly intended that the invention be limited only in terms of the appended claims.