Television – Basic receiver with additional function – Multimode
Reexamination Certificate
2000-12-06
2004-04-20
Lee, Michael H. (Department: 2614)
Television
Basic receiver with additional function
Multimode
C348S443000, C348S452000, C348S097000
Reexamination Certificate
active
06724433
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention is related to processing image sequences, and in particular, to methods and systems for converting an image sequence intended to be displayed at a first frame rate to an image sequence intended to be displayed at a second frame rate.
2. Background
As is well known, motion film is typically exposed and viewed at 24 film frames per second (fps). By contrast, NTSC video, which applies to television, is typically recorded and played back at 29.97 video fps. The selection of 29.97 fps for video is based on the frequency of electricity in the United States, which is 59.94 Hertz (Hz) or cycles per second. Video typically includes two fields per frame, and therefore, there are typically 59.94 fields per second.
For television, the NTSC color video standard specifies that 525 lines of information are scanned at a rate of 29.97 fps, therefore, each field scans 262.5 horizontal lines. However, typically only approximately 480 lines per frame, or 240 lines per field, are active or illuminated and contain actual picture information. The two fields of a video frame are often referred to as being “interlaced.” The lines of information from the two fields of a respective frame interlace, i.e., alternate, to produce the frame. Thus, one field can contain the odd lines of a frame and the other field can contain the even lines of a frame. The two fields are also respectively referred to as “odd” and “even” fields. In addition, the NTSC video standard is not always used. Many users use proprietary standards that are similar to the NTSC video standard. For example, where a frame is encoded by only one field, the resulting video sequence can include frames with 240 lines of resolution at 60 frames per second or 240 lines of resolution at 30 frames per second.
It is a common practice in the movie and television industry to convert from the film format to the NTSC video format so that filmed works can be broadcast and displayed on a television set. Clips of filmed work are also often transferred to a video format, such as the NTSC video format, because video formats are convenient to store and view as well. Such a conversion is known as a “telecine” process, which typically converts 24 film fps to 30 video fps video (in addition to the resizing or letterboxing to accommodate the difference in screen aspect ratio).
To convert 24 fps of film to 30 fps of NTSC video, duplicate or repeated fields are inserted o “pad” the 24 fps to 30 fps. The first film frame is converted into 2 video fields (1 even field and 1 odd field), the second film frame is converted into 3 video fields (2 even fields and 1 odd field), with two of the video fields being the same, the third film frame is converted into 2 video fields, the fourth film frame is converted into 3 video fields, with two of the video fields being the same, and so on. Thus, the video field to film frame pattern is “2, 3, 2, 3,” where an extra video field is inserted for every other film frame. As a result, 4 frames of film convert to 5 corresponding frames of video. This is referred to as a “three-two (3:2) pull down.” To return the 30 fps of video to the original 24 fps of film, a reverse process, termed inverse telecine, is performed, where frames of video convert to 4 corresponding frames of video. Prior methods rely extensively on manual intervention to perform the inverse telecine process.
One significant difficulty encountered in performing inverse telecine is handling edits, slow motion, special effects sequences, or other special cases, wherein the 2, 3, 2, 3 pattern is interrupted. For example, because of an edit or abort during final assembly, the 2, 3, 2, 3 pattern may be interrupted in the middle and restarted as follows 2,3,2,[edit] 2, 3, 2, 3. To correctly return or convert this pattern to the original film pattern, a user locates the pattern break and conventionally resynchronizes the sequence by manually deleting one or more fields. This is a time consuming and expensive process, and in particular, makes difficult the accurate performance of the inverse telecine process on a large number of video clips in a short period of time.
Because of the difficulties encountered in performing the inverse telecine process, the video format is often retained when displaying a clip on a computer. However, the video format can be wasteful because the duplicate frames needlessly occupy bandwidth. Further, the display of duplicate frames causes motion in the clip to transition in a jerky or erratic manner. In addition, where video fields are interlaced, the interlacing of fields based on film frames from different times can produce artifacts, which are visible on a progressively scanned monitor, such as a computer video monitor.
SUMMARY OF THE INVENTION
The present invention is generally directed to automated methods and systems for converting image streams having a first frame rate to a second frame rate without the need for user intervention. Embodiments of the present invention obviate the effects of a telecine process, wherein additional frames are added to accomplish the frame rate conversion. In one embodiment, a statistical analysis of the differences between pixels in adjacent frames or groups of frames is performed to detect a telecine pattern, thereby identifying which frames to remove.
In another embodiment, where frames are encoded using both even and odd video fields, a statistical analysis of the differences between adjacent fields detects the telecine pattern, identifies which frames to remove, and identifies frames that are candidates for re-interleaving. The novel process disclosed herein can detect and delete the duplicate frames of the telecine process for video sequences with interlaced or non-interlaced frames, and/or of various resolutions.
Video image streams are frequently converted from a film format to a video format through a process known as a telecine process. Although the telecine process allows a sequence originally taken in film at 24 fps to be stored in a video format at 30 fps and displayed on a television monitor, the process typically results in duplicative frames, jittery motion, and interleaving of disparate frames. By providing a technique to automatically perform an inverse telecine process to substantially return the sequence to the film format, the picture quality improves and the bandwidth needed to transmit the processed sequence is reduced.
The techniques for performing the automated inverse telecine processes can be implemented in a server connected to the Internet or other network. The Internet allows a variety of users to communicate with the server. A user can upload, in real time or from a storage device, a first video sequence to the server. The server processes the uploaded video sequence either substantially in real time or in the background. While processing in real time or after processing in the background, users can download the processed video sequence from the server.
In addition, one embodiment of the present invention automatically detects whether the incoming video sequence is encoded in a single field or in multiple fields by counting the number of lines per frame and comparing the count to a predetermined amount.
Where the frames have been encoded in single fields, i.e., wherein a frame is composed of one field, the process computes comparisons of the adjacent frames in the sequence. The comparison can be made on all the pixels of each frame, or on a portion of the pixels, such as every other pixel, every fourth pixel, or some other interval of pixels. A history of the comparisons is maintained. One embodiment compares both the luminance and the chrominance components of a pixel. Another embodiment compares only the luminance component.
The pixels can be compared in a variety of ways. For example, the computation of the comparison can include summations of the absolute differences between pixels, summations of the squares of differences between pixels, and the like. In one embodiment, the summation is furthe
Knobbe Martens Olson & Bear LLP
Lee Michael H.
RealNetworks, Inc.
Stewart Steven
Tran Trang U.
LandOfFree
Automated inverse telecine conversion does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Automated inverse telecine conversion, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Automated inverse telecine conversion will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3190689