CN1174645A - Picture transfer process and device - Google Patents

Picture transfer process and device Download PDF

Info

Publication number
CN1174645A
CN1174645A CN 94195223 CN94195223A CN1174645A CN 1174645 A CN1174645 A CN 1174645A CN 94195223 CN94195223 CN 94195223 CN 94195223 A CN94195223 A CN 94195223A CN 1174645 A CN1174645 A CN 1174645A
Authority
CN
China
Prior art keywords
image
film
sequence
cloth
width
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN 94195223
Other languages
Chinese (zh)
Inventor
兹比格纽·赖布茨津斯基
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
CFB Centrum fur neue Bildgestaltung GmbH
Original Assignee
CFB Centrum fur neue Bildgestaltung GmbH
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by CFB Centrum fur neue Bildgestaltung GmbH filed Critical CFB Centrum fur neue Bildgestaltung GmbH
Priority to CN 94195223 priority Critical patent/CN1174645A/en
Publication of CN1174645A publication Critical patent/CN1174645A/en
Pending legal-status Critical Current

Links

Images

Landscapes

  • Image Processing (AREA)

Abstract

A process is disclosed for producing a second picture sequence (target picture sequence) having a second picture frequency and a corresponding second picture interval from a first picture sequence (source picture sequence) with a first picture frequency and a corresponding first picture interval. The generated picture sequence serves in particular to give to an observer the visual impression of an event happening in time. At least part of the pictures of the target picture sequence is obtained by additive superposition of at least two pictures of the source picture sequence that follow each other in time and are evaluated with an intensity weighting factor. All pictures of the target picture sequence are obtained by superimposing at least two pictures of the source picture sequence evaluated picture by picture on the basis of individually calculated intensity weighting factors.

Description

The method and apparatus of image transitions
In film and TV industry, exist multiple different picture system standard.These standards are given for producing the required number that is recorded in single image on video recording or the film band of moving frame.A kind of image document of standard often need be play with the projector equipment of another kind of standard, therefore this image document must be processed into the image document that is applicable to the projection equipment standard.This process is referred to as image transitions.
The present invention relates to the conversion of TV, computer or film image.These images are produced by the recording system of a kind of specific standard (being given per second picture number), and will be by optical projection system (the being different per second picture number) projection of another kind of standard.In present patent application, these different recording systems are referred to as " picture system ".Existing picture system mainly contains following 6 kinds:
1.NTSC-TV:60 images/sec (" U.S. " standard),
2.PAL TV:50 images/sec (" Europe " standard),
3.Film 16:32 images/sec (archives sheet),
4.Film 24:48 images/sec (cinema film standard),
5.Film 60:120 images/sec (reading scan system),
6. computer picture: the per second picture number is able to programme.Below these systems are described in detail respectively.
1.NTSC TV: the NTSC nominal value is per second 30 width of cloth " image ".But every width of cloth image has comprised half alternate figure of 2 " frames ".This two frames, half image is formed a width of cloth ntsc television image in the following manner: first half figure provides the image information of all odd-numbered lines, and second half figure then provides the image information of all remaining each row, the i.e. image information of even number line.Therefore, to be also referred to as be " interlacing scan " image to the NTSC image.Because every frame is being represented the different time intervals, so every frame should be considered as discrete image.Can think that thus the NTSC per second has 60 width of cloth images.
2.PAL TV: the PAL nominal value is per second 25 width of cloth " image ".With the ntsc television image similarity, every width of cloth image has also comprised half alternate figure of 2 " frames ", and also is " interlacing scan ".Equally, because every frame is being represented the different time intervals, so every frame should be considered as discrete image.Can think that thus the PAL per second has 50 width of cloth images.
3.Film 16: the film before sound film is produced with per second 16 width of cloth " image ".These images are positively to see from the sheet band of archives film.But in actual film production and projection process, per second has image exchange 32 times, wherein has only 16 exchanges to show image.Other 16 exchanges are because shutter closes generation in the forward (FWD) process is black " image ".This can find out from the secret note band between unexposed portion on the cinefilm or two width of cloth exposure images.During animation effect in analyzing motion picture projection, these invisible " picture black " roles are identical with visual picture.Can think that thus the Film16 per second has 32 width of cloth images.
4.Film the nominal value of 24:Film 24 is per second 24 width of cloth " image ".These images are positively to see on film is with.Similar with Film 16, in actual film production and projection process, in fact there is each second the image of a multiple amount to produce, promptly per second has image exchange 48 times.Wherein have only 24 exchanges to show image.Other 24 exchanges are because shutter closes generation in the forward (FWD) process is black " image ".Can think that thus Film 24 per seconds have 48 images.
5.Film the nominal value of 60:Film 60 is per second 60 width of cloth " image ".These images also are positively to see from the film band.In actual film production and projection process, in fact image exchange there is 120 times each second.Wherein have only 60 exchanges to show image.Other 60 exchanges are because shutter closes generation in the forward (FWD) process is black " image ".Can think that thus Fihm 60 per seconds have 120 width of cloth images.
6. computer picture: computer graphic similarly is manually to produce the image of unreality " photography ".Image motion (animation) is a computational process, applicable to any picture system.When specific computer image system of design also will be transferred in another picture system, they should be considered as a system.
High-resolution (" high-definition ") TV will be the improvement of NTSC and PAL standard as the future of TV.It works in accurately on the per second picture number (per second 50 width of cloth images or per second 60 width of cloth images accurately).NTSC and PAL have a slight image deviations rate for the benchmark of 50 width of cloth per second images and 60 width of cloth per second images.For the image transitions between these two systems, it is important is actually 60 width of cloth NTSC images and 50 width of cloth PAL images and can be exactly finishes at one time, just finishes (speak by the book is 1.0010 seconds) in a second.
If when showing its image with former recording system, the dynamic effect that image shows on screen is always making us satisfied effect in varying degrees.
For original system was recorded, the dynamic effect that can reach with NTSC and Film 60 projections was best.Because these picture systems have the highest per second picture number.The per second picture number is high more, and human eye is good more for the perception of dynamic effect, and the continuity of its action just is level and smooth more.
For recording a dynamic process, the poorest original image system of effect is Film 16.For Film16, its actual movement effects can't have been seen (32 images/sec) on former cinefilm, because those original image projection equipment have not existed." acceleration " effect of this file film is owing to use current equipment, i.e. the result (Film 24) of 48 images/sec projection.This invention can provide a kind of possibility by its original speed playback for making these films.
In the broadcast of film and TV, can produce the image disruption of two kinds of distortions.A kind of distortion is so-called stroboscopic distortion (" strobe-flicker "), and another kind of distortion is the what is called distortion (" jitter ") of beating.The stroboscopic distortion only is owing to recording or too low caused for producing the used image rate of continuous images motion during projected images.Because the basic fundamental difference that adopts, so it is also different to be presented at the type of all kinds of stroboscopics on the screen monitor in movie or television.But the distortion effect that this class distortion and image are produced when a system is transformed into another system is irrelevant.
Flating (" image jitter ") then is because the image conversion method images of used pure machinery is caused: in the target image system, systematically remove and repeat to show the interruption of the motion process that single image can cause on the film being recorded.
Because this distortion is relevant with the image transfer rate, with very short interval repetition, thereby produce a kind of rocking, the image impression of shake in varying degrees.United States Patent (USP) 1,815,455 (Waller) and 5,153,620 (Songer) are devoted to eliminate the stroboscopic distortion.Although as above said, the stroboscopic distortion seemingly is necessary to discuss tout court these two patents not owing to image transitions produces because in these two patents this to have described with this invention be method relevant on the surface.
The Waller method attempts to be shown on the same width of cloth image by image that will be right after previously and the image that is right after later, in order to reduce the chatter phenomenon.This overlapping projection has caused a kind of interesting visual effect " layer of moving image is formatted ".
This layer is formatted, and effect can not solve stroboscopic (" strobe-flicker ") or (" the jitter ") problem of beating.The patent of Songer has been described two width of cloth images has been superimposed as piece image with identical ratio, attempts in order to reduce the chatter distortion when the motion picture projection.This stack has caused fuzzy (" blurred ") of image.It can not solve stroboscopic (" strobe-flicker "), can not solve (" the jitter ") problem of beating.
The method purpose of Waller and Songer is not that the image document that is produced replay in another kind of system is come out, because the picture number in its set time unit does not change.Dual or multiexposure, multiple exposure under constant image rate can not solve the image transitions problem, also can cause projected images debase thus, i.e. the image quality decrease (unintelligible) of relative original image.
Now, every kind of picture system all should be able to be shown the image of being recorded by another kind of picture system.If the quantity of original image is shown out (under the situation at Film16) by a new picture system simple and mechanically, then can produce the serious motion and the distortion of sound at playback time.For example, if the video recording that projection is recorded by the NTSC system type in the PAL system then can produce the phenomenon that slows down.And the video recording that projection is recorded by the PAL system type in a NTSC system then can produce the phenomenon of quickening.In order to keep the original motion visual effect, need make the picture number of recording image data be adapted to the picture number of projection system, the process of this adjustment picture number is referred to as conversion or conversion.The conversion of the image that is adopted from a kind of picture system to another kind of picture system up to now can cause flating (" jitter ").To elaborate its reason below:
When object of eye-observation, what see is the piece image in a certain moment.This width of cloth figure is in motion continuously.Yet still there is not the technology that to record this class image, the substitute is such continuous motion is recorded with a series of discrete rest images.
For can produce the impression of continuous motion on screen, human eye must be able to be observed continual (60 width of cloth images/sec preferably can a be arranged) rest image sequence, so that can produce a kind of illusion of continuous motion.This is the basic principle of all images and image switching system.The interruption of any picture system (for example lack and go or multiimage, or the time of image arranges inaccurate) all can produce the interference to the movement vision perception.If take away piece image wherein, then can in the process of motion, produce a vacancy.If repeat piece image wherein, then can in the process of motion, produce a kind of " stagnating ".In order to make example with the dial plate of clock and watch here to the description of doing by the variety of problems that is produced in the mechanical transfer process to simplify.
The object of observing is the second hand of table now.Suppose that the second hand of table produced by a picture system, and its per second only claps piece image, whole recording process had 4 seconds.Every width of cloth image in the recorded material has shown the step that second hand is goed ahead.If whole image data is shown with another system, only needed 3 width of cloth images (being every width of cloth image 1.33 seconds) in its per 4 seconds and used mechanical conversion method in image transitions, then piece image wherein has been removed, a step of second hand can be removed thus, can show second hand beating in its motion process on the screen.This beating is caused owing to lacking image (lacking that step that is removed).
Still in same example, if recording system is to record 3 width of cloth images in 4 seconds, then every width of cloth image is the interval demonstration second hand with 1.33 seconds.If this image document is shown in a per second is shown the system of 4 width of cloth images, get the time period in 4 seconds and observe, and used mechanical transform method, then need to repeat to show the image that a width of cloth is recorded.Just a second hand step is wherein repeated projection, can show stagnate (" freeze ") of motion in second like this on the screen.This " stagnating " effect is because repeating step is caused.
Remove beat or freeze beyond, existing image has also been represented wrong visual information in above two examples.Removing (image jump) under the situation of image, every width of cloth image is projection in 1.33 seconds process, but in fact this width of cloth image the process that should show be 1 second.In the process that image need repeat (freeze frame), the time that every width of cloth image shows is 1 second but the process of this image representative should be 1.33 seconds.
Therefore use the method for mechanical switch, correct image time and movable information can not be provided.All are called as beat (" jitter ") by the caused interference of mechanical conversion.Although existing picture system is with 1 images/sec or 1.33 images/sec work, but with 32,48,50 or 60 images/sec work, it is identical by " jitter " effect that mechanical conversion produced with foregoing effect.
Although many in the world companies have carried out diversified trial, attempt to solve " jitter " problem, also do not find gratifying solution so far.
United States Patent (USP) 3 511 567 (Dejoux) and 4 889 423 (Trumbull) are at the image transform in film and the television industry.The content of these patents will be done reference in this article.Because these conversion are based upon on the basis of deletion and multiimage, so in present patent application these methods are called mechanical transform method.The image conversion method of these document descriptions is based upon on the following principle: when original video recording image document contains too much images, then with unnecessary image-erasing.When the contained image of original video recording image document is not enough, then to allow some image repeat, so that produce enough picture number.
The image conversion method that is proposed in the Dejoux patent is put into practice in television industry for many years.But these transform methods can not solve " jitter " problem, and in fact opposite Dejoux method has caused " jitter ".
Trumbull has described and a kind of a moving image converting with higher image switching rate has been become a transform method with moving image of low image switching rate.This method is by superposeing and removing some images and finish.The method of Trumbull can not solve " jitter " problem, because the time relationship of original image system and timing are not retained.Compare with the method for Dejoux, its image document through conversion has just shown the quantitative passage of time with respect to the original image motion.Two kinds of methods have all been destroyed the time relationship of output source map picture, the shake " jitter " that produces moving image thus when image transitions.
This problem can be illustrated with following further example:
Fig. 1 has shown a sequence that 4 width of cloth images are arranged, i.e. the image of being made up of 4 width of cloth exposed plates.The image that produces on the film is a circulatory motion object.For convenience's sake, time course with the number of degrees (°) provide, 360 ° of sequences that show one 4 width of cloth image wherein, the time span of its representative is 1/6 second.As shown in Figure 1, this circulatory motion object is because its motion and formed an image that has extended in the time interval (45 ° or 1/48 second) of an image exposure.There is equal time gap (or angle) at the center of the every width of cloth in this 4 width of cloth image apart from other image, and this time gap (or angle) is equivalent to the switching rate of image.Because the hypothesis object is in the linear movement of doing at the uniform velocity, their adjacent mid point distance equates.The Edge Distance of adjacent image also equates.When these images were observed with corresponding recording technology, the observer can perceive a kind of uniform linear movement of object.
Fig. 2 has provided a film image sequence, and it is that a same annular object is done uniform linear movement.It is carried out recording of 60 width of cloth images/sec, use known method to convert it to Film24 then.As can be seen from Figure 2: all relative original image of each width of cloth in 4 width of cloth images has been extended.These images that extended will be owing to using motion picture technique be retained to get off.As can be seen from the figure, have overlappingly among first width of cloth and second width of cloth figure, and the distance of its second width of cloth figure and the 3rd width of cloth figure is obviously drawn back.When this film was observed on Film 24 projectors, the observer can see in figure one and figure two transition regions and an object of which movement slowly occurs, a kind of acceleration, uneven object of which movement occurred in image two and image three transition regions.This gives the impression of observer's image in shake (" jitter ") on the whole.
Before the present invention, thus image conversion method, all be based on various theories of mechanics as conversion between NTSC and PAL.In other words, piece image or be removed, or be repeated.
The mechanical image conversion method of modified form recently all be devoted to " smoothly " this image shake, jump over or stagnate.Yet these smoothing methods can not be dealt with problems rightly, for the reduction that smoothly can cause definition quality of " jitter ", and the wrong time that can not solve every width of cloth image concern problem.
In order under the situation that does not cause " jitter ", to keep correct time movement relation, must find out new image conversion method, they should be able to make image produce image in conversion and recording process.These images can have identical characteristic with the former identical image of recording, so all must produce new image under these situations.What the present invention proposed is the conversion and the conversion method that can keep original image time relationship characteristic.
The present invention relates to a kind of image is transformed into the method for second picture system from first picture system, this method can not resemble and cause that image disruption relevant with image motion and time change the known mechanical switch method.First sequence that comes from first picture system (being the original image system) contains multiple image.Similarly, second sequence that comes from second picture system (i.e. the new sequence that will produce) also contains multiple image.Every width of cloth image in new second picture system that constitutes all is to be formed by stacking by a kind of special relationship each other by at least two width of cloth images in first picture system, and promptly first image sequence is that weighted factor by separately reflects to the time gap between the newly-generated image.Can use in this way from the image of first picture system and be combined into second picture system, thereby get rid of flating (" jitter "), stagnate and jumped over phenomenon in a kind of mode of level and smooth smoothness.This image conversion method can be used for the mutual conversion such as image between following each picture system: ntsc television, and the PAL TV, Film 16, Film24, Film 60 (" Showscan " system) and computer picture.
Corresponding to method of the present invention generally is to produce each width of cloth image in the target image sequence fully again and the image of original existence is not the part of target sequence.At this, for to produce in each target sequence image should find out in the output sequence for two width of cloth correspondence image, they are the most alike in the image that will produce on the point at this moment.What should find out equally is time consistency degree between these images.This time consistency degree has been determined the percentage of image participation generation target sequence image in the output sequence.Different with the method that exists so far is that this percentage becomes with single image in the given image sequence (calling the primary image sequence in the following text).But be used for producing the method double exposure (film) or the electronic image mixing (TV) of new images.When using these two kinds of methods, be to realize with any special measures respectively for the weighting of output image.
Image conversion method of the present invention also intends not considering the single image in the output image sequence when producing target image sequence.But this way matches at this and purpose of the present invention, does not promptly produce at playback time and is interrupted sense.The time interval when these images are in goal systems and provide " picture black " particularly is perhaps when they do not belong to and will produce the image time during adjacent that two width of cloth image.The following image that will remove in the output image sequence of some change over conditions obtains restriction with regard to making in practice like this, be that a higher picture system of image rate and goal systems are under the situation of the lower film image system of image rate (for example, 16 change over condition from NTSC to Film) promptly in output system.
Advantage of the present invention and strong point can be by more clearly embodying in conjunction with following description and accompanying drawing, wherein
Fig. 1 is a schematic diagram, the effect that it has been produced when having shown a cyclic motion object electricity consumption cinema system projection.These images link up.
Fig. 2 is a schematic diagram, the effect that it has been produced when having shown a cyclic motion object electricity consumption cinema system projection.These images are to come from the TV signal conversion by the known image conversion method in advance.
Fig. 3 is a schematic diagram, the effect that it has been produced when having shown a cyclic motion object electricity consumption cinema system projection.This is the situation after TV signal is changed by the method in this invention.
Fig. 4 a is a time diagram, and it has shown TSC-system formula picture system.
Fig. 4 b is a time diagram, and it has shown the pal mode picture system.
Fig. 5 is that figure is split in a rotation, and it has shown NTSC and PAL picture system and the relation of the time series between them.
Fig. 6 is that figure is split in a rotation, and it has shown NTSC and Film 24 picture systems and the relation of the time series between them.
Fig. 7 is that figure is split in a rotation, and it has shown Film 60 and Film 24 picture systems and the relation of the time series between them.
Fig. 8 is that figure is split in a rotation, and it has shown Film 60 and Film 16 picture systems and the relation of the time series between them.
Fig. 9 is that figure is split in a rotation, and it has shown Film 60 and PAL picture system and the relation of the time series between them.
Figure 10 is that figure is split in a rotation, and it has shown Film 24 and Film 16 picture systems and the relation of the time series between them.
Figure 11 is that figure is split in a rotation, and it has shown NTSC and Film 16 picture systems and the relation of the time series between them.
Figure 12 is that figure is split in a rotation, and it has shown PAL and Film 16 picture systems and the relation of the time series between them.
Figure 13 is that figure is split in a rotation, and it has shown PAL and Film 24 picture systems and the relation of the time series between them.
Figure 14 is that figure is split in a rotation, and it has shown NTSC and Film 60 picture systems and the relation of the time series between them.
Figure 15 is that figure is split in a rotation, and it has shown that NTSC and PAL picture system are by an altered time series relation of zero-time deviator.
Figure 16 is a simplified block diagram, and it has shown the first Analysis of Nested Design scheme of the present invention.
Figure 17 is a simplified block diagram, and it has shown the second Analysis of Nested Design scheme of the present invention.
Fig. 6 has shown vertical shading line black sector to Figure 14, expression be black " image " of not exposure.
If film or video tape do not have the scene of taking moving, when promptly not moving, then there is not the shake (" jitter ") of image in the image that is produced by mechanical image conversion method.It is visible that the shake of image (" jitter ") has only when motion occurring.This just why the scene of some specific projections can produce good or bad optics impression.Because (for example, shaking of video camera inclined to moving with camera lens and stretched or the like in the influence of numerous variations.), the very difficult intensity of determining " jitter ".Because production is a width of cloth width of cloth single image, they do not contain the visual information about motion and time, are very difficult to proofreading and correct of " jitter " afterwards therefore.
Fig. 3 has provided the method corresponding to first kind of way of realization of the present invention.Wherein for adjacent ntsc television image of time corresponding to the equitant exposure intensity of each image area image of time interval, change with respect to the time interval of the film image that will produce (Film 24).According to this way of realization of the present invention, film image is used following map intensity and is produced:
Fig. 1=87.5%1A=12.5%1B
Fig. 3=37.5%2A+62.5%2B
Fig. 5=87.5%3B+12.5%4A
Fig. 7=37.5%4B+62.5%5A
Wherein do not use regional 5A and 7B.
Should notice that at this prerequisite of simplifying these exposure intensities is, the time interval that film image 1 is placed in also is the position that the interval 1A of television image begins.The professional can find out the possibility that also has other, promptly simply the original position of film image 1 is passed with respect to television image 1A, can produce different percentage relations, but its principle is constant, be the degree of overlapping difference of film time for exposure interval corresponding to time interval, the principle that provides area image intensity in proportion is constant.
According to method shown in Figure 3, each width of cloth film image is the information of adjacent television image of the time of being applied in, and produces by the image intensity weighting.The weight of wherein every width of cloth television image has reflected in time overlapping of the time interval in television image territory and film image time interval.
As shown in Figure 3, the film image as a result 1 of generation comprised a dark relatively elongation the relative weak image of image of ring-type object of which movement with one.The former is the position that is positioned at the position of relative TV territory Figure 1A, and the latter is with respect to the position in the position of television image 1B, and an image that has " shade " that has extended is so arranged.The inventor confirms, observes the image of this having " shade ", is that the perceptual effect of the film image that produces with Fig. 1 is identical to the impression of the position of a moving object concerning the observer.Can be with similar method with the image among television image territory 2A and the 2B, the exposure intensity in according to the present invention combines with it, and the time stack of its television image territory 2A and 2B is with respect to film image 2.So " digraph " result who produces has produced a kind of impression that more is close to image area 2B relevant position to object.
As mentioned above, elite of the present invention is that every image sequence that will produce with given standard can be synthesized by the image of recording with another kind of standard.Wherein original and image sequence that will produce should act on the common time shaft and its common time scale is arranged.The image that will produce is to produce new images by the image in equitant with its time or the contiguous original sequence at every turn.At this, all distributed a weighted factor for employed image in each original sequence, it has determined the contribution of this image in new images.This weighted factor be with the degree of time-interleaving and new images that will produce and original image in the distance dependent that on time shaft, drawn back.
The first step of deriving these weighted factors is to set up one for output and the common time reference of goal systems.Be that time relationship between the single image of output system or goal systems need only be by analyzing the length between these two time points at this basic thought, it is just passable that the time span that synchronous images switches promptly takes place.Therefore the entire image transfer process can be regarded as a kind of periodic repetition.It is made up of some " basic " image sequences (hereinafter referred to as the primary image sequence).Also be adapted to all cycles of this image transitions for the weighted factor that is used to a primary image sequence to be calculated.Length and the weighted factor that can how long reuse a primary image sequence are definite by the situation of image transitions.
For each picture system and each change over condition, picture number in its primary image sequence can be by following method simple computation: by two image rates that participate in the converted image systems, find out their least common multiple, then divided by corresponding image rate, the picture number of primary image sequence in two respective image systems that resulting two numbers are given.In calculating the image rate that should use be that image rate that can draw the accurate image exchange of per time unit.When application image speed, should be taken into account time response at the horizontally interlaced image (" Interlaced ") of video recording in the image, simultaneously will be image exposure in film image or the situation when not exposing (" black square image ") count.When calculating the length of primary image sequence, should be taken into account an important factor: for the situation of all images conversion, if wherein the given primary image sequence number of at least one film image system is an odd number, then should be with the length doubles of primary image sequence.Because otherwise can make two primary image sequences that are connected produce different effective (not deceiving) film image numbers.This rule in fact should be applied in the conversion between all film image systems.
The important change over condition of in actual applications each of picture number in the primary image sequence.
As mentioned above, elite of the present invention is that every image sequence that will produce with given standard can be synthesized by the image of recording with a kind of standard.Wherein original and image sequence that will produce should act on the common time shaft and its common time scale is arranged.The image that will produce is to produce new image by the image in equitant with its time or the contiguous original sequence at every turn.Use for each at this that employed image has distributed a weighted factor in original sequence, it has determined the contribution of this image in new images.This weighted factor be with the degree of time-interleaving and new images that will produce and original image in the distance dependent that on time shaft, drawn back.
The first step of deriving these weighted factors is to set up one for output and the common time reference of goal systems.Be that time relationship between the single image of output system or goal systems only need be by analyzing the length between these two time points at this basic thought, it is just passable that the time span that synchronous images switches promptly takes place.Therefore the entire image transfer process can be regarded as a kind of periodic repetition.It is by some " basic " image sequences-form hereinafter referred to as the primary image sequence.Also be adapted to all cycles of this image transitions for the weighted factor that is used to a primary image sequence to be calculated.Length and the weighted factor that can how long reuse a primary image sequence are definite by the situation of image transitions.
For each picture system and each change over condition, picture number in its primary image sequence can be by following method simple computation: by two image rates with reference to the converted image system, find out their greatest common divisor (ggT), then divided by corresponding image rate f S, T:
n s，t＝f s，T/ggT(f s，T)
Resulting two number n S, tThe given picture number of primary image sequence in the two corresponding picture systems.In calculating the image rate that should use be to draw per time unit's that image rate of image exchange accurately.When application image speed, should be taken into account the time response of the horizontally interlaced image (" Interlaced ") of video recording in the image and image exposure in film image or the situation when not exposing (" black square image ") count.When calculating the length of primary image sequence, should be taken into account an important factor: for the situation of all images conversion, if wherein the given primary image sequence number of at least one film image system is an odd number, then should be with the length doubles of primary image sequence.Because otherwise can make two primary image sequences that are connected produce different effective (not deceiving) film image numbers.This rule in fact should be applied in the conversion between all film image systems.
The important change over condition of in actual applications each of picture number in the primary image sequence.
Be positioned at time point t among the target image sequence T iNew synthetic i width of cloth image F I, T(t i) be to be positioned at time point t by two width of cloth adjacent with this time point jAnd t J+1Sequence of source images S in image, F J, s(t j) and F J+ 1, s(t J+1), produce by following relation:
F i，T(t i)＝C j，s×F j，s(t j)+C j+1，s×F j+1，s(t j+1)
Wherein, i and j are integers, C J, s, C J+1, sBe to be respectively the intensity weighted coefficient of j and the j+1 width of cloth (source) image, they can be calculated as follows and draw:
C j，s＝1-(|t i-t j|)/(t j+1-t j)，
C j+1，s＝1-(|t j+1-t i|)/(t j+1-t j)。
The method of a synthetic new images has comprised more following basic steps:
By a markers determine one for first image sequence (sequence of source images) and second image sequence (target image sequence) that will produce common time scale.
2. find out with to customized type transmission or playback time, every width of cloth image residing position on time scale in source images and the target image sequence.The position of these images is to be represented by the characteristic point of their time interval (should adopt central point).
3. for every width of cloth image that will produce in the target image sequence, find out in the sequence of source images corresponding with it picture position on time scale, also be those and its adjacent picture position of interval central point of time.
4. calculate the distance of the time interval characteristic point (should adopt central point) of respective image in first image sequence to the time interval central point of the target image sequence image that will produce.
5. the image calculation in each corresponding sequence of source images is drawn an intensity or the amplitude weighting factor with image-related normalization, it and from the target image sequence that will produce between the image area of image characteristic point to the distance dependent of characteristic point the image area of itself.
6. each picture material that will produce is to multiply by superimposed formation behind separately the weighted factor by the respective image in the sequence of source images in the target image sequence.
Therefore by series of features identical (picture material is different certainly), isometric primary image sequence constitutes target image sequence.Its time length is by an above-described as far as possible little integer image, and it just contains expands corresponding image in first and second image sequence.Be in order to produce each image of target primary image sequence wherein with image-related weighted factor, according to foregoing method, calculate under the situation of output image sequence and target image sequence considering that " picture black " interval and/or horizontally interlaced image are present in.
For further setting forth the principle of the used method of the present invention, developed one notional " device " here, hereinafter referred to as " rotation is split ".Use this " rotation is split " can be distributed to the mathematical value of image in each new images system, has provided a kind of approach that makes the result of calculation sharpening of real image conversion gained thus.
" rotation is split " used the principle of foregoing watch hands.In two different picture systems, a sequence is the number of presentation video and the time of demonstration, at this moment between in from two systems minimum possible integer number purpose image can pass through.For example relatively PAL and NTSC-as shown in Figs. 4a and 4b-first complete sequence has 6 figure and at PAL 5 figure arranged at NTSC.The time span of this sequence is 0.1 second.The length of this sequence in the TSC-system formula (0.0167 images/sec multiply by 6 images and equals 0.1 second) and in pal mode (0.02 images/sec multiply by 5 images and equals 0.1 second) all be 0.1 second.
In an image conversion process, must in the time period of a sequence, produce new image and constantly repeat this fixing sequence.Such as in above-mentioned example, NTSC is to 5 new figure (equally also being in 0.1 second) in pal mode by 6 original graph (in 0.1 second) in NTSC to the conversion of pal mode.Conversion from PAL to the TSC-system formula is to producing 6 new figure in the TSC-system formula by 5 original graph among PAL conversely.Method need produce new image thus, and makes the constantly fixing repetition of each film image sequence.
For understanding the operation principle of " rotation is split ", people can imagine a white indicator by 360 ° of such speed rotations of each sequence on the black dial plate.When recording this indicator, can record 6 width of cloth images at each whole circle with the TSC-system formula.In the exposure process of every width of cloth image, the indicator of white has passed through 60 ° splits (360 °/6 images), sees Fig. 4 a.When this indicator is recorded with pal mode, can record 5 width of cloth images at each whole circle.In the exposure process of every width of cloth image, the indicator of white has passed through 72 ° splits (360 °/5 images), sees Fig. 4 b.
Then each figure in this sequence has a mathematical value in this way.Can help to provide mathematical relationship between the image by illustrated method.NTSC-PAL: sequence length is 0.1 second NTSC among Fig. 4 a
From 0 to 60 ° of figure #1=
From 60 to 120 ° of figure #2=
From 120 to 180 ° of figure #3=
From 180 to 240 ° of figure #4=
From 240 to 300 ° of figure #5=
PAL (a=preamble sequence, the follow-up sequence of b=) from 300 to 360 ° of (360 °=0 °) Fig. 4 b of figure #6=
Figure from 288 to 360 ° of #5a=(360 °=0 °)-(the last piece image of preamble sequence)
From 0 to 72 ° of figure #1=
From 72 to 144 ° of figure #2=
From 144 to 216 ° of figure #3=
From 216 to 288 ° of figure #4=
Figure from 288 to 360 ° of #5=(360 °=0 °)
Figure from 0 to 72 ° of #1b=(first width of cloth image of follow-up sequence)
By each figure is used such mathematical value, just might find out the mid point (" central timing factor ") of the time series factor of each image with following method.NTSC image mid point (see Fig. 4 a):
Scheme #1=30 °
Scheme #2=90 °
Scheme #3=150 °
Scheme #4=210 °
Scheme #5=270 °
Scheme #6=330 ° of PAL image mid point (seeing Fig. 4 b):
Fig. 5 #A=-36 ° (from 324 ° of preamble sequence)
Scheme #1=36 °
Scheme #2=108 °
Scheme #3=180 °
Scheme #4=252 °
Scheme #5=324 °
Scheme #1b=396 ° (from 36 ° of follow-up sequence)
These numerical value have been set up the prerequisite of calculating single image time relationship in two picture systems and corresponding weighting factor.With this process of diagrammatic representation, now the principle of " rotation is split " is done an expansion, be about to unified being synthesized on " rotation an is split " figure of conversion " rotation is split " of two picture systems of participation conversion.Can see at once that from such width of cloth schematic diagram the figure of which two output system schemes in time accordingly corresponding to a primary image sequence of goal systems.Fig. 5 has provided NTSC to the conversion between the PAL.As example, provided the image corresponding relation between the picture system below and described the calculating of its weighted factor.
The mid point of every width of cloth figure of a system can be observed from the angle of another system now.Mutual observation in NTSC and pal mode picture system angle can easily be found out from Fig. 5, can see that wherein PAL sequence and NTSC sequence are overlapped on the scale of a circumferential.For example from the NTSC angle, then NTSC image #2 (is the center with 90 °) is prior to 18 ° of PAL image #2 (is the center with 180 °), but lags behind 54 ° of PAL image #1 (is mid point with 36 °) in time.Can be by all images in the following analysis NTSC system with the method: A.NTSC#1 (is mid point with 30 °) be ahead of 6 ° of PAL#1 (is mid point with 36 °), and stagnates
The back in PAL#5a (with-36 ° be mid point, come from presequence) 66 °, B.NTSC#2 (is mid point with 90 °) is ahead of 18 ° of PAL#2 (is mid point with 180 °), and
Lag behind 54 ° of PAL#1 (is mid point with 36 °), C.NTSC#3 (is mid point with 150 °) is ahead of 30 ° of PAL#3 (is mid point with 180 °),
And lag behind 42 ° of PAL#2 (is mid point with 180 °), D.NTSC#4 (is mid point with 210 °) is ahead of 42 ° of PAL#4 (is mid point with 252 °),
And lag behind 30 ° of PAL#3 (is mid point with 180 °), E.NTSC#5 (is mid point with 270 °) is ahead of 54 ° of PAL#5 (is mid point with 324 °),
And lagging behind 18 ° of PAL#4 (is mid point with 252 °), F.NTSC#6 (is mid point with 330 °) is ahead of PAL#1b, and (with 396 ° is that mid point is (at next
In the sequence 36 °)) 66 °, and lag behind 6 ° of PAL#5 (is mid point with 324 °).
The center of each image also can be analyzed from PAL " angle ".For example PAL image #2 (is mid point with 108 °) lags behind 18 ° of NTSC image #2 (is mid point with 90 °) in time, and is ahead of 42 ° of NTSC#3 (is mid point with 150 °) in time.Image in all PAL systems can be done as above with the method and analyze: A.PAL#1 (is mid point with 36 °) lags behind 6 ° of NTSC#1 (is mid point with 30 °), and surpasses
Preceding in 54 ° of NTSC#2 (is mid point with 90 °); B.PAL#2 (is mid point with 108 °) lags behind 18 ° of NTSC#2 (is mid point with 90 °), and
Be ahead of 42 ° of NTSC#3 (is mid point with 150 °); C.PAL#3 (is mid point with 180 °) lags behind 30 ° of NTSC#3 (is mid point with 150 °),
And be ahead of 30 ° of NTSC#4 (is mid point with 210 °); D.PAL#4 (is mid point with 252 °) lags behind 42 ° of NTSC#4 (is mid point with 210 °),
And be ahead of 18 ° of NTSC#5 (is mid point with 270 °); E.PAL#5 (is mid point with 324 °) lags behind 54 ° of NTSC#5 (is mid point with 210 °), and
Be ahead of 6 ° of NTSC#6 (is mid point with 270 °).
For corresponding image consistency in time can precentagewise be provided, must be with they normalization, also promptly divided by the represented number of degrees scope of the image in the output image system.Under the situation of conversion from PAL to NTSC, the time difference of its corresponding image should be divided by 72 °.Under situation conversely, then divisor should be 60 °.Can draw following numerical value thus: more leading 8.3% than PAL#1 by PAL to NTSC:A.NTSC#1, and lag behind 91.7% than PAL#5a; B.NTSC#2 is more leading 25% than PAL#2, and lags behind 75% than PAL#1; C.NTSC#3 is more leading 41.7% than PAL#3, and lags behind 58.3% than PAL#2; D.NTSC#4 is more leading 58.3% than PAL#4, and lags behind 41.7% than PAL#5; E.NTSC#5 is more leading 75% than PAL#5, and lags behind 25% than PAL#5; F.NTSC#6 is more leading 91.7% than PAL#1b, and lags behind 8.3% than PAL#5.Lag behind 10% to PAL:A.PAL#1 than NTSC#1 by NTSC, and than NTSC#2 leading 90%; B.PAL#2 lags behind 30% than NTSC#2, and than NTSC#3 leading 70%; C.PAL#3 lags behind 50% than NTSC#3, and than NTSC#4 leading 50%; D.PAL#4 lags behind 70% than NTSC#4, and than NTSC#5 leading 30%; E.PAL#5 lags behind 90% than NTSC#5, and than NTSC#6 leading 10%;
These data have shown that image in two width of cloth output image systems has much with the difference of corresponding target image system performance on its time characteristic.They also will contribute for synthetic this image in this ratio.Be easy to thus find out that a little difference between the output image and the new image that will synthesize on express time point will produce a high weighting that trends towards the former.Similarly a big difference will produce a very little weight.The weighted factor of an image has provided the complementary of a percentage of passing by the time between output image and target image.Can draw its center by the above-mentioned formula that provides thus by two original images (j and j+1) of input data needs, and target image (i).
For the conversion that (substantially) sequence of the value (precentagewise) that each image drew is used between PAL and TSC-system formula as follows: generate by 91.7%PAL#1 and 8.3%PAL#5a to NTSC:A.NTSC#1 by PAL; B.NTSC#2 is generated by 75 %PAL# 2 and 25%PAL#1; C.NTSC#3 is generated by 58.3%PAL#3 and 41.7%PAL#2; D.NTSC#4 is generated by 41.7%PAL#4 and 58.3%PAL#3; E.NTSC#5 is generated by 25 %PAL# 5 and 75%PAL#4; F.NTSC#6 is generated by 8.3%PAL#1b and 91.7%PAL#5.Generate by 90 %NTSC# 1 and 10%NTSC#2 to PAL:A.PAL#1 by NTSC; B.PAL#2 is generated by 70 %NTSC# 2 and 30%NTSC#3; C.PAL#3 is generated by 50 %NTSC# 3 and 50%NTSC#5; D.PAL#4 is generated by 30%NTSC#4 and 70%NTSC#3; E.PAL#5 is generated by 10 %NTSC# 5 and 90%NTSC#6.
These percentages are based upon shown in Figure 5, on equitant " rotation is split " schematic diagram.
By the image conversion process that the present invention did, do not cause " jitter " effect, and every width of cloth image of new images system has all shown correct (in the time course) time.
Use above-mentioned mathematical principle to draw, especially be adapted to the relation of image transitions between picture system: draw PAL (as mentioned above) from NTSC, draw NTSC (as mentioned above) from PAL, draw NTSC from Film 16, draw Film 16 from NTSC, draw NTSC, draw Film24 from NTSC from Film 24, draw Film 24 from Film 16, draw Film 16 from Film 24, draw Film 60, draw Film 16 from Film 60 from Film 16, draw Film 60 from Film 24, draw Film 24 from Film 60, draw Film 24, draw PAL from Film 24 from PAL, draw PAL from Film 16, draw Film 16 from PAL, draw Film 60, draw PAL from Film 60 from PAL, draw Film60 from NTSC, draw NTSC from Film 60.Obviously, a professional can draw from above-mentioned principle for similarly transmission between other system, comprising the computer graphical that produces the different images switching rate.
But provide the conversion between the NTSC-PAL that comes out with the conversion method analogy of said system below.
For example can do the conversion between Film 24 and the NTSC.Shown in Figure 6 is that schematic diagram is split in the rotation that an image sequence transmits.Its basic sequence time remaining is 0.0833 second.5 NTSC images and 4 Film 24 images (comprising two substantial images and two " black " images) are arranged in this time interval.The data of NTSC in this image sequence are as follows: A. figure #1=is mid point from 0 ° to 72 ° with 36 °; B. scheming #2=from 72 ° to 144 °, is mid point with 108 °; C. scheming #3=from 144 ° to 216 °, is mid point with 1 80 °; D. scheming #4=from 216 ° to 288 °, is mid point with 252 °; E. scheming #5=from 288 ° to 360 °, is mid point with 324 °.
The data of Film 24 in this image sequence (a=preamble sequence, the follow-up sequence of b=) are as follows: A. figure #3a=is from 180 ° to 270 °, is mid point (promptly-135 °, this is the last piece image of preamble sequence) with 225 °; B. scheming #1=from 0 ° to 90 °, is mid point with 45 °; C. scheming #2=from 90 ° to 180 °, is mid point (one black " image ") with 135 °; D. scheming #3=from 180 ° to 270 °, is mid point with 225 °; E. scheming #4=from 270 ° to 360 °, is mid point (360 °=0 °-one black " figure with 315 °
Picture "); F. scheming #1b=from 0 ° to 90 °, is mid point (first width of cloth image of follow-up sequence) with 45 °.
From Fig. 6, can derive the following relationship between NTSC and the Film24.Film24 is called as Film in following table; (represent with the number of degrees) from Film24 to NTSC: A.NTSC#1 is ahead of Film#19 °, and lags behind 171 ° of Film#3a; B.NTSC#2 is ahead of 117 ° of Film#3, and lags behind 63 ° of Film#1; C.NTSC#3 is ahead of 45 ° of Film#3, and lags behind 135 ° of Film#1; D.NTSC#4 is ahead of 153 ° of Film#1b, and lags behind 27 ° of Film#3; E.NTSC#5 is ahead of 81 ° of Film#1b, and lags behind 99 ° of Film#3.24 (representing with the number of degrees): A.Film#1 lag behind 9 ° of NTSC#1 from NTSC to Film, and are ahead of 63 ° of NTSC#2; B.Film#2 should be one black " image "; C.Film#3 lags behind 45 ° of NTSC#3, and is ahead of 27 ° of NTSC#4; D.Film#4 should be one black " image ".(not needing NTSC#5).
The number of degrees expression way that will reach " in advance " " hysteresis " is converted to the percentage expression mode, then provides following numerical value.Be ahead of Film#1 5% from Film 24 to NTSC:A.NTSC#1, and lag behind Film#3a 95%; B.NTSC#2 is ahead of Film#3 65%, and lags behind Film#1 35%; C.NTSC#3 is ahead of Film#3 25%, and lags behind Film#1 75%; D.NTSC#4 is ahead of Film#1b 85%, and lags behind Film#3 15%; E.NTSC#5 is ahead of Film#1b 45%, and lags behind Film#3 55%.Lag behind NTSC#1 12.5% from NTSC to Film 24:A.Film#1, and be ahead of NTSC#2 87.5%; B.Film#2 should be one black " image "; C.Film#3 lags behind NTSC#3 62.5%, and is ahead of NTSC#4 37.5%; D.Film#4 should be one black " image ".(not needing NTSC#5).
Percentage can draw when changing between NTSC and Film 24 image datas thus, the following end value of each image in the image sequence: from Film 24 to NTSC:A.NTSC#1 is to be made of 95 % Film# 1 and 5% Film#3a; B.NTSC#2 is made of 35% Film#3 and 65% Film#1; C.NTSC#3 is made of 75 % Film# 3 and 25%Film#1; D.NTSC#4 is made of 15% Film#1b and 85% Film#3; E.NTSC#5 is made of 55% Film#1b and 45% Film#3; Be to constitute from NTSC to Film 24:A.Film#1 by 87.5% NTSC#1 and 12.5% NTSC#2; B.Film#2 should be one black " image "; C.Film#3 is made of 37.5% NTSC#3 and 62.5% NTSC#4; D.Film#4 should be black " image " (not needing NTSC#5).
Be used for Film 16 and split figure shown in Fig. 7 (Fig.7) with the overlapping rotation of Film 60 image sequences conversion.The time remaining of basic sequence is 0.0833 second in its conversion.Film 24 basic sequences contain 4 width of cloth images (wherein 2 meaningful in addition 2 width of cloth are black " figures ") and Film 60 sequences have 10 width of cloth images (wherein 5 meaningful other 5 width of cloth are black " figure ").Should consider aforesaid special case at this: though primary image combined sequence that can be littler, i.e. 2 width of cloth Film 24 and 5 width of cloth Film, 60 images.But Film 60 be this means that the picture number that two image sequences that join will expose does not wait, thereby cause the image arrangement different because of sequence.Thus, the picture number of two systems should be doubled.Being that the data of Film 60 in this image sequence: A. schemes #1=from 0 ° to 36 ° below, is mid point with 18 °; B. scheming #2=from 36 ° to 72 °, is mid point (black " image ") with 54 °; C. scheming #3=from 72 ° to 108 °, is mid point with 90 °; D. scheming #4=from 108 ° to 144 °, is mid point (black " image ") with 126 °; E. scheming #5=from 144 ° to 180 °, is mid point with 162 °; F. scheming #6=from 180 ° to 216 °, is mid point (black " image ") with 198 °; G. scheming #7=from 216 ° to 252 °, is mid point with 234 °; H. scheming #8=from 252 ° to 288 °, is mid point (black " image ") with 277 °; I. scheming #9=from 288 ° to 324 °, is mid point with 306 °; J. scheming #10=from 324 ° to 360 °, is mid point with 342 °
(360 °=0 °-Hei " image ").
Being that the data (a=preamble sequence, the follow-up sequence of b=) of Film 24 in this image sequence: A. schemes #3a=from 180 ° to 270 ° below, is that (promptly-135, this is preposition preface to mid point with 225 °
The last piece image of row); B. scheming #1=from 0 ° to 90 °, is mid point with 45 °; C. scheming #2=from 90 ° to 180 °, is mid point (one black " image ") with 135 °; D. scheming #3=from 180 ° to 270 °, is mid point with 225 °; E. scheming #4=from 270 ° to 360 °, is mid point (360 °=0 °-one black " figure with 315 °
Picture "); F. scheming #1B=from 0 ° to 90 °, is mid point (first width of cloth image of follow-up sequence) with 45 °.From Fig. 7, can derive the following relationship between Film 60 and the Film 24: be ahead of 27 ° of Film 24#1 from Film 24 to Film 60:A.Film 60#1, and lag behind 153 ° of Film 24#3a; B.Film 60#2 should be one black " image "; C.Film 60#3 is ahead of 135 ° of Film 24#3, and lags behind 45 ° of Film 24#1; D.Film 60#4 should be one black " image "; E.Film 60#5 is ahead of 63 ° of Film 24#3, and lags behind 117 ° of Film 24#1; F.Film 60#6 should be one black " image "; G.Film 60#7 is ahead of 171 ° of Film 24#1b, and lags behind 9 ° of Film 24#3; H.Film 60#8 should be one black " image "; I.Film 60#9 is ahead of 99 ° of Film 24#1b, and lags behind 81 ° of Film 24#3; H Film 60#10 should be one black " image "; A.Film 24#1 lags behind 27 ° of Film 60#1, and is ahead of 45 ° of Film 60#3; B.Film 24#2 should be one black " image "; C.Film 24#3 lags behind 63 ° of Film 60#5, and is ahead of 9 ° of Film 60#7; D.Film 24#4 should be one black " image ".
(not needing Film 60#9).Number of degrees value is converted to percent value then draws following result: more leading 15% from Film 24 to Film 60:A.Film 60#1 than Film 24#1, and lag behind 85% than Film 24#3a; B.Film 60#2 should be one black " image "; C.Film 60#3 is more leading 75% than Film 24#3, and lags behind 25% than Film 24#1; D.Film 60#4 should be one black " image "; E.Film 60#5 is more leading 35% than Film 24#3, and lags behind 65% than Film 24#1; F.Film 60#6 should be one black " image "; G.Film 60#7 is more leading 95% than Film 24#1, and lags behind 5% than Film 24#3; H.Film 60#8 should be one black " image "; I.Film 60#9 is more leading 55% than Film 24#1b, and lags behind 45% than Film 24#3; J.Film 60#10 should be one black " image ".Lag behind 37.5% from Film 60 to Film 24:A.Film 24#1 than Film 60#1, and than Film 60#3 leading 62.5%; B.Film 24#2 should be one black " image "; C.Film 24#3 lags behind 87.5% than Film 60#5, and than Film 60#7 leading 12.5%; D.Film 24#4 should be one black " image ".
(not needing Film 60#9)
Be used between Film 60 and the Film 24 conversion, as follows for the percent value of the final weighted factor of each figure in the picture system: from Film 24 to Film 60 (percentage): A.Film 60#1 is made of 85% Film 24#1 and 15% Film 24#3a; B.Film 60#2 should be one black " image "; C.Film 60#3 is made of 25% Film 24#3 and 75% Film 24#1; D.Film 60#4 should be one black " image "; E.Film 60#5 is made of 65% Film 24#3 and 35% Film 24#1; F.Film 60#6 should be one black " image "; G.Film 60#7 is made of 5% Film 24#1b and 95% Film 24#3; H.Film 60#8 should be one black " image "; I.Film 60#9 is made of 45% Film 24#1b and 55% Film 24#3; J.Film 60#10 should be one black " image ".From Film 60 to Film 24 (percentage): A.Film 24#1 is made of 62.5% Film 60#1 and 37.5% Film 60#3; B.Film 24#2 should be one black " image "; C.Film 24#3 is made of 12.5% Film 60#5 and 87.5% Film 60#7; D.Film 24#4 should be one black " image ".
(not needing Film 60#9)
Fig. 8 has provided the overlapping rotation that is used for Film 16 and the conversion of Film 60 image sequences and has split figure.The time remaining of basic sequence is 0.25 second in this conversion.8 width of cloth images (wherein 4 meaningful in addition 4 width of cloth are black " figures ") are arranged in Film 16 sequences and Film 60 sequences have 30 width of cloth images (wherein 15 meaningful other 15 width of cloth are black " figure ").Also should follow the principle that the picture number with two figure basic sequences doubles at this, so that make the primary image sequence have constant exposure frame number.List the percent value that Film 16 and the conversion of Film 60 image sequences are drawn below: from Film 16 to Film 60 (percentage): A.Film 60#1 is by 81.7% Film 16#1 and 18.3% Film 16#7a structure
Become; B.Film 60#2 should be one black " image "; C.Film 60#3 is made of 8.3% Film 16#3 and 91.7% Film 16#1
; D.Film 60#4 should be one black " image "; E.Film 60#5 is made of 35% Film 16#3 and 65% Film 16#1; F.Film 60#6 should be one black " image "; G.Film 60#7 is made of 61.7% Film 16#3 and 38.3% Film 16#1
; H.Film 60#8 should be one black " image "; I.Film 60#9 is made of 88.3% Film 16#3 and 11.7% Film 16#1
; J.Film 60#10 should be one black " image "; K.Film 60#11 is made of 15% Film 16#5 and 85% Film 16#3; L.Film 60#12 should be one black " image "; M.Film 60#13 is by 41.7% Film 16#5 and 58.3% Film 16#3 structure
Become; N.Film 60#14 should be one black " image "; O.Film 60#15 is by 68.3% Film 16#5 and 3 1.7% Film 16#3 structure
Become; P.Film 60#16 should be one black " image "; Q.Film 60#17 is made of 95% Film 16#5 and 5% Film 16#3; R.Film 60#18 should be one black " image "; S.Film 60#19 is by 21.7% Film 16#7 and 78.3% Film 16#5 structure
Become; T.Film 60#20 should be one black " image "; U.Film 60#21 is by 48.3% Film 16#7 and 51.7% Film 16#5 structure
Become; V.Film 60#22 should be one black " image "; W.Film 60#23 is made of 75% Film 16#7 and 25% Film 16#5; X.Film 60#24 should be one black " image "; Y.Film 60#25 is by 1.7% Film 16#1b and 98.3% Film 16#7 structure
Become; Z.Film 60#26 is a width of cloth black " image "; AA.Film 60#27 is by 28.3% Film 16#1b and 71.7% Film 16#7
Constitute; BB.Film 60#28 is a width of cloth black " image "; CC.Film 60#29 is made of 55% Film 16#1b and 45% Film 16#7
; DD.Film 60#30 is a width of cloth black " image ".Film 16 to Film 60 (precentagewise meter): A.Film 16#1 is by 31.25% Film 60#1 and 68.75% Film 60#3 structure
Become; B.Film 16#2 is a width of cloth black " image ", (not need Film 60#5 and Film 60#7); C.Film 16#3 is by 56.25% Film 60#9 and 43.75% Film 60#11
Constitute; D.Film 16#4 is a width of cloth black " image ", (Film 60#13 and Film 60#15 not need); E.Film 16#5 is by 81.25% Film 60#17 and 18.75% Film 60#19
Constitute; F.Film 16#6 is a width of cloth black " image ", (Film 60#21 not need); G.Film 16#7 is by 6.25% Film 60#23 and 93.75% Film 60#25
Constitute; H.Film 16#8 is a width of cloth black " image ", (Film 60#27 and Film 60#29 not need).
As can be seen from Figure 8, the some width of cloth images in Film 16 basic sequences time scale (digital watch) go up and Film 60 primary image sequences in three width of cloth doublings of the image, for example Film 16#7 and Film 60#23, #25 and #27.
This example and similar situation, as NTSC to Film 16 or PAL to the conversion of Film 16, also belong to category of the present invention.Promptly synthesize piece image in the fresh target image sequence with two adjacent images of above time in the original sequence, weighted factor wherein also is associated to the distance that produces the image mid point with used image time interval mid point.Corresponding device thereof (herein to the description of reference Figure 16 with Figure 17) also should be able to be finished in design to three width of cloth in the original image or more images content mixing to a width of cloth new images.
Fig. 9 has provided the rotation that is used for image transitions between Film 60 (" Showscan ") and PAL system and has split figure.Sequence time shown in Figure 12 is continuously 0.1 second.This is corresponding to 5 width of cloth PAL and 12 width of cloth Film, 60 images (wherein 6 the meaningful images and 6 width of cloth black " image ") are arranged in the sequence.
Here also be immediately from every sequence picture number draw the master data that is used for calculating with image-related weighted factor, i.e. time interval and central point thereof and illustrate their relativeness clear in the drawings.Thus again can be according to the general description of front with PAL-NTSC, NTSC-PAL etc. are that the explanation of example has drawn the time relationship between output and target image sequence.
Listed below when 60 conversions of PAL and Film, be used for effective numerical value of each single image weighted factor of composing images sequence: be to constitute from PAL to Film 60:A.Film 60#1 by 70.8% PAL#l and 29.2% PAL#5a; B.Film 60#2 is a width of cloth black " image "; C.Film 60#3 is made of 54.2% PAL#2 and 45.8% PAL#1; D.Film 60#4 is a width of cloth black " image "; E.Film 60#5 is made of 37.5% PAL#3 and 62.5% PAL#2; F Film 60#6 is a width of cloth black " image "; G.Film 60#7 is made of 20.8% PAL#4 and 79.2% PAL#3; H.Film 60#8 is a width of cloth black " image "; I.Film 60#9 is made of 4.2% PAL#5 and 95.8% PAL#4; J.Film 60#10 is a width of cloth black " image "; K.Film 60#11 is made of 87.5% PAL#5 and 12.5% PAL#4; L.Film 60#12 is a width of cloth black " image ".From Film 60 to PAL:A.PAL#1 is to be made of 65% Film 60#1 and 35% Film 60#3; B.PAL#2 is made of 45% Film 60#3 and 55% Film 60#5; C.PAL#3 is made of 25% Film 60#5 and 35% Film 60#7; D.PAL#4 is made of 5% Film 60#7 and 95% Film 60#9; E.PAL#5 is made of 85% Film 60#11 and 15% Film 60#1b.
Figure 10 is that figure is split in the rotation that is used for image transitions between Film 24 and Film 16 systems.Its basic sequence time remaining is 0.125 second.The basic sequence of Film 24 contain 6 width of cloth images (wherein 3 meaningful, and in addition 3 width of cloth are black " figure ").The basic sequence of Film 16 contain 4 width of cloth images (wherein 2 meaningful, and in addition 2 width of cloth are black " figure ").
The central point of the corresponding time interval that is used for changing and they also is that the picture number by image sequence draws and can reads from Figure 10 fully.Thus, the following composition rule of every width of cloth image in the time of can providing the conversion of 16 of Film 24 and Film: from Film 16 to Film 24:A.Film 24#1 is to be made of 91.7% Film 16#1 and 8.3% Film 16#3a; B.Film 24#2 is a width of cloth black " image "; C.Film 24#3 is made of 58.3% Film 16#3 and 41.7% Film 16#1; D.Film 24#4 is a width of cloth black " image "; E.Film 24#5 is made of 25% Film 16#1b and 75% Film 16#3; F.Film 24#6 is a width of cloth black " image ".From Film 24 to Film 16A.Film 16#1 is to be made of 87.5% Film 24#1 and 12.5% Film 24#3; B.Film 16#2 is a width of cloth black " image "; C.Film 16#3 is made of 37.5% Film 24#3 and 62.5% Film 24#5; D.Film 16#4 is a width of cloth black " image ".
Figure 11 has provided the rotation that is used for image transitions between NTSC and Film 16 picture systems and has split figure.Its basic sequence time remaining is 0.25 second at this.Contain 15 width of cloth NTSC images in the NTSC basic sequence and the basic sequence of Film 16 contains 8 width of cloth images (4 meaningful image and 4 width of cloth are black " figure ").Image time interval on time shaft (reach herein all rotations split be annular among the figure), the position of its central point and the relation between them also all are that the picture number by image sequence draws and can draw from figure fully.
According to above-mentioned rule and the detailed example of analogy given, the following composition rule of every width of cloth image in the time of can drawing 16 of NTSC and Film conversion: from Film 16 to NTSC:A.NTSC#1 is to be made of 88.3% Film 16#1 and 11.7% Film 16#7a; B.NTSC#2 is made of 15% Film 16#3 and 85% Film 16#1; C.NTSC#3 is made of 41.7% Film 16#3 and 58.3% Film 16#1; D.NTSC#4 is made of 68.3% Film 16#3 and 31.7% Film 16#1; E.NTSC#5 is made of 95% Film 16#3 and 5% Film 16#1; F.NTSC#6 is made of 21.7% Film 16#5 and 78.3% Film 16#3; G.NTSC#7 is made of 48.3% Film 16#5 and 51.7% Film 16#3; H.NTSC#8 is made of 75% Film 16#5 and 25% Film 16#3; I.NTSC#9 is made of 1.7% Film 16#7 and 98.3% Film 16#5; J.NTSC#10 is made of 38.3% Film 16#7 and 71.7% Film 16#5; K.NTSC#11 is made of 55% Film 16#7 and 45% Film 16#5; L.NTSC#12 is made of 81.7% Film 16#7 and 18.3% Film 16#5; M.NTSC#13 is made of 8.3% Film 16#1b and 91.7% Film 16#7; N.NTSC#14 is made of 35% Film 16#1b and 65% Film 16#7; O.NTSC#15 is made of 61.7% Film 16#1b and 38.3% Film 16#7
.Be to constitute from NTSC to Film 16:A.Film 16#1 by 56.25% NTSC#1 and 43.75% NTSC#2; B.Film 16#2 is a width of cloth black " image ", (NTSC#3 and NTSC#4 not need); C.Film 16#3 is made of 81.25% NTSC#5 and 18.75% NTSC#6; D.Film 16#4 is a width of cloth black " image ", (NTSC#7 not need); E.Film 16#5 is made of 6.25% NTSC#8 and 93.75% NTSC#9; F.Film 16#6 is a width of cloth black " image ", (NTSC#10 and NTSC#11 not need); G.Film 16#7 is made of 31.25% NTSC#12 and 68.75% NTSC#13
; H.Film 16#8 is a width of cloth black " image ", (NTSC#14 and NTSC#15 not need);
The rotation that is used for image transitions between PAL and Film 16 picture systems is split figure as shown in figure 12.Its basic sequence time remaining is 0.5 second at this.In this basic sequence, contain 25 width of cloth PAL images and 16 width of cloth Film, 16 images (8 meaningful images wherein, and 8 width of cloth are black " images " in addition).
Also to be above-mentioned method obtain from the picture number of image sequence (selected graphic technique complete sequence always corresponding 360 °) and can read from figure with the data of its central point and their time relationship about two system diagram pictures are interval.The composition rule of every width of cloth image of basic sequence is as follows during 16 of PAL and Film conversion: from Film 16 to PAL:A.PAL#1 is to be made of 91% Film 16#1 and 9% Film 16#15a; B.PAL#2 is made of 23% Film 16#3 and 77% Film 16#1; C.PAL#3 is made of 55% Film 16#3 and 45% Film 16#1; D.PAL#4 is made of 87% Film 16#3 and 13% Film 16#1; E.PAL#5 is made of 19% Film 16#5 and 81% Film 16#3; F.PAL#6 is made of 51% Film 16#5 and 49% Film 16#3; G.PAL#7 is made of 83% Film 16#5 and 17% Film 16#3; H.PAL#8 is made of 15% Film 16#7 and 85% Film 16#5; I.PAL#9 is made of 47% Film 16#7 and 53% Film 16#5; J.PAL#10 is made of 79% Film 16#7 and 21% Film 16#5; K.PAL#11 is made of 11% Film 16#9 and 89% Film 16#7; L.PAL#12 is made of 43% Film 16#9 and 57% Film 16#7; M.PAL#13 is made of 75% Film 16#9 and 25% Film 16#7; N.PAL#14 is made of 7% Film 16#11 and 93% Film 16#9; O.PAL#15 is made of 39% Film 16#11 and 61% Film 16#9; P.PAL#16 is made of 71% Film 16#11 and 29% Film 16#9; Q.PAL#17 is made of 3% Film 16#13 and 97% Film 16#11; R.PAL#18 is made of 35% Film 16#13 and 65% Film 16#11; S.PAL#19 is made of 67% Film 16#13 and 33% Film 16#11; T.PAL#20 is made of 99% Film 16#13 and 1% Film 16#11; U.PAL#21 is made of 31% Film 16#15 and 69% Film 16#13; V.PAL#22 is made of 63% Film 16#15 and 37% Film 16#13; W.PAL#23 is made of 95% Film 16#15 and 5% Film 16#13; X.PAL#24 is made of 27% Film 16#1b and 73% Film 16#15; Y.PAL#25 is made of 59% Film 16#1b and 41% Film 16#15.Be to constitute from PAL to Film16:A.Film 16#1 by 71.9% PAL#1 and 28.1% PAL#2; B.Film 16#2 is a width of cloth black " image ", (PAL#3 not need); C.Film 16#3 is made of 59.4% PAL#4 and 40.6% PAL#5; D.Film 16#4 is a width of cloth black " image ", (PAL#6 not need); E.Film 16#5 is made of 46.9% PAL#7 and 53.1% PAL#8; F.Film 16#6 is a width of cloth black " image ", (PAL#9 not need); G.Film 16#7 is made of 34.4% PAL#10 and 65.6% PAL#11; H.Film 16#8 is a width of cloth black " image ", (PAL#12 not need); I.Film 16#9 is made of 21.9% PAL#13 and 78.1% PAL#14; J.Film 16#10 is a width of cloth black " image ", (PAL#15 not need); K.Film 16#11 is made of 59.4% PAL#16 and 90.6% PAL#17; L.Film 16#12 is a width of cloth black " image ", (PAL#18 and PAL#9 not need); M.Film 16#13 is made of 96.9% PAL#20 and 3.1% PAL#21; N.Film 16#14 is a width of cloth black " image ", (PAL#22 not need); O.Film 16#15 is made of 84.4% PAL#23 and 15.6% PAL#24; P.Film 16#16 is a width of cloth black " image ", (PAL#25 not need).
Figure 13 has provided the rotation that is used for image transitions between PAL and Film 24 picture systems and has split figure.Its basic sequence time remaining is 0.5 second.Contain 25 width of cloth PAL images and 24 width of cloth Film, 24 images (wherein 12 the meaningful images and 12 width of cloth are black " images ") in the basic sequence.
Time interval and affiliated central point are drawn by picture number in the image sequence equally.Represent to find out corresponding time relationship between the adjacent and overlaid image of the time in the single image and output image sequence of to produce in the image sequence according to the respective graphical among Figure 13.The composition rule of every width of cloth image of basic sequence is as follows during 24 of PAL and Film conversion: from Film 24 to PAL (representing with percentage): A.PAL#1 is to be made of 99% Film 24#1 and 1% Film 24#23a; B.PAL#2 is made of 47% Film 24#3 and 53% Film 24#1; C.PAL#3 is made of 95% Film 24#3 and 5% Film 24#1; D.PAL#4 is made of 43% Film 24#5 and 57% Film 24#3; E.PAL#5 is made of 91% Film 24#5 and 9% Film 24#3; F.PAL#6 is made of 39% Film 24#7 and 61% Film 24#5; G.PAL#7 is made of 87% Film 24#7 and 13% Film 24#5; H.PAL#8 is made of 35% Film 24#9 and 65% Film 24#7; I.PAL#9 is made of 83% Film 24#9 and 17% Film 24#7; J.PAL#10 is made of 31% Film 24#11 and 69% Film 24#9; K.PAL#11 is made of 79% Film 24#11 and 21% Film 24#9; L.PAL#12 is made of 27% Film 24#13 and 73% Film 24#11; M.PAL#13 is made of 75% Film 24#13 and 25% Film 24#11; N.PAL#14 is made of 23% Film 24#15 and 77% Film 24#13; O.PAL#15 is made of 71% Film 24#15 and 29% Film 24#13; P.PAL#16 is made of 19% Film 24#17 and 81% Film 24#15; Q.PAL#17 is made of 67% Film 24#17 and 33% Film 24#15; R.PAL#18 is made of 15% Film 24#19 and 85% Film 24#17; S.PAL#19 is made of 63% Film 24#19 and 37% Film 24#17; T.PAL#20 is made of 11% Film 24#21 and 89% Film 24#19; U.PAL#21 is made of 59% Film 24#21 and 41% Film 24#19; V.PAL#22 is made of 7% Film 24#23 and 93% Film 24#21; W.PAL#23 is made of 55% Film 24#23 and 45% Film 24#21; X.PAL#24 is made of 3% Film 24#1b and 97% Film 24#23; Y.PAL#25 is made of 51% Film 24#1b and 49% Film 24#23.24 (percentage is represented) from PAL to Film: A.Film 24#1 is made of 97.9% PAL#1 and 2.1% PAL#2; B.Film 24#2 is a width of cloth black " image "; C.Film 24#3 is made of 89.6% PAL#3 and 10.4% PAL#4; D.Film 24#4 is a width of cloth black " image "; E.Film 24#5 is made of 81.2% PAL#5 and 18.8% PAL#6; F.Film 24#6 is a width of cloth black " image "; G.Film 24#7 is made of 72.9% PAL#7 and 27.1% PAL#8; H.Film 24#8 is a width of cloth black " image "; I.Film 24#9 is made of 64.6% PAL#9 and 35.4% PAL#10; J.Film 24#10 is a width of cloth black " image "; K.Film 24#11 is made of 56.2% PAL#11 and 43.8% PAL#12; L.Film 24#12 is a width of cloth black " image "; M.Film 24#13 is made of 47.9% PAL#1 and 52.1% PAL#14; N.Film 24#14 is a width of cloth black " image "; O.Film 24#15 is made of 39.6% PAL#15 and 60.4% PAL#16; P.Film 24#16 is a width of cloth black " image "; Q.Film 24#17 is made of 31.3% PAL#17 and 68.7% PAL#18; R.Film 24#18 is a width of cloth black " image "; S.Film 24#19 is made of 22.9% PAL#19 and 77.1% PAL#20; T.Film 24#20 is a width of cloth black " image "; U.Film 24#21 is made of 14.6% PAL#21 and 85.4% PAL#22; V.Film 24#22 is a width of cloth black " image "; W.Film 24#23 is made of 6.3% PAL#23 and 93.7% PAL#24; And X.Film 24#24 is a width of cloth black " image ", (PAL#25 not need).
Figure 14 has provided the combination rotation that is used for image transitions between NTSC and Film 60 picture systems and has split figure.Contain 1 width of cloth NTSC image (territory) and 2 width of cloth Film, 60 images (frame) in the basic sequence, wherein the latter alternately has content images and black " figure ".Image siding-to-siding block length, the time relationship of center position and two primary image sequences can be directly drawn and can be found out easily from Figure 14 by the combination of picture number in the primary image sequence.
Can draw each image rule in the following respective objects sequence in generating by corresponding output sequence thus: form by 75%Film 60#1 and 25%Film 60#1b from Film 60 to NTSC:NTSN#1.Form by the NTSC#1a of 75 %NTSN# 1 and 25% to Film 60:A.Film 60#1 from NTSC.B.Film 60#2 is the black figure of a width of cloth.
At this, these two primary image sequences start simultaneously with combination " rotation is split " figure that the basis is calculated in conduct.Though this is special circumstances of image relative distribution, yet be to be worth adopting, because have only the synchronism of beginning before and after conversion that could guarantee film in this case.Certainly also should allow the slight variations of primary image sequence relative distribution, so that realize optimization to certain particular conversion situation.Skew between consequent two primary image sequences is also corresponding caused new picture centre and the weighted factor that changed (corresponding relation between the image can not become in the corresponding output image sequence but goal systems first width of cloth image and time go up.It is to be got by the image sequence that begins synchronously).The picture centre of being given in formula is tried to achieve by combination " rotation is split " figure, and has considered the skew between two primary image sequences among the figure, and then above-mentioned given weighted factor computing formula can remain unchanged.Figure 15 has provided the legend of NTSC-PAL conversion.
Flexible method is to keep the above-mentioned picture centre that draws and obtain weighted factor by the computing formula of having considered skew.The formula of this moment is:
C j，s＝1-|t j-(t i+d)|/(t j+1-t j)
C J+1, s=1-|t J+1-(t i+ d) |/(t J+1-t j) wherein, d=skew, increment.
Displacement d is defined herein as the starting point passing of this point of its output system relatively of the primary image sequence of goal systems.The positive number symbolic representation is offset clockwise.According to this, the skew relative and NTSC primary image sequence of PAL primary image sequence is+24 ° among Figure 15.This class skew should not caused the change of output image sequence image distribution itself by first width of cloth image in the primary image sequence that will produce.Therefore the skew of first width of cloth picture centre only limits in the defined scope in center of the output image system diagram picture that is associated on the time by two width of cloth in the goal systems.Thus, for change over condition shown in Figure 15, skew should be limited in-6 ° to+54 ° scope.
The given in the manner described above ratio of image in the original image system is transferred in the new picture system goes.Transmission between two picture systems comprises two steps: to preparation and the actual image transitions of wanting image transmitted.
To the preparation of wanting data transmission is that the visual information that is lacked in the data is rebuild." interlacing scan " television image is necessary that the part that will be lacked earlier rebuilds before conversion, because the image that every width of cloth television image all is " interlacing " (palisade) in black line.To this, there is known reconstruction to lack the method for image, it is applied to drawing static (" freeze frame ") image (also the television image territory being called television image in present patent application) from the television image sequence.
For the image transitions step, have two kinds of conversion methods at least: a kind of is the double exposure method.From a kind of image sequence to another kind of picture system during conversion, two adjacent images in the original image system can be used optical printer to beat on the picture in the new images system with different exposure (with respect to the time for exposure or the intensity of above-mentioned weighted factor of being given and ratio).
From a kind of television image system when another television image system does transmission, can will be mixed into a width of cloth picture in the new images system with different electronic signal amplitudes (with respect to above-mentioned computational methods of being given and ratio) from the television image of two adjacent reconstructions of television image system output.
Figure 16 has provided a functional unit schematic block diagram that works in the conversion equipment 100 of digital form.It is used for from one with the first image exchange frequency f sBe recorded in the first recording medium M sOn the first image sequence S to the second image exchange frequency f TBe recorded in the second recording medium M TOn the conversion of the second image sequence T.
Set with a ntsc television video camera 101 at this and on video tape, to produce telerecording S with the image rate of per second nominal value 30 (reality 60) width of cloth image.It is fed to conversion equipment 100.
This conversion equipment comprises a same reproducing device (first video tape recorder) 102 by the work of TSC-system formula, it has a control member 102a, a central markers 103, the control input end of its output and remaining member (108 and 109 and video camera 101 exception, also has reproducing device 112, it does not belong to conversion equipment 100) be connected, the second serial full images memory 106 that the first serial full images memory 105 that 104, one inputs of a CPU (microcontroller) are connected with television camera 102 data output ends and an input are connected with first memory 105 data output ends.This conversion equipment has also further comprised one and has been connected with the output of memory 105 and 106 data output ends and CPU 104, have two discrete controlled path 107a, 107b's, be exclusively used in this digital mixing arrangement 107, the 3rd serial full images memory 110 that input is connected in the mixing arrangement data output end, an input keyboard 108 that is connected with the processing unit input, display 109 that is connected with the processing unit output and one have control member 111a's, press the recording and/or reproducing apparatus (second video tape recorder) 111 of pal mode work.
One dish becomes the video tape of target image sequence T with per second nominal value 25 (actual 50) width of cloth picture recording with this video tape recorder 111 under pal mode, after providing through conversion equipment 100, and finally can be for 112 projections of PAL reproducing device.
Elder generation is with video tape M before transfer process begins sPut into video tape recorder 102, the program of corresponding aforementioned NTSC-PAL conversion is set and adjusts Control Parameter (as side-play amount etc.) according to circumstances by input keyboard 108.Then starting markers 103 and (by it synchronously) then starts video tape recorder 102 and resets and start video tape recorder 111 and record.
The playback work of video tape recorder 102 is to be controlled by processing unit 104 (externally under the acting in conjunction of clock 103).It is operated in and allows and processing unit 104, video memory 105,106,110 together single width handle under the state of image.Be subject to processing unit 104 equally and control, feed a width of cloth (the j width of cloth) image to the first full images memory 105 from video tape recorder 102 at every turn.It is transferred to the second full images memory when the sequence of source images image ([j+1] width of cloth) that is next arrived is presented in video tape recorder 102.Meanwhile, (j+1) width of cloth image is deposited in memory 105.Like this, on each time point all with image among the file layout that the is applicable to digital processing two width of cloth sources video recording all set S, in order to synthetic i width of cloth target image.
After two memories 105 and 106 were deposited figure and finished, processing unit 104 was clapped the weight coefficient C that calculates the j width of cloth and (j+1) width of cloth image under the TSC-system formula from j under the clock signal control of markers 103 J, sAnd C J+1, s, be used for the i width of cloth image of synthetic pal mode.The conversion program of finishing this step is to place memory 104a with the form of showing.Meanwhile, processing unit has also correspondingly been adjusted the signal amplitude value of path 10 7a and 107b.Then, the content in the memory 105 and 106 be weighted with adjusted amplitude and in mixing arrangement 107 by the pixel addition.Result of calculation is deposited in the 3rd full images memory 110.Here, the i width of cloth image in the He Cheng PAL image sequence is recorded by second video tape recorder 111 by the cadence signal of markers 103 in this way.
Synthetic image can be observed on display 109.If necessary, also can carry out manual intervention by 108 pairs of transfer processes of keyboard.
Described treatment progress constantly repeats, and has recorded again to extremely with Phase Alternation Line system until the whole image sequence that provides with the TSC-system formula (an one " film ").Rightabout image transitions PAL-NTSC, process similarly.Similarly, considering specially, under the prerequisite of the conversion parameter of being determined by program, the conversion between the computer picture mode also is like this.
Figure 17 is the functional unit schematic block diagram of part by the conversion equipment 200 of traditional photography technical method work.It is used for from one with the first image exchange frequency f S "And I between corresponding image area S 'Be recorded in the first recording medium F S 'On the first image sequence S ' to the second image exchange frequency f TAnd I between corresponding image area TBe recorded in the second recording medium F TOn the conversion of the second image sequence T '.
In this hypothesis with a Film 16 video cameras 201 with the image rate of per second 16 width of cloth images at sheet band F S 'Last generation record film S '.It is fed to conversion equipment 200.
Similar to the composition form of the above device 100, this conversion equipment has comprised a markers 202, a CPU (microcontroller) 203, a conversion table memory 203a, an input keyboard 204 that is connected with processing unit 203 inputs and a display 205 that is connected with processing unit 203 outputs.Can control and follow the tracks of transfer process by them.Conversion equipment 200 has also further comprised the camera work of a special use and has mixed copy and double exposure device 206.Its two control input ends are connected with the output of CPU 203.It has last discrete controlled exposure device 206a of two times, 206b, and one is used for former film F S 'Conveyer 206c and one be used to copy film F TConveyer 206d.
One dish is mixing in the copy device like this, for example uses Film 24 standards of per second 24 width of cloth images, and target image sequence T ' is photographed film F TOn, after providing through conversion equipment 200, finally can be for 207 projections of a Film 24 projectors.
Before a transfer process begins, earlier former band Fs ' put into and mix copy device 207, by the program of input keyboard 204 set handling unit 203 corresponding to this transfer process (with respect to the conversion of the Film 16 to Film 24 that describes in detail previously).Then start the conveyer 206c and the 206d that treat exposed plate that markers 103 and (synchronous by it) are used for former film and meet the Film24 standard then.
By the control of sending sheet, make and mix the processing that copy device 206 can be made camera work to single image two films.Thereby, can make former film F S 'In two width of cloth adjacent image B J, s 'And B J+1, s 'Can expose to treating exposure image sequence image B I, TOn.This mixing copy device can make two images in the adjacent former film all lay respectively at any time among exposure device 206 and the 206b textural.They are connected with a binary channels time for exposure controller 206e.Time controller links to each other with the output of processing unit 203.
Under the control of the clock signal of markers 103, processing unit 203 is clapped to call from j at every turn and is stored in the memory 203a, is used for synthesizing the i width of cloth image of Film 24 standards, the intensity weighted coefficient C of the j width of cloth and (j+1) width of cloth image under Film 16 standards J, s 'And C J+1, s 'And the time for exposure of corresponding calculation exposure device 206a and 206b.Wherein numerical value j places the counter of conveyer 206c or 206d to provide by an internal counter (not marking among the figure) or one.Then, by time for exposure controller 206e and exposure device 206a and 206b the i image of target sequence T ' is exposed and the sheet band is forwarded.
Described treatment progress constantly repeats, and has taken the photograph again to extremely with the Film24 standard until the whole image sequence that provides with Film 16 standards.Rightabout image conversion process similarly.Similarly, considering specially, under the prerequisite of the conversion parameter of being determined by program, the conversion between the computer picture mode also is like this.The same also conversion between Film 24 and the Film 60 in addition.
In described device, also can add a telerecording device that is used to record composograph in addition.Can be used for mixed process is carried out selectable manual intervention by the playback of display 205 and the control of keyboard.
Exposure device also can be designed to the device of a controllable intensity, and weighted factor is by weakening light corresponding to its numerical value to realize in this device.
For as Figure 16 and device shown in Figure 17, its processing unit and affiliated ancillary equipment can be one special, target when under a stable condition, connecing, microprocessor-controlIed member is as a personal computer.
Handling and/or generating the image sequence (comprising video recording-film, the conversion of film-video recording and film-film) that is provided on the film material also is feasible according to similar device shown in Figure 17 by digital form.For this reason, look concrete condition input have additional an A/D image converter (class such as cinematographic projector and video tape recorder or for recording the projected image CCD-array of face of will record a video) as for the perspective plane also/or set up one for recording the motion picture technique fore device of composograph sequence, such as (resolution is high as far as possible) display at output.For the application that is associated with the D/A conversion, can adopt special (also known) digital control laser and holographic exposure device and optical printer according to circumstances to this type of application.

Claims (15)

1. image conversion method, be used for producing second image sequence (target image sequence) that has between the second image exchange frequency and corresponding second image area, this second image sequence (target image sequence) is to get from have first image sequence between the first image exchange frequency and corresponding first image area, consequent image sequence is suitable to the visual effect of observer with the time dimension incident, in target image sequence, having a part of image at least is to obtain with the intensity weighted stack one by one by adjacent image of time in minimum two width of cloth sequence of source images, being characterized as of this method:
The all images of target image sequence all is the images by at least two width of cloth sequence of source images, calculates its intensity weighted factor respectively earlier, and weighted superposition obtains again.
2. according to the method for claim 1, it is characterized by: piece image residing position in time course is to be represented by the position of a time interval characteristic point, according to this, image occurs in image sequence with given when the time comes transmission or playback standard at transmission or playback time.
3. according to the method for claim 2, it is characterized by: characteristic point is the mid point of time interval.
4. according to the method for claim 2 or 3, it is characterized by:
Determined source images and the common zero-time and the time course of the target image sequence that will produce by a markers;
Try to achieve in time course the position of each width of cloth image in the source images and the target image sequence that will produce;
For the image that each width of cloth in the target image sequence will produce, its position in time course is to try to achieve by some width of cloth images contiguous in the sequence of source images, especially by two contiguous width of cloth images;
The time gap or the skew of image time interval are tried to achieve according to their characteristic points in adjacent image in the sequence of source images and the target image sequence that will produce;
It is specific all to calculate an image for each width of cloth image in the adjacent image in the sequence of source images, normalization the intensity weighted factor, it with sequence of source images in image time interval image time interval in the target image sequence that will produce time gap or be offset relevant.And the image in the target image sequence that will produce multiply by after the corresponding intensity weighted factor superimposed again and generate by adjacent image in the sequence of source images.
5. according to the method for one of aforementioned claim, it is characterized by: sequence of source images is stored on a kind of recording medium, and/or target image sequence is stored on a kind of recording medium.
6. according to the method for one of aforementioned claim, it is characterized by: the stack of image is to carry out pixel phase Calais by the image that provides with digital form to finish, and wherein, the amplitude of sum signal is adjusted according to the corresponding strength weighted factor.
7. according to the method for claim 5 and 6, it is characterized by: to analog form, especially the recorded material of obtaining by mechanical camera style has been carried out digitlization earlier before doing stack.
8. according to the method for one of claim 1 to 5, it is characterized by: image overlay is by the width of cloth picture on the recording medium is finished with the adjacent image multiexposure, multiple exposure of sequence of source images.Wherein exposure intensity and time for exposure are adjusted by the corresponding strength weighted factor.
9. according to the method for one of claim 2 to 8, it is characterized by: be positioned at characteristic point t among the target image sequence T iThe new synthetic i width of cloth image of time interval is to have characteristic time point t by two width of cloth adjacent with this time point jAnd t J+1Sequence of source images S in image F J, s(t j) and F J+1, S (t J+1) produce by following relation:
F i，T(t i)＝C j，s×F j，s(t j)+C j+1，s×F j+1，s(t j+1)
Wherein, i and j are integers, C J, s, C J+1, sBe the intensity weighted coefficient of j and j+1 source images respectively, and be calculated as follows into the normalization weighted factor:
C j，s＝1-(|t i-t j|)/(t j+1-t j)
C j+1，s＝1-(|t j+1-t i|)/(t j+1-t j)。
10. according to the method for one of aforementioned claim, it is characterized by: second image sequence produces with the same manner by one at every turn, have same length basic sequence constituted, its time length is to determine with the as far as possible little integer width of cloth image that can comprise first and second image sequences, be used for producing the weighted factor of every width of cloth image in second image sequence, be according to calculating between the first and second image exchange frequencies and first and second image areas, wherein should be taken into account between " deceiving " image area that is present in first and/or second image sequence and/or the situation of interlaced picture.
11. image conversion apparatus, be used for producing second image sequence (target image sequence) that has between the second image exchange frequency and corresponding second image area, this second image sequence (target image sequence) is to get from have first image sequence between the first image exchange frequency and corresponding first image area, consequent image sequence is suitable to the visual effect of observer with the time dimension incident, in target image sequence, having a part of image at least is to obtain with the intensity weighted stack one by one by adjacent image of time in minimum two width of cloth sequence of source images, and this device has:
One is used for the transmission or the replay device that source images are transmitted or resets with between the first image exchange frequency and first image area;
One with the cause image in the aliasing sequence of source images and form mixing arrangement as the vision-mix of image in the target image sequence selectively;
One is used for recording and recording or transmitting device of transmitting with the target image sequence of output result between the second image exchange frequency and second image area as mixing arrangement;
One is used for transmission and resets the markers and the synchronizer of the Synchronization Control of mixing and recording arrangement.It is characterized by:
This device is provided with a computing unit that is connected with synchronizer with markers, be used for calculating with the image-related intensity weighted factor and by the signal that markers and synchronizer provide and carry out the stack of sequence of source images image, it also is provided with one and has the mixing arrangement that can regulate its controller to the weighted factor in each mixed process one by one, and this controller is connected with the output of aforementioned calculation unit.
12. device according to claim 11, it is characterized by: in order to keep at least one width of cloth in two width of cloth images in the target image sequence at every turn, at least should be provided with an electronic image memory, its data output end is connected with mixing arrangement, and this mixing arrangement is to establish with the picture material pixel addition that digital form provides for realizing.
13. the device according to claim 11 or 12 is characterized by: mixing arrangement is connected to an image to analogue recording with the electronic image memory and carries out digitized digitlization fore device.
14. the device according to claim 11 is characterized by:
This mixing arrangement is designed to have the optics copy device of two discrete controlled exposing units.Its control input end is connected with the output of computing unit, the adjacent image that is recorded in the sequence of source images on first recording medium is imported by copy device, and is exposed on second recording medium by exposure intensity and/or the time that corresponding weighting factor is regulated.
15. the device according to claim 11 to 13 is characterized by: the electronic image mixing arrangement intends connecting optical printer.
CN 94195223 1994-12-06 1994-12-06 Picture transfer process and device Pending CN1174645A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 94195223 CN1174645A (en) 1994-12-06 1994-12-06 Picture transfer process and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 94195223 CN1174645A (en) 1994-12-06 1994-12-06 Picture transfer process and device

Publications (1)

Publication Number Publication Date
CN1174645A true CN1174645A (en) 1998-02-25

Family

ID=5039717

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 94195223 Pending CN1174645A (en) 1994-12-06 1994-12-06 Picture transfer process and device

Country Status (1)

Country Link
CN (1) CN1174645A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1960490B (en) * 2005-11-04 2010-08-18 腾讯科技（深圳）有限公司 Method for converting GIF file to SWF file
CN102111542A (en) * 2009-12-28 2011-06-29 索尼公司 Image processing apparatus, imaging apparatus, image processing method, and program
CN1689325B (en) * 2002-10-01 2012-05-30 汤姆森许可贸易公司 Implicit weighting of reference pictures in a video encoder

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1689325B (en) * 2002-10-01 2012-05-30 汤姆森许可贸易公司 Implicit weighting of reference pictures in a video encoder
CN1960490B (en) * 2005-11-04 2010-08-18 腾讯科技（深圳）有限公司 Method for converting GIF file to SWF file
CN102111542A (en) * 2009-12-28 2011-06-29 索尼公司 Image processing apparatus, imaging apparatus, image processing method, and program
CN102111542B (en) * 2009-12-28 2014-11-26 索尼公司 Image processing apparatus, imaging apparatus and image processing method

Similar Documents

Publication Publication Date Title
CN1136737C (en) Image conversion and encoding techniques
CN1490643A (en) Method for image treating apparatus, recording media and program
CN1701596A (en) Image pickup apparatus and synchronization signal generating apparatus
CN1113320C (en) Method of converting two-dimensional images into three-dimensional images
CN101069417A (en) Image processing method, image processing apparatus, image processing program, and image file format
CN1195374C (en) Imaging signal processing method and device
CN1264048C (en) Image processer, image photographic system and image display system
TWI423165B (en) Image processing apparatus, image processing method, and storage medium
JPH10509853A (en) Method and apparatus for image-to-scheme conversion
CN1930519A (en) Image pickup device
TW201143355A (en) Apparatus and method for video processing
CN1556938A (en) Display device and display method
CN1595973A (en) Method and system for converting interlacing video stream to line-by-line video stream
CN1127856C (en) Motion image synthesizing device
CN1685704A (en) Image pickup apparatus, photographing method, and storage medium recording photographing method
CN1625268A (en) Solid-state image pickup device and image signal processor
CN1244077C (en) Video information producing device
CN1174645A (en) Picture transfer process and device
US20130120530A1 (en) Image processing apparatus and method and program
CN1671213A (en) Processing apparatus and computer program for adjusting gamma value
CN1313712A (en) Method and device for producing stereo picture
CN1655025A (en) Imaging method and imaging apparatus
JP2012216939A (en) Image processing device, image processing method and program
JP2006313965A (en) Photographic apparatus, and photographed image processing method, and program
CN1708980A (en) Image pickup device

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C02 Deemed withdrawal of patent application after publication (patent law 2001)
WD01 Invention patent application deemed withdrawn after publication