EP0985212A2

EP0985212A2 - Method of and arrangement for recording and reproducing video images

Info

Publication number: EP0985212A2
Application number: EP98900649A
Authority: EP
Inventors: Wilhelmus Hendrikus Alfonsus BRÜLS
Original assignee: Koninklijke Philips Electronics NV
Current assignee: Koninklijke Philips NV
Priority date: 1998-02-02
Filing date: 1998-02-02
Publication date: 2000-03-15

Abstract

A method is described for recording video images and the associated sound with a reduced resolution, in such a manner that the recorded information can be reproduced by means of standard equipment. Video frames are filtered so as to derive browse sets having a reduced number of pixels. A number of 16 browse sets are stored in a frame memory in order to form a mosaic frame of 4x4 browse sets, after which the mosaic frame is recorded. Audio samples are processed to form audio browse samples: the MSH of a first audio sample is used as the MSH of an audio browse sample and the MSH of a ninth audio sample is used as the LSH of this audio browse sample, which is subsequently recorded. In this way, the information required to edit a recording is reduced by a factor of 16.

Description

Method of and arrangement for recording and reproducing video images

Video recording for the recording of image and sound of scenes is generally known today. A camera converts image and sound into electric signals which are stored on a carrier such as a magnetic tape; a playback apparatus can read these signals from the carrier and generates a signal which can be applied to a reproducing apparatus, such as a monitor or a television set, in order to be reproduced.

Conventional practice was to record these electric video signals in analog form. Nowadays digital recording is also known, for which compression techniques such as DVC can be used. In both cases the recording may be regarded as a sequence or series of recorded images with associated sound, which images may have been provided with a unique identification number.

In the field of image recording and reproduction there is a need to be able to arrange the recorded images in a sequence which deviates from the recorded sequence. This is the case, for example, with an amateur filmer who wishes to make a holiday film, but it also applies to a professional film-maker who can or wishes to use only a part of his recordings. Prior to the actual pasting the user should view and judge the recorded sequence and decide which parts of this sequence he wishes to use for recording on a target carrier and in what order; this process will be referred hereinafter as "editing".

Conventional editing is effected by simply playing the carrier with the recordings on a playback apparatus, viewing the reproduced images, and taking notes. Images are pasted in a conventional manner by playing back a carrier with recordings by means of a playback apparatus and by applying the electric signals generated by this playback apparatus to a recording apparatus to record them on another carrier. To paste a recording at a desired location the user should position the recording head of the recording apparatus at the desired location with respect to the target carrier, position the playback head of the playback apparatus at the beginning of the desired recording on the source carrier, and subsequently start the playback apparatus and the recording apparatus at the same time. However, this is a rather intricate, time-consuming and inaccurate procedure.

Therefore, in order to achieve a greater ease of use, computer programs have been developed for the editing of recordings. The recordings are then read into a computer memory. Subsequently, the user views the images read into the computer memory on a monitor or television set for the purpose of editing and gives edit commands to the computer in accordance with his edit decisions. In response to the edit commands entered by the user the computer then reads the images from the memory in the sequence determined by the user, the image signals being available on a signal output of this computer in order to be recorded on the target carrier.

It will be evident that these tasks can be performed by a dedicated computer but also by a suitably programmed standard computer such as a PC. Hereinafter, such a computer will also be referred to as an editing computer. The use of such a computer presents various problems.

First of all, the required amount of storage capacity is comparatively large. For example, the storage of video recordings having a length of 1 hour requires approximately 11 GB when the digital signals have been compressed in accordance with the DVC format. It is to be noted that hereinafter it is assumed that the recordings have been compressed in accordance with the DVC format but it will be evident that the invention is not limited thereto. The use of other compression techniques is also conceivable. Alternatively, the recordings may have been recorded in analog form, but in order to save storage capacity, it is prefened to use a compression technique such as DVC for reading into the computer memory. Secondly, it takes much time to read these video recordings into the memory. Normally, the playback apparatus being used can play the source carrier only at the "normal" speed, which implies that the read-in of said video recordings then takes 1 hour. It is true that some playback equipment is capable of playing at two or even four times the "normal" speed, but this equipment is fairly expensive. Thirdly, a comparatively high bit rate is required for editing the images read into the memory. When the digital signals have been compressed in accordance with the DVC format the bit rate during read-out from the memory is approximately 25 Mbit/s. This also holds for the read-in of the recordings.

It is an object of the invention to provide a solution for the afore- mentioned problems.

The present invention is based on the recognition of the fact that during editing the user does not need images having the full resolution. It is sufficient to view low- resolution images, after which the actual editing operation is performed upon full-resolution images. On the basis of this it would in principle be possible to derive low- resolution images from the original recordings of full-resolution images. Such a procedure is useful if the invention is used in conjunction with existing recordings. When the images are read into the memory of the editing computer the original recordings are then played back, the computer thus receiving a sequence of full-resolution images. The computer has then been programmed to turn each full-resolution image received into a reduced-resolution image and to store this reduced-resolution image in the computer memory. This already yields the advantage that less storage capacity is required. During editing these reduced-resolution images are used, which has the advantage that the bit rate is comparatively low. However, in accordance with the invention dual recording is effected during the recording of the images: of each image to be recorded a full-resolution version is recorded and a version of strongly reduced resolution, the full-resolution image and the low- resolution image being provided with mutually conesponding identification numbers. During the read-in of images into the editing computer the original low-resolution images are played back, which also has the advantage that the bit rate during reading into the computer memory is comparatively low. Eventually, the edit commands selected by the user will be performed upon the original high-resolution images.

As already stated, a recording comprises a sequence of discrete images, the normal reproduction time for an individual image being fixed; this time is 1/25 s when the PAL standard is used. The reproduction of a low-resolution image would then also require 1/25 s. In accordance with a further aspect of the present invention a multiplicity of N low-resolution images is combined to form a single combination image of normal resolution, in the form of a so-called mosaic image. In a manner comparable to that described hereinbefore, this combination can be effected by the editing computer during the read-in of the images into the computer memory, which already yields the advantage that viewing a recording takes less time. However, preferably this combination is already effected during said dual recording, which has the advantage that during the read-in of the combination images into the memory of the editing computer, which combination images can be read in by means of a standard playback apparatus, the required time has been reduced by a factor N. Indeed, now one combination image is read and loaded into the computer memory in 1/25 s, which combination image corresponds to N original low-resolution images.

The afore-mentioned and other aspects, characteristic features and advantages of the present invention will be elucidated with the aid of the following description of a prefened embodiment of the invention, with reference to the drawings, in which:

Figure 1 is a block-schematic diagram which shows an arrangement in accordance with the present invention; Figure 2 diagrarnmaticaUy illustrates the composition of a mosaic frame;

Figure 3 is a block-schematic diagram of an editing computer as regards video;

Figure 4 diagrammatically illustrates the composition of an audio browse sample; and Figure 5 is a block-schematic diagram of a recording arrangement as regards audio.

Figure 1 shows a block-schematic diagram of a recording arrangement 1 in accordance with the present invention. The recording arrangement 1 has a digital signal input 2 arranged to receive digital video information. This digital video information can be furnished, for example, by a recording apparatus 5, such as a camera, to be connected to the input 2. However, this video information can also be supplied by a playback apparatus 6 for playing back existing recordings, to be connected to the input 2.

It is to be noted that the digital signal input 2 can be used in conjunction with a signal-delivery apparatus, such as a camera 5 or a playback apparatus 6, if such an apparatus is adapted to supply digital video signals. For the cooperation with a signal- delivery apparatus 5, 6 adapted to supply analog video signals the arrangement 1 may be provided with an analog signal input 2' connected to a converter 7 for supplying digital video signals, which converter 7 can have its output coupled to the digital input 2. For the cooperation with a signal-delivery apparatus 5, 6 adapted to supply compressed digital video signals the arrangement 1 can be provided with a second digital signal input 2", which is connected to a decompression means 8 for supplying digital video signals, which decompression means 8 can have its output coupled to the digital input 2. The inputs 2 and/or 2' and/or 2" can be external connections.

It is to be noted that hereinafter the recording arrangement 1 will be described as a separate arrangement but that the recording arrangement 1 can also form part of a camera 5. In that case the input terminals may be dispensed with and the input 2 is internally coupled to an image-signal generating means of the camera.

The recording arrangement 1 comprises two recording channels 10, 20. The first recording channel 10 will be refened to as the normal recording channel and is adapted to produce a standard digital video recording on a carrier 11, which is suitably a magnetic tape. For this purpose, the normal recording channel 10 comprises a first compression means 12, which has an input coupled to the input 2, and first recording means 13, which are coupled to the compression means 12 to receive a compressed digital video signal and to effect a recording of compressed digital video signals on the carrier 11.

The compression means 12 can be a known compression means, for example adapted to effect compression in accordance with the JPEG format, the MPEG format or the DVC format. Since the present invention is very suitable for use in conjunction with DVC compression, it is assumed hereinafter that in each case DVC compression is used. As the nature and construction of the compression means 12 fall beyond the scope of the present invention and an expert does not require knowledge thereof for a proper understanding of the present invention, while furthermore compression means are known per se and the present invention can be applied using a known compression means, said means will not be described in further detail. For similar reasons the recording means 13 will not be described in more detail. It is to be noted only that the digital video signals recorded on the carrier 11 can be reproduced by means of a standard playback apparatus such as a digital video recorder.

The second recording channel 20, refened to as the browse-recording channel, is adapted to effect a digital video recording in an intermediate memory 21, which is suitably a hard-disk included in the recording arrangement 1. For this purpose, the browse-recording channel 20 comprises a filter means 22, which has an input coupled to the input 2, a frame memory 23, a second compression means 24, and second recording/read means 25 coupled to the second to the second compression means in order to receive a compressed digital video signal and in order to effectuate a recording of compressed digital video signals in the intermediate memory 21.

For the same reasons as mentioned in the foregoing, the second compression means 24 and the second recording/read means 25 will not be described in more detail. It is to be noted merely that the second compression means 24 is preferably adapted to operate in accordance with the same compression format as the first compression means 12. Furthermore, it is to be noted that the second recording/read means 25 and the intermediate memory 21 can together be formed by a standard hard-disk unit.

The recording of digital video signals in the intermediate memory 21 will now be described, for which also reference is made to Figure 2.

Video recordings take the form of a sequence of successive separate pictures, referred to as "frames". Each frame may be regarded as a set of picture elements or pixels, the number of pixels in a vertical direction being referenced py and the number of pixels in a horizontal direction being referenced p_jj. As is known, a television set does not display the pixels of a frame at the same time but the frames are written in the form of (py) horizontal lines, writing being effected in an interlaced fashion.

It is obvious that a video signal which represents a frame should comply with a multitude of rules, which rules are together referred to as a video standard or video format. An example of such a format is PAL; writing one complete frame takes 1/25 s in the case of the PAL format. When a video signal complies with a standard it can be applied directly to a reproducing apparatus adapted to this standard; a video signal which represents a frame but which does not comply with the video standard of a reproducing does not yield normal images during reproduction on this reproducing apparatus, in which case the images are usually not identifiable by a viewer. Hereinafter, the term "frame" is assumed to refer not only to a set of pixels which fill a whole screen but also to a video signal which complies with a predetermined video standard.

The upper part of Figure 2 shows a series of successive frames numbered N, N+ 1, N+2 etc. In said first recording channel 10 these frames N, N+ 1, N+2 etc. are recorded as successive full frames and, in addition, identification information such as an identification number (N) is recorded for each frame. In the second recording channel 20, however, the successive frames are first subjected to a filter operation by the filter means 22, as a result of which the number of pixels is reduced. The filter factor in a horizontal direction is designated a_H and the filter factor in a vertical direction is designated a_v. As a result of said filter operation the resolution of the frame in a horizontal direction is reduced by said factor a_H and the resolution of the frame in a vertical direction is reduced by said factor a_v. This results in a set of pixels refened to as a "browse set". The number of pixels in a browse set is smaller than the number of pixels in a frame. The number of pixels in a horizontal direction of a browse set can be written as PH/SH ^an^^{i me} number of pixels in a vertical direction of a browse set can be written as p_v/a_v. The total number of pixels in a browse set can therefore be written as (p_H.p_v)/a_H.a_v).

In principle, a_v and a_H may differ from one another and can be arbitrary, preferably integral, numbers. Preferably, a_v and a_H are 4 because filtering can then be effected properly and simply in combination with existing compression systems operating with DCT blocks of 8x8 pixels. Said filter operation can be effected by each time removing one pixel from a series of successive pixels and skipping others. Said filter operation can also be effected by each time averaging a number of pixels. In said prefened case that a_v and a_H are 4, a first browse pixel could be formed as an average of 16 pixels (4x4) at the top left in a frame etc. Since such a method for the conversion from one resolution to the other is known per se and it will be obvious to the expert how such a filter means 22 can be constructed, a further description of such a filter means 22 will not be given.

The browse set N derived from said frame N is stored in a part of the frame memory 23. A frame memory is a memory having a capacity which is adequate for the storage of one complete frame and is known per se.

Likewise, a browse set N+ 1 is derived from the next frame N+1 and is also stored in a part of the frame memory 23 but at a location other than the location where the preceding browse set N has been stored. For controlling said operations the recording arrangement 1 has a control unit 40, as will be evident to the expert. It will be obvious that it is thus possible to store a total of a_v O a_H browse sets in the frame memory 23. Figure 2 illustrates that 16 successive sets N, N+ 1,

N+2, , N+ 15 are preferably stored in the frame memory 23 in such a manner that first four successive browse sets N, N+1, N+2, N+3 are stored adjacent one another, below these the next four successive browse sets adjacent one another etc. However, another storage sequence is also possible.

Thus, a frame with the pixels of 16 successive browse sets is built up in the frame memory 23; such a frame is refened to as a mosaic frame. When such a frame is displayed on a display screen, the individual browse sets appear adjacent and below one another as thumbnail images, as illustrated in Figure 2. It is to be noted that for the sake of clarity the browse sets and the mosaic frame in Figure 2 have been enlarged in proportion to the original frames.

When the frame memory 23 is full the mosaic frame stored therein is compressed by the compression means 24 and is stored in the intermediate memory 21 by means of the recording/read means 25. After this, a new sequence of 16 successive browse sets is loaded into the frame memory 23 and subsequently the relevant mosaic frame is compressed and stored in the intermediate memory 21, which process is repeated as long as frames are being recorded on the carrier 11.

Once the described recording operation has been completed each frame of the recordings has thus been stored twice: one time as a complete full-resolution image on the carrier 11 and one time as a browse set of reduced resolution, forming part of a mosaic frame in the intermediate memory 21. Obviously, there is a 1-to-l relationship between the complete images on the carrier 11 and the browse sets in the intermediate memory 21. Preferably, an identification number (N) is assigned to each frame and browse set, respectively, during recording of the frames on the carrier 11 and during storage of the browse sets in the intermediate memory 23, which identification number is stored together with these frames and browse sets.

With the scope of the present invention the mosaic frames and the complete frames can be provided on two separate carriers which are associated with one another. For this purpose it is useful, for example, to construct the intermediate memory 21 as a storage means which can be loaded directly into a PC in order to be read, such as for example a PCMCIA card with hard disk, known per se. However, for given uses it is desirable that the mosaic frames from the intermediate memory 21 are also written onto the carrier 11. Preferably, an initial part of the carrier 11 is left free for this purpose when recording commences. Upon completion of the recording the carrier 11 is rewound to the beginning and the mosaic frames are read from the intermediate memory 21 by the second recording/read means 25, which transfer the information thus read to the first recording means 13, which record the received information on said initial part of the carrier 11. To this end, the second recording/read means 25 are coupled to the first recording means 13. If a_H and a_v are both 5 one mosaic frame in the intermediate memory 21 conesponds to 25 frames on the carrier 11, which conesponds to one second of real-time video (PAL). A one-hour video recording on the carrier 11 (90,000 frames) then conesponds to 3600 mosaic frames, which can be recorded on an initial part of the carrier 11 having a length of 2.4 minutes. As stated above, both a_H and a_v are preferably 4; in that case a one- hour video recording on the carrier 11 conesponds to 5625 mosaic frames, which can be recorded on an initial part of the carrier 11 having a length of 3.75 minutes.

Although in the foregoing the recording arrangement 1 has been illustrated as being constructed by means of hardware components, it will be evident that it is alternatively possible to integrate parts of the recording arrangement 1. It is also possible to realize the recording arrangement 1 in the form of a suitably programmed standard computer such as a PC.

Moreover, it will be evident that it is also possible to produce a series of mosaic frames with an existing recording made on a carrier 11. To this end, the carrier 11 can be played back by means of a read apparatus connected to one of said inputs 2, 2', 2". It is likewise possible to construct the first recording means 13 as recording/read means, an output of the recording/read means 25 acting as read means being connected to the filter 22.

If, as stated hereinbefore, the mosaic frames have been stored on the carrier 11 in accordance with the same format as the full frames, the present invention yields the important advantage that a user can view the mosaic frames using standard equipment, such as a standard playback apparatus and a standard TV set. The mosaic frames are decompressed by the playback apparatus in a standard fashion and the decompressed video signal is applied to the TV set. The display screen of the TV set then displays a sequence of mosaic frames. If the playback apparatus has been set to the normal playback speed, the mosaic frames are reproduced in PAL format at a rate of 25 mosaic frames per second.

Thus, by means of standard equipment the user can rapidly get an overview of the recordings stored on the carrier 11.

Hereinafter, the editing and pasting of recordings on the carrier 11 will be described with reference to Figure 3. An editing computer 100, which can be a suitable programmed standard computer such as a PC, having a central processing unit 140, has a terminal 102 for the connection of a recording/reproducing apparatus 53, for example a digital video recorder. In a first stage the recording/reproducing apparatus 53 is operated in a read mode in order to read the recorded mosaic frames from the carrier 11 and to apply the information thus read to the terminal 102, the compressed mosaic frames being stored in an intermediate memory 110 of the editing computer 100, which intermediate memory can be, for example, a hard disk or RAM.

An important advantage of the present invention is that the time for reading the mosaic frames into the intermediate memory 110 of the editing computer 100 is comparatively short. If the recording/reproducing apparatus 53 is operated at the normal speed (real time) the read-in of the mosaic frames conesponding to one-hour real-time full video frames requires only 3.75 minutes if a_H and a_v are both 4. Another important advantage of the present invention is that this requires only 700 MB storage capacity.

In a second stage the user can view and judge (edit) the image information stored in the intermediate memory 110. There are various possibilities of doing so. In the first place, the user can simply display the recordings as successive mosaic frames on a display screen 120. It is then possible to choose from different reproducing speeds, varying from still to the normal speed (25 mosaic frames per second).

In the second place, the user may opt to display only one browse set at a time rather than to display several browse sets at the same time on the display screen 120. A mosaic frame is decoded by decoding means 121 and the decoded mosaic frame is stored in a frame memory 122. The browse set selected by the user is read from the frame memory 122 and displayed on the display screen 120. When the browse set is displayed with its intrinsic size only l/16th part of the display screen is utilized. However, it is likewise possible to operate the arrangement 100 in a scaled-up mode in which the number of pixels both in a horizontal direction and in a vertical direction is doubled by interpolation. Such a scaled-up browse set occupies l/4th part of the display screen.

When the browse sets are thus displayed the 16 successive browse sets can be read sequentially from the frame memory 122 and displayed on the display screen. This can be effected in real time, which means that 25 browse sets per second (PAL) are displayed in succession. Since a decoding operation is required only once every 16 browse sets, i.e. once for each mosaic frame, decoding can readily be effected by means of a software implementation. Displaying several browse sets at a time makes it easier for the user to get an overview of the successive browse sets. In the case of real-time display of the browse sets it is also possible to reproduce sound, as will be explained hereinafter.

In each of said display modes the user can take decisions as regards edit operations. For this, the user utilizes a list of edit commands, which list is for example entered into the editing computer 100 by means of a keyboard 141 and stored in an edit- command memory 142 of the editing computer 100. Typically, an edit command has three parameters: fragment number i, start frame A(i), and end frame B(i). This means that the ft video fragment to be edited starts with the browse set bearing the identification number A(i) and ends with the browse set bearing the identification number B(i). In a monitor mode the user can view the result of his edit commands. The browse sets are then reproduced on the screen 120 in the sequence dictated by the edit commands, the associated sound being also reproduced.

When the user is satisfied he gives a definitive edit command to the editing computer 100. In the definitive editing stage the editing computer 100 will perform edit commands upon the full frames recorded on the carrier 11, i.e. the ft video fragment to be edited begins with the full frame bearing the identification number A(i) and ends with the full frame bearing the identification number B(i). First of all, the full frames are loaded into a memory of the editing computer 100, for which purpose the intermediate memory 110 can be used. Subsequently, the full frames are read out of the memory in the sequence given by the edit commands and are applied to an output of the arrangement 100, which output can be connected to the same terminal as the input 102, thus enabling the output sequence of full images to be recorded on a target carrier by the recording/reproducing apparatus 53.

It is to be noted that if the images have been encoded in accordance with the DVC format the images can be stored in the memory in encoded form and can be output in the same encoded form, as a result of which it is not necessary to perform any decoding and encoding operations while edit commands are carried out.

Furthermore, it is to be noted that the editing computer 100 can be adapted to load all the images from the carrier 11 into the memory while the full images are being read from this carrier. This could be effected while the user is editing the browse sets. However, it is also possible that the editing computer 100 is adapted not to read the full images from the carrier until after the definitive edit command has been given and then to utilize this definitive edit command by first deriving a list of the desired frames from the list of edit commands and, subsequently, to load only these desired frames into the memory, which also reduces the storage capacity required during the actual editing operation.

The recording method and editing method described above only relate to the video components of the recordings. This could be satisfactory in situations where the sound of the recordings is not relevant or is not used at all, for example because the sound is added to the edited recordings at a later instant. However, it also conceivable that the accompanying sound is to be edited during editing of recordings and this sound should therefore be monitored; an example of such a situation is a report with spoken comments, or an interview. Therefore, it is a further object of the present invention to provide a method of recording audio with reduced resolution, the recorded audio being suited for reproduction by means of standard playback equipment. A method of reducing the audio data by a factor of 16, which conesponds to said prefened reduction of the video information by a factor of 4x4, will be described with reference to Figure 4. From the description it will be apparent to an expert how other reduction factors can be obtained.

In digital audio recording a sequence of digits is recorded on the carrier, which may each be regarded as a sample of the sound amplitude. In Figure 4 this is illustrated by representing the (mono) audio recording as a series of juxtaposed samples, which successively bear a sample identification number M, M+ l, M+2, ... etc. Figure 4 more in particular shows two such audio recordings side by side for the purpose of stereo recording, one for the left-hand sound channel (L) and one for the right-hand sound channel (R). In the DVC format this sampling is normally effected 48,000 times per second (per channel).

In order to obtain a reduction factor of 16 it is possible, in principle, to skip each time 15 samples and to record each 16^th sample. This conesponds to sampling at a rate of 3 kHz, which makes it possible to reproduce sound in the spectrum from 0-1500 Hz. However, in accordance with the present invention a better sound quality can be achieved by recording only a predetermined number of the most significant bits of each sample.

Hereinafter, the example is described of a 16-bit sound recording, i.e. a recording in which each sample comprises 16 bits, which enables a signal-to-noise ratio (SNR) of 96 dB to be obtained. Of these 16 bits the first 8 bits will be refened to as the most significant half (MSH) of the sample and the last 8 bits will be refened to as the least significant half (LSH) of the sample.

In an embodiment of the present invention one 16-bit browse sample is derived from a sequence of 16 successive audio samples M, M+ l, ... M+15 by using the MSH of the first sample as the MSH of the browse sample and by using the MSH of the ninth sample (M+ 8) as the LSH of the browse sample.

In Figure 4 the MSHs of the samples of the left-hand channel thus used for composing browse samples are successively referenced 1L, 2L, 3L, 4L, ..., and the MHSs of the samples of the right-hand channels thus used for composing browse samples are successively referenced 1R, 2R, 3R, 4R, ... The structure of the browse audio recording is illustrated diagrammatically in the lower part of Figure 4, which shows for example that the first browse sample shown for the left-hand channel has an MSH consisting of MSH 1L and an LSH consisting of MSH 2L.

This sampling is thus based on taking 8-bit samples at a sampling rate of 6 kHz, which enables a sound reproduction with an SNR of 48 dB in the range of 0-3 kHz (telephone quality).

Now reference is made to Figure 5, which illustrates the audio recording by means the recording arrangement described with reference to Figures 1 and 2.

In addition to the video information the input 2 also receives the audio information. In the first recording channel 10 this audio information is recorded on the carrier 11 in a standard fashion by first audio recording means 33. In the second recording channel 20 a browse sample is defined from a sequence of 16 successive audio samples in a buffer memory 34, which browse sample is applied to second audio recording means 35 for recording in an audio intermediate memory 31. It is to be noted that the audio intermediate memory 31 and the previously mentioned video intermediate memory 21 can form part of the same memory device such as a RAM or a hard disk, and that the second audio recording/read means 35 and the previously mentioned audio recording/read means 25 are in fact the same recording/read means. After completion of a recording the browse samples are read out of the audio intermediate memory 31 by the second audio recording/read means 35 and are recorded on the carrier 11 by the first audio recording means 33.

It has been explained above that the browse video can be reproduced by means of a standard playback apparatus, the successive frames being displayed in the form of a mosaic frame. In accordance with an important aspect of the present invention the browse audio can also be reproduced, together with said browse video, by means of this standard playback apparatus. When the playback apparatus is operated at the normal playback speed the browse samples are interpreted as "normal" samples, which means that substantially only the MSH of each browse sample contributes to the sound reproduction: the LSH of each browse sample may be regarded as a kind of noise at a level of -48 dB. The audio browse samples reproduced at the "normal" speed (i.e. 48,000 per second) consequently conespond to an original sound version whose speed is 16 times as high (which speed is exactly equal to the factor 16 video acceleration) with a sound quality of 1.5 kHz and 48 dB.

As explained, the browse sets can be reproduced in real time during the reproduction of browse recordings by means of the editing computer 100 (Figure 3). The editing computer 100 is adapted to convert the audio browse samples recorded on the carrier 11 into edit samples, and to reproduce these edit samples in real time, as follows. Each browse sample is converted into two edit samples, the MSH of the first edit sample being similar to the MSH of this browse sample and the MSH of the second edit sample being similar to the LSH of this browse sample; the LSH of both edit samples is made equal to zero. The edit samples thus computed are reproduced with a frequency of 6 kHz. Referring to Figure 4, it will be evident that this results in a sound reproduction at "normal" speed, with a quality of 3 kHz and an SNR of -48 dB.

The present invention thus provides a method of recording image and sound recordings with reduced resolution, which resolution reduction amounts to a factor of 16 in the described example, which is a prefened embodiment. The method can be applied to existing recordings but can be used with great advantage for making new recordings, in which case recording in accordance with the present invention is effected parallel to a normal recording. In accordance with an important aspect of the present invention said resolution reduction is effected in such a manner that the number of video frames to be recorded, as well as the number of audio samples to be recorded, is reduced but these video frames and audio samples are recorded in a normal manner, which yields the important advantage that these reduced-resolution video frames and these reduced-resolution audio samples can be read and reproduced by means of normal equipment. Moreover, each reduced-resolution video frame (mosaic frame) includes information of each individual original video frame, in such a manner that of each individual original video frame a scaled-down version is reproduced when the reduced-resolution video frames are read and reproduced by means of normal equipment. It will be evident to the expert that the protective scope of the present invention as defined in the Claims is not limited to the embodiments described and shown in the drawings. For example, in a variant in which the recording arrangement 1 forms part of a camera 5 adapted to supply compressed digital video signals, the compression means 12 in the first recording channel 10 can be dispensed with, in which case the video signals, before they reach the filter means 22 and the control means 40, are decompressed by a decompression means 8 coupled between the input 2 and the inputs of the filter means 22 and the control means 40.

As already stated, the successive mosaic frames are recorded on a carrier in accordance with a video standard, so that when this carrier is played they can be reproduced directly on a standard reproducing apparatus and then produce a mosaic of a plurality of successive thumbnail images. If said intermediate memory 21 has the form of a carrier, such as said example of a PCMCIA card with hard disk, recording in said intermediate memory can also be effected in accordance with a video standard, but if said intermediate memory 21 functions merely as an intermediate memory and actual recording is effected on a carrier 11 with the aid of the recording means 13, it is not necessary to effect the recording in said intermediate memory in accordance with a video standard.

As is known to the expert, compression in accordance with the DVC standard includes a process in which the pixels are subjected to an operation refened to as a "shuffle" operation. Allowance is to be made for this during the generation and recording of the mosaic frames if it is desired that the recorded mosaic frames are to be displayed. This can be achieved in different manners, as will be evident to those skilled in the art. For example, in the recording arrangement 1 as illustrated in Figure 1, an inverse shuffle means is coupled between the frame memory 23 and the second compression means 24 and a shuffle means is coupled between the second recording/read means 25 and the recording means 13. Such inverse shuffle means and shuffle means are known per se and require no further description.

Claims

1. A method of recording video images, comprising the steps of:

(a) supplying a sequence of video frames of normal resolution;

(b) deriving from each video frame a browse set of reduced resolution;

(c) writing a plurality of successive browse sets into a frame memory at different locations to generate a mosaic frame;

(d) storing successive mosaic frames in an intermediate memory.

2. A method as claimed in Claim 1, in which after completion of the recording of the supplied sequence of mosaic frames in the intermediate memory, the mosaic frames are read out of the intermediate memory and are recorded on a carrier as a sequence in conformity with a video standard.

3. A method as claimed in Claim 1, in which the supplied sequence of video frames is recorded on a carrier; and in which after completion of the recording of the mosaic frames in the intermediate memory, the mosaic frames are read out of this intermediate memory and are recorded on a carrier as a sequence in conformity with a video standard.

4. A method as claimed in Claim 3, in which the sequence of mosaic frames is recorded on the same carrier on which the sequence of video frames has been recorded.

5. A method as claimed in Claim 4, in which an initial part of the carrier is left free when the sequence of video frames is recorded on said carrier, and in which the sequence of mosaic frames is recorded on said free initial part of said carrier.

6. A method as claimed in any one of the preceding Claims, in which in step (b) a browse set is generated by filtering a video frame by a factor a_H in a horizontal direction and by a factor a_v in a vertical direction, a_H and a_v being preferably integers; and in which in step (c) a mosaic frame is generated by writing a_H browse sets side by side into the frame memory and by writing a_v browse sets below one another into the frame memory.

7. A method as claimed in Claim 6, in which a_H and a_v are both 4.

8. A method as claimed in any one of the preceding Claims, in which in addition a sequence of audio samples is supplied in step (a), in which a browse sample is derived from a plurality of successive audio samples, and in which the successive audio browse samples are recorded in an intermediate memory.

9. A method as claimed in Claim 8, where appendant to one of the Claims 2- 5, in which the successive audio browse samples are recorded on said carrier together with the mosaic frames.

10. A method as claimed in Claim 8, where appendant to one of the Claims 6- 7, in which each audio sample comprises B bits, B being preferably equal to 16; in which the first B/2 bits of an audio sample are refened to as MSH and in which the last B/2 bits of an audio sample are refened to as LSH; and in which an audio browse samples is derived from (a_H.a_v) successive audio samples by taking the MSH of the first audio sample and using it as the MSH of the audio browse sample to be derived, and by taking the MSH of the (l +(a_H.a_v)/2)^th audio sample and using it as the LSH of the audio browse sample to be derived.

11. An arrangement for making video recordings using the method as claimed in any one of the preceding Claims, comprising: a first recording channel (10) including means (12, 13) for recording video frames received at an input (2) onto a carrier (11) in a standard fashion; a second recording channel (20) including an intermediate memory (21) and means (24, 25) for writing browse sets derived from said video frames into the intermediate memory (21), which second recording channel (20) also includes a filter (22) and a frame memory (23).

12. An arrangement as claimed in Claim 11, in which the first recording channel (10) also includes means (33) for recording audio samples on the carrier (11), and in which the second recording channel (20) also includes means (35) for writing audio browse samples into an intermediate memory (31).

13. An arrangement as claimed in Claim 11 or 12, including means (25; 35) for reading the browse sets written into intermediate memory (21) and the audio browse samples written into the intermediate memory (31), and including means (13, 33) for recording the information thus read onto the carrier (11).

14. An arrangement as claimed in any one of the Claims 11-13, which arrangement forms part of a camera (5).

15. An editing computer (100), comprising:

- a central processing unit (140);

- an intermediate memory (110);

- input means (141) for entering edit commands; - an edit-command memory (142);

- decoding means (121);

- a frame memory (122); in which the editing computer is adapted to receive a sequence of mosaic frame and, optionally, a sequence of audio browse samples, and to store the received information in the intermediate memory (110); in which the editing computer is adapted to read a mosaic frame from the intermediate memory (110) and, after decoding, to store it in the frame memory (122); in which the editing computer is adapted to successively supply the individual browse sets of the mosaic frame stored in the frame memory (122), for display on a display screen (120); in which the editing computer is adapted to read an audio browse sample from the intermediate memory (110) and to derive two audio samples therefrom, the MSH of the first audio sample being similar to the MSH of the read-out audio browse sample while the LSH of the first audio sample being zero, and the MSH of the second audio sample being similar to the LSH of the read-out audio browse sample while the LSH of the second audio sample is zero; in which the editing computer is adapted to supply the audio samples thus generated at a rate suitable for sound reproduction.

16. An editing computer as claimed in Claim 15, which editing computer is adapted to reproduce said browse sets and said sound in a sequence determined by information stored in said edit-command memory (142).

17. An editing computer as claimed in Claim 15 or 16, which editing computer is adapted to receive a sequence of full-resolution video frames and audio samples, conesponding to said mosaic frames and audio browse samples and store them in the intermediate memory (110), and to read said full-resolution video frames and audio samples from said intermediate memory (110) and to supply at an output for editing in a sequence determined by the information stored in said edit-command memory (142).

18. An editing computer as claimed in Claim 17, which editing computer is adapted to store in the intermediate memory (110) only those of the received full-resolution video frames and audio samples which, on the basis of the information stored in said edit- command memory (142), should be used for editing.