Note: Descriptions are shown in the official language in which they were submitted.
- wo 92/0~6~8 ~ S9 PCI/NL91/00177
. .
Read device
.:,
; The invention relates to a read device for reading a record carrier
on which coded pictures have been recorded, which coded pictures represent
- ~ pictures made up of a matrLlc of P rows of Q pixels, Q and P being integers, the
. ratio between the dimension of the pixels in a first direction and the dimension of -
5 the pixels in a second direction substantially perpendicular to the first direction
being substantially unity, the read device comprising a read unit for reading the
recorded coded pictures and for supplying coded pixels representing the pixels, a
picture memory unit comprising a picture memory having memory locations,
storage means for storing in the memory locations of the picture memory coded
10 pixels representing the pixels, and fetching means for reading the information
from the memory locations of the picture memory, the picture memory unit
being adapted to supply the coded pixels in a sequence related to the sequence of
receipt, the read device further comprising means for converting the supplied
coded pixels into a picture information signal suitable for a picture display unit
15 for displaying a representation of the coded picture stored in the picture memory
and represented by the picture inforrnation signal.
Such a system is known, inter alia from the book "Compact Disc
Interactive, a designer's overview", published by Kluwer (ISBN 9020121219).
20 This book describes the so-called CD-I system. This system enables
representations of coded pictures stored on a "Compact Disc" to be read and to
be displayed on a display screen.
A problem encountered with such systems is the storage capacity
required for the picture memory. With CD-I systems it is customary to store a
25 coded picture comprising coded pixels, each representing a pixel whose ratio
between the dimensions in the horizontal direction and the dimensions in the
Yertical direction is substantially equal to unity (1.05 for PAL and 1.19 for
.', . .
''
'.
' . ' ~ ` ' . . ! : - . ' ' . , : ' ~: : ' . . ., ! . . . ,. - . : . .. .. . . . .
-WO 92/0:,6~8 PCl`/NL91/00l77
2~
' :
-,NTSC). When a representation of the coded picture is displayed on a standard
TV set it is advisable to select the number of picture lines of the coded picture to
be substantially equal to the number of picture lines on the TV display screen.
During reproduction a coded picture is read from the CD-I disc
and is stored in a memory. The coded picture stored in the picture memory is
then converted into a video signal suitable for the TV set. The memory locationsin the picture memory are arranged in rows and columns in the customary way.
For practical reasons known per se the number of rows and columns is
preferably equal to a power of 2. The picture memory may then simply comprise
- 10 commercially available integrated memory circuits. In view of the above-
mentioned requirement as regards the number of coded picture lines and the
- . dimensions of the memory a picture memory of 512 by 512 memory locations is
required for displaying a representation on a TV set in the case that the coded
picture is a picture having an aspect ratio of 2/3, as is customary in photography.
15 Indeed, the power of 2 which is approximately equal to the number of picture
`~ lines on the display screen is 512, so that it is preferable to select 512 for the
number of rows in the memory. The number of pixels per picture line then
- becomes equal to 768. The next power of 2 greater than 768 is 1024, so that the
number of columns in the memory is preferably e~ual to 1024.
It is one of the objects of the invention to provide a read device
which enables a memory with a smaller storage capacity to be used and which
yet enables the above-mentioned requirements as regards the dimensions of the
memory to be met. It is another object to realise this without any appreciable loss
of picture quality on a display screen of a standard TV system.
A further object of the invention is to provide rneans enabling an
enlarged and/or rotated representation of a coded (sub) picture to be obtained in a
simple manner.
- In a system as defined in the opening paragraph the first-
mentioned object is acnieved in that the read device comprises picture conversion
30 means for converting the read-out coded pictures into an adapted coded picture
comprising P' rows of Q' coded pixels representing pixels whose ratio between
- the dimension in the first direction and the dimension in the second direction is
:
- - - - -- - - . . - ~.
.. . . . . . .
WO 92/05658 PCI/NL91/00177
: 3
--- substantially equal to Q/P, which picture conversion means are arranged in a data
path between the read unit and the means for converting the supplied coded
pixels into the picture information signal.
The usP of the picture conversion means enables the adapted
. 5 coded picture to be stored in a memory whose number of columns and number of
rows are equal to each other and are a power of 2. In practice, this malces it
possible to use a memory which simply comprises integrated memory circuits
and whose storage capacity is optimised for the coded picture. Moreover, in the
case of reproduction of an enlarged representation of a part of the coded picture
the use of the picture conversion means makes it simple to adapt the number of
coded pixels to the storage capacity of the picture memory. -
If the number of picture lines of the coded picture is selected to
be a power of 2 or a multiple thereof an embodiment, which is attractive becauseof its simplicity, is characterised in that the picture conversion means comprise a
sample rate converter for converting a series of a specific number of consecutive
- coded pixels into an adapted series of an adapted number of pixels, which sample
rate converter is arranged between the read unit and the picture memory unit.
Another embodiment of the device is characterised in that the
picture memory unit is adapted to modify the relationship between the sequence
in which the coded pixels are supplied by the picture memory and the sequence
in which the coded pixels are received, in such a manner that the orientation ofthe picture represented by the supplied pixels is rotated through an angle of 90degrees relative to the orientation corresponding to the original relationship
between said sequences.
The combination of the picture conversion means and the means
. for modifying the relationship between the picture-memory read-in and read-out
sequence has the advantage that it is simply possible to obtain a rotated
representation of the picture with the display screen of a display-screen unit
being filled correctly. The screen is filled appreciably better than in the case of
rotation without the use of picture conversion.
A further embodiment is characterised in that P' substantially
corresponds to the number of useful picture lines per picture as defined in the
.. ' .~. :;
- ,
WO 92JOa6a8 P~/NL91/00177
-:
f~s~ 3
: 4
. PAL TY standard, the SECAM TV standard or the NTSC TV standard. In this
way it is achieved shat a very good picture quality of the displayed representation
can be obtained by means of a picture memory having a very small storage
capacity.
:. S All this makes it possible to use a picture memory of 512x512
memory locations, which is a reduction by a factor of 2 in comparison with ~he
- storage capacity of the picture memory used customarily used in Ihe ~own CD-I
system. Moreover, the speed with which the coded pictures are retrieved from
the picture memory is only 2/3 of the retrieval speed normally attainable with the
known CD-I systems.
The reduction of the number of pixels from 768 to 512 pixels per
line does not visibly affect the quality of the landscape-format representation on
an N~SC TV set or a PAL TV set. This is because the resolution of the display
screens used in these sets substantially corresponds to the bandwidth of the
picture information signal, which is approximately 4 to 5 MHz. This corresponds
to a resolution of approximately 400 to S00 pixels per picture line.
The invention will now be described in more detail, by way of
example, with reference to Figures 1 to 39, in which
Figures la, lb and lc show a picture-storage system, a picture
retrieval and reproduction system, and a simplified picture retrieval and
- reproduction system respectively,
Figure 2 shows a suitable format for recording picture
information on a record carrier,
-- Figure 3 illustrates a suitable coding of picture information,
: Figure 4 illustrates a suitable residual coding to be used for in
the coding of picture information,
Figure 5 shows how the display screen is filled if a picture is
30 rotated withou~ a picture conversion being applied,
Figure 6 shows a format of a subfile containing a residually
coded picture,
.,` ,~
: ' ' - - - - ,
- WO 92/0~i6~8 9 PCT~/NLS~1/00177
~ . .
; . 5
- Figure 7 shows a record c~rrier on which recorded coded picture
. lines have been arranged in a suitable manner,
. Figure 8 shows a picture composed of picture lines,
Figure 9 illustrates a number of different picture processing
~ 5 functions,
- - Figure 10 shows an embodiment of a retrieval and reproduction . - ::
system capable of displaying picture information in accordance with preferential
. reproduction settings,
- Figure 11 shows a suitable format for recording preferential ~ . :
; - 10 reproduction settings on the record carrier, ~:
Figure 12 shows a suitable format for storing preferential
. reproduction settings in a non-volatile memory,
Figure 13 shows a mosaic picture composed of sixteen low-
- resolution images,
Figure 14 shows in greater detail an embodiment of the
. simplified picture retrieval and reproduction system,
Figure 15 shows an embodiment in which control data groups
can be arranged in packets,
Figure 16 shows a data extraction circuit for use in the picture .
20 retrieval and reproduction system shown in Fig. 14,
; Figure 17 shows in greater detail an embodiment of the picture
storage system,
Figure 18 shows a recording unit for use in the picture storage
system,
. 25 Figure 19 diagrammatically illustrates the CD-ROM XA format,
Figure 20 shows a suitable organisation of the record carrier if
. - .
the picture information has been recorded in accordance with a CD-I forrnat,
Figures 21, 23 and 24 show suitable configurations of picture
lines of absolutely coded pictures for a number of different resolutions if the
: ~ 30 recorded information has been divided into blocks in accordance with a CD-I ' ~;
format,
.:
. _ . . . ... . , . , . ~,, .
WO 92/0~6~8 PCI/NL91/00177
` 6
Figure 22 shows a picture made up of picture lines to illustrate
the configuration shown in Fig. 21,
Figure 25 shows an example of a picture processing unit,
Figures 26 and 27 illustrate picture pr~cessing functions so be..
performed by the picture processing unit,
Fi~ure 28 shows an embc~iment of a read device,
Figures 29 and 31 diagrammatically show examples of a
simplified picture processing unit,
- Figure 30 illustrates the operation of the simplified picture
10 processing unit shown in Figures 29 and 31, and
Fi ,ures 32 to 39 illustrate picture processing functions which can
be performed with the device in accordance with the invention.
'
Figure la shows a picture storage system 12 in which the
- 15 invention can be used. The picture storage system 12 comprises a picture
scanning unit 1 for scanning pictures on a picture carrier 3, for example a strip-
shaped photographic negative or slide. The picture scanning device I further
. ~ comprises a picture coding unit for coding the picture information obtained upon
scanning. The coded picture information is recorded on a record carrier 184 by
20 means of a recording unit 5 under control of a control unit 4. Prior to recording
: the control unit 4 can apply an optional picture processing, for example to
enhance, correct or edit the picture representation defined by the coded pictureinformation. For this purpose the control unit may comprise picture processing
means which are known ~r se. The recording unit 5 may comprise, for
- 25 example, an optical, a magnetic or a magneto-optical recording device. In view
of the high storage capacity of optical~and magneto-optical record carriers it is
preferred to use an optical or a magneto-optical recording device. The control
unit 4 may comprise a computer system, for example a so-called "personal scomputer" or a so-called work station with suitable hardware and application
software.
; Figure lb shows a picture retrie-al and reproduction system for
retrieving and displaying representations of coded pictures stored on the record
. .:. .
WO 92/0~6~8 ~ PCr/NL91/00177
, . .
:
- carrier 184 by means of the picture storage system 12. The picture retrieval and
reproduction system 13 comprises a read unit 6 for locating and reading out
selected coded pictures under control of a control unit 7. Representations of
- coded pictures thus read can be displayed on a picture display unit. Such a
S picture display unit may comprise a display screen 8, which for exarnple forms
part of the control unit 7, or an electronic image printer 9 for generating a hard -
copy 15 of a representation of the read-out coded picture. The picture retrievaland reproduction system 13 may further comprise an additional recording device
5a, by means of which the coded picture information read by means of the read
device 6, after an optional picture processing operation performed by the control
unit 7 for the purpose of enhancement, correction or editing. The control unit in
- the picture retrieval and reproduction system 13 may comprise a computer
- system, for example a "Personal Computer", or a work station with suitable
hardware and application software. Although such a system is very suitable for
the control task to be perforrned and the optional picture processing it has thedrawback that it is comparatively expensive.
- In general, it is desirable to have such an expensive computer
system for the control unit in conjunction with the electronic image printer 9
because of the complexity of the control and picture processing functions.
However, if it is merely desired to display selected coded pictures on a displayscreen, the computing capacity a~d storage capacity of a computer system in the `
form of a personal computer or work seaeion are high in comparison with the
control functions to be performed. In that case ie is preferred to employ a
simplified conerol unit with a limited computing and storage capacity and a
- 25 limited data processing speed.
Figure lc shows such a simplified picture retrieval and
reproduction system 14. This simplified system 14 comprises a display unit 10
. and a picture retrieval and read unit 11 cornprising the read unit 6. A control unit
' for controlling the retrieval and read operation and, if applicable, a limited
, 30 picture processing can be accommodated in one of the units 10 and 11, but
suitably in ;he unit 11. When the control unit is accommodated in the retrieval
.
WO 92/05658 ',~ PCT/Nl,91/00177
; and read unit 11 it is possible to employ, arnongst others, a standard TV set or
monitor unit for the picture display device.
This is an advantage, in particular for consumer uses because the
consumer then merely has to purchase the retrie~al and read device to display the
5 representations of the pictures.
As a result of their comparatively high cost the picture storage
system 12 shown in Fig. la and the picture retrieval and reproduction system 13
shown in Fig. Ib are particularly suital~le for central uses, for example in
photoprocessing laboratories.
For recording coded picture information it is preferred to record
the information on the record carrier in a predetermined format and order.
Figure 2 shows a suitable format and order, in which files containing coded
picture inforrnation bear the references IPl, ..., IPn. Hereinafter the files IPl,
` O . . . ,IPn will be referred to as picture files. Moreover, a plurality of control files
15 BB have been recorded. These files contain control data which is used for
controlling the read-out of the coded picture information, for the purpose of
performing optional picture processing operations on the picture information read
and for the puIpose of displaying representations of the coded picture
information. It is to be noted that part of the control data may be included in the
20 picture files. Preferably, this part of the control data is the part relatingspecifically to the control of the read-out, display and picture processing of the
coded picture information contained in the relevant picture file. The advantage of
this is that the required control data becomes available at the instant at which it is
needed, i.e. at the instant at which the picture file is read.
Apart from the picture files Ip and the associated control files BB
. it may be desirable in a number of cases to record files with additional
. information, for example audio information or text information. Such audio
and/or text information may relate to, for example, coded picture inforrnation
~4 and can then be reproduced or displayed when the representations of the relevant
30 coded picture information are displayed. The files with additional information are
referenced ADD and may be recorded, for example, after the coded picture
information.
~ WO 92/Os6~8 ~r ~ ~?6;2rg PCr/NL91/00177
. . .
., 9 :
For every picture stored the picture files contain a plurality of
:- subfiles, which each define a representation of the sarne scanned picture, the
resolutions of the representations defined by these coded pictures being different.
In ~igure 2 the different subfiles for the picture file IPI bear the references
S TV/4, TV, 4TV, 16TV, 64TV, 256TV. The subfile IV defines a representation
of the scanned picture with a resolution corresponding substantially to a standard
- NTSC or PAL TV picture. Such a picture may comprise, for example, 51~ lines
of 768 pixels each. The subfile TVl4 represents the scanned picture with a
resolution which in the horizontal and the vertical direction has been reduced
- 10 linearly by a factor of 2 relative to the resolution of the picture represented by
the subfile TV. The subfiles 4TV, 16TV, 64TV and 256TV define picture
representations whose horizontal and vertical resolution has been increased
linearly by a factor of 2, 4, 8 and 16 respectively. It is to be noted that,
-~ ~ obviously, the number of subfiles may be selected differently. For example, each
`- 15 picture file may comprise the subfiles TV/16, TV/4, TV, 4TV and 16TV.
Preferably, the subfiles are arranged in such a way that the resolutions of the
- representations defined by the successive coded pictures increase (linearly) in
steps of 2. During reproduction, when the consecutive subfiles are generally read
.: successively, it is then simple to first display a representation of a picture of low
resolution and, subsequently, to replace this representation wholly or partly by : ~:
representations of the same picture of each time increasing resolution. This hasthe advantage that the waiting time before a picture representation appears on the
'~ display screen is minimized. Indeed, on account of the limited amount of
.- information needed for this, the read-out time of a coded picture defining a low-
resolution representation is short in comparison with the read-out time of
encoded pictures defining higher-resolution representations. ;
A generally knowr representation of pictures is that in which the
picture is composed of a matrix Gl small areas of constant luminance value
and/or constant color val,ue. In this representation it is customary to select the
areas of constant color value to ~e larger than the areas of constant luminance
value.
,
. ~' .
: .- . .
WO 92/05658 ~ r~ ^9 PCr/NL91/û0177
. .
An area of constant color value will be referred to hereinafter as
a color pixel and an area of constant luminance value will be referred to
hereinafter as a luminance pixel. A row of color pixels of a width equal to the
full picture width will be referred to hereinafter as a color picture line. A row of
luminance pixels of a width equal to the full picture width will be referred to
hereinafter as a luminance picture line. A picture represented by luminance
picture lines and color picture lines can be defined simply by a coded p;cture by
assigning to each luminance pixel and color pixel a digital code specifying the
relevant luminance value and color values.
-
lD Figure 3 by way of illustration shows the structure of a picture of
color pixels and luminance pixels. rhe luminance pixels bear the reference signs(Y2 1; ...; YK~l R-l) rne color pixels bear the reference signs (C1 1; ...;
CK R). It is to be noted that in Fig. 3, as is customary, the dimensions of the
color pixels in the horizontal and the vertical direction is twice as large as the
15 dimensions of the luminance pixels. This means that the resolution of the color
information in the horizontal and the vertical dire tion is a factor of two lower
than the resolution of the luminance infonnation.
A suitable picture coding is that in which a digital code or digital
codes is/are assigned to every luminance pixel and every color pixel, the code(s)
20 defming the absolute value of the luminance component Y and the absolute
. values of the colour-difference components U and V respectively. Such a coding
will be referred to hereinafter as an absolute picture coding. Preferably,
representations of a number of low-resolution pictures are recorded as absolutely
coded pictures. This enables the picture information to be recovered in a simple25 manner. This is particularly advantageous for the simplified picture retrieval and
- reproduction system 14, because this enables the price of such a system, which is
intended for the consumer market, to be kept low by the use of simple picture
decoding systems.
The use of a picture file with a number of absolutely coded
30 pictures of different resolutions simplifies the reproduction of representations of
-- composite pictures, where a representation of a small low-resolution picture is
displayed within the outline of a representation of a higher-resolution picture. .
.. . .. .. ~ .. . .~ :
WO 92/0~658 ~ 3 `~9 P~/NL91/00177
The reproduction of such a representation of a composite picture is referred to as
"Picture in Picture" (or "PIP")~ Moreover, recording a plurality of absolutely
coded pictures defining representations OI the same picture with different
resolutions simplifies the reproduction of enlarged representations of details of a
5 coded picture. Such a function is also referred to as the IELE-function (or
ZOOM-function). The availability of absolutely coded pictures with different
resolutions implies that for some of the 'I~LE functions and PIP functions the
- required picture information is directly available and need not be derived by
- means of additional picture processing operations to be performed by complex
10 circuits.
In the recording of picture information it is customary to record
the coded pixels in rows (or lines) or sometimes in columns. Recording in lines
: is to be preferred because in the customarily used picture display units the
, picture information should be presented in the form of lines.
When the absolutely coded pictures are recorded in the subfiles `; ~ --
- TV/16, TVt4 and TV it is preferred not to record consecutive coded picture lines
contiguously. Such method of arTanging the recorded information is frequently
referred to as "interleaving". The advantage of such a method is that if a
comparatively great part of the information cannot be retrieved owing to defects20 of the disc or other causes, it reduces the likelihood that two adjacent picture
~, lines in the representation of the coded picture are reproduced incorrectly.
Representations with faults in adjacent picture lines are comparatively difficult to
restore. This is not case with representations in which erroneously read pixels (or
a picture line) are situated between two correctly read picture lines. In that case
25 the erroneously read pixels (or picture line) can be replaced simply by pixels (or
a picture line) derived from one or both adJacent picture lines. It is to be noted
that erroneously read pixels can also be restored readily by the use of so-called
error-correction codes. The correction of errors on the basis of such error-
- correction codes is comparatively intricate and is therefore less suitable for,use in
30 the simplified picture retrieval and reproduction system 14, in which the use of
complex circuits should be avoided whenever possible in view of the resulting , ~ -
high cost. ~ i
.
~'
-- WO 92/0s6s8 ~ PCI/NL91tO0177
2~
12
In the case that the picture inforrnation is recorded on a disc-
- shaped record carrier with a spiral track the part of the track needed for
- recording a coded picture will occupy a plurality of turns of the spiral track. In
view of a simple restoration of erroneously read picture lines it is then desirable
5 that the coded picture lines defining adjacent picture lines in the representation of
the picture to be reproduced do not adjoin each other neither in the track
- direction (also referred to as tangential direction) nor in a direction transverse to
` the track (also referred to as radial direction~, which will be explained with
reference to Figures 7 and 8.
ID Figure 7 shows a disc-shaped record carrier 70 on which picture
80 composed of consecutive picture lines ll, ..., In has been recorded in a spiral
track 71 in the form of a series of absolutely coded picture lines BLal, BLa3,
BLaS, BLa7, BLa9, BLall, BLal3, BLa2, BLa4, .... The absolutely coded
- picture lines BLal, .. , BLal3 represent the picture lines ll, .. , 113 respectively.
- lS The absolutely coded picture lines have been recorded in such a way that the
:. information of consecutive picture lines is not contiguous neither in a radial nor
in a tangential direction. The reference numeral 72 refers to an unreadable discportion, also called disc defect. The defect shown extends over more than one
turn of the spiral track 71. Since the coded picture lines defining adjacent picture
- 20 lines of the representation do not adjoin one another neither radially nor
tangentially this prevents coded picture lines which define adjacent picture lines
in the representation from being read incorrectly as result of the occurrence ofdisc defects. It is to be noted that for the sake of clarity the length occupied by
the coded picture lines BLa on the recording is shown to be substantially greater
than in reality. In practice, it occurs comparatively often that a disc defect
occupies a plurality of consecu~ively recorded coded picture lines. In view of the
requirement that adjacent picture lines should not be defined by adjacently
- recorded coded picture lines the order of the absolutely cod d picture lines in the
track depends shongly on the length of the turns of the spiral track and on the
length required for recording an absolutely coded picture line. Arrangements
-` suitable for recording absolutely coded picture lines will be described in more
detail further on in the description.
:
- : :
, . ~ ~ .. , . ~ .
- WO 92/0~6ag PCI~NL91/00177
13
For high resolutions the storage of absolutely coded picture
information has the drawback that the arnount of information to be recorded is
very large. For such high-resolution pictures a residual coding is very suitable.
In such a residual coding differences between the signal value of the pixels of the
~; 5 high-resolution picture and the signal value of the corresponding part of the
lower-resolution picture are determined ~md subsequently encoded
To illustrate this coding method Fig. 4 shows one luminance
pixel Y of a low-resolution picture and four luminance pixels Yl l'; Y2 l'; Yl 2'
and Y2 2' cf the corresponding higher-resolution picture in the case that the
10 horizontal and the vertical resolution is increased by a factor of 2. Instead of the - ;-
absolute luminance value of the luminance pixels Y~ --, Y2 2' the residual
coding encodes differences (hereinafter referred to as residual values) between
. the luminance values of the luminance pixels Yl 1'~ --, Y2 2' and the luminance
. - 9 pixel Y. In this way the residual values of a complete picture can be determined
15 both for the luminance and for the color information. As the number of residual
values equal to zero or being very small is large in comparison with the number
of large residual values a substantial data compression can be obta~ned by
applying an additional coding in which the residual values aTe non-linearly
quantized and are subsequently subjected to, for exarnple, a Huffman coding.
A residually coded picture can be used as a basis for a new
residual coding for a picture with further increased resolutions. Thus, by
recording one absolutely coded low-resolution picture and a series of residuallycoded pictures of increasing resolutions in compressed forrn it is possible to
record a plurality of coded pictures defining representations of the same picture
25 with increasing resolutions. In the picture file IP1 shown in Fig. 2 the pictures in
the subfiles TV/4 and IV are absolutely coded and the pictures in the subfiles ~ ;
4Tv, 16TV, 64TV and 256TV are residually coded, with non-linear quantization
and Huffman coding. Such a coded picture will be briefly referred to hereinafter.- as a residually coded picture.
The color information is also coded residually in a way similar to
the luminance information. However, the horizontal and the vertical resolution of
the consecutive residually coded color information increases by a factor of four
-
':
WO 92/05658 pcr/Nl~9l/oo177
r~ r~ 3 14
- instead of by a factor of two as with the luminance inforrnation. This means that
a picture file containing only residually coded luminance information and no
: color inforrnation (4TV and 64TV) alternates with a picture file containing both
- residually coded luminance information and residually coded color information
- 5 (16TV and 256TV), see Fig. 2. Leaving out the color inforrnation in the subfiles
4TV and 64TV reduces the required storage capacity and the access time to the
coded picture information in the picture file. However, the absence of the colorinformation in the subfiles 4TV and o41~V need not adversely affect the picture
quality during reproduction. This is because during the reproduction of a
10 representation of a coded picture for which no color information has been
recorded the color inforrnation of the next coded picture defining a representation
of higher resolution or the color information of the preceding coded picture
defining a representation of lower resolution can be utilised. In order to reduce
the access time to the required picture information, utilising the fact that the15 luminance information is perceptively more important than the color information,
it is to be preferred to record the luminance information Y in the subfiles 16TVand 256TV contiguously to the chrominance inforrnation U, V in the subfiles
4TV and 64TV, as is illustrated for the file IP* in Fig. 2.
However, it is to be noted that during the reproduction of the
20 representation of the coded picture recorded by means of the subfile 4TV it is
also possible to use the color information of the subfile TV or the complete color
information of the subfile 16TV.
As already stated, it is customary to record the coded pixels line
by line.
When the residual coding described above is used, using a non-
linear quanti~ation and Huffman coding, the residual values are represented by
means of codes of varying length. This means that the space required for
recording the residually coded picture lines is variable. Therefore, the position at
- which the beginning of the residually coded pi ture line is recorded is not
- 30 unambiguously defined ~y the beginning of recording of the first coded picture
line of a coded picture. This complicates the selective read-out of the coded
picture lines, for example only those coded picture lines needed to carry out a
. .
WO 92/056~8 pcr/NL91/ool77
3 'Jf~
.. ` . 15
- TELE function. This problem can be mitigated by recording a line number LN
(see Fig. 6) at the beginning of each coded picture line BL and line
synchronization code LD. The line synchronization code may be, for example, a
- unique bit combination which does not occur within the series of Huffman codes
5 representing information of the residuaily coded pictuIe elements. It is to be- noted that the addition of the line synchronization codes LD and line numbers
LN has the additional advantage that it facilitates the read synchronization andsignificantly reduces error propagation after an erroneously read residual code.A very fast retrieval of selected coded picture lines can be
L0 achieved in that the addresses at which the recordings of coded picture lines on
the record carrier begin are recorded on the record carrier in a separate control
file, preferably at the beginning of each subfile. In Fig. 6 these addresses have
- been indicated, by way of example, as ADLN#1, .. , ADLN#1009 in the control
file IIDB at the beginnin" of the subfile 4TV. The picture line information in the
- 15 form of the series of residually coded picture lines is insented in the section
APDB of the subfile 4TV. (The section APDB represents the actual picture
information within the subfile 4TV).
Generally, when searching for the starting points of the picture
lines on the record carrier during a coarse search process a read element is
moved relative to the record carrier to a position at a short distance before the
starting point where the recording of the coded picture line begins. Subsequently,
a fine search process is carried out in which, while the record carrier is scanned
with a speed corresponding to the normal read speed, the beginning of the ~ -
recording of the selected residually coded picture line is awaited, after which
reading of the selected coded picture line is started. The accuracy with which the
read element can be positioned relative to the record carrier during the coarse
search pr~}cess is limited and in optical data storage systems it is generally much ~ -
greater than the distances between the positions at which the recordings of
successive coded picture lines on the record carrier begin. Therefore, it is
: 30 preferred to store only the start addresses of a limited number of coded picture
lines whose starting points of recording are spaced apart by a distance
- substantially equal to the accuracy with which the read element can be positioned
.
. ' ' . , ''... ' ' 1 ~ . ' ,, , ' , '.'.' , ' . ~ ' :
,, ... . . . . . . . - . .. . . ; ..
wo 92/05658 PCl/NL91/00177
3~ 9 16
during a coarse search process. This enables the inforrnation of selected coded
- picture lines within a stored coded picture to ~e located and read rapidly without
an unnecessarily large arnount of space being needed for the storage of address
- data. In the case of a disc-shaped record carrier the average search accuracy
S during a coarse search process, in which the read element is radially moved over
the disc, is by definition equal to half the length of one turn of the disc, which
means that the distances between the positions specified by addresses
substantially correspond to half the length of one turn of the disc when disc-
shaped record carriers are used.
The stored coded pictures generally define a number of pictures
in landscape format (i.e. for a faithful reproduction the picture should be
displayed in an orientation in which the width of the picture is larger than thehei,,ht of the picture) and a number of pictures in portrait format (i.e. for a
faithful reproduction the picture should be displayed in an orientation in whichthe height of the picture is larger than the width of the picture).
By way of illustration Figure 1 shows an image carrier with some
pictures in landscape format (2a, 2b, 2c and 2d) and one picture in por~rait
format (2e). On the record carrier all the coded pictures are recorded as thoughthey were representations of pictures in landscape format. This is in order to
enable a uniform picture scanning to be used without the necessity to detect
whether the scanned picture is of the landscape or portrait type and to change
over the scanning and/or picture processing depending upon the detection result.However, this means that during reproduction the representations of portrait
- format pictures will be displayed in an incorrect rotated position. This can be
- 25 precluded by providing a possibility to assign a rotation code to the recorded
coded pictures, which code indicates whether the representation should be rotated
during reproduction and, if this is the case, whether the representation should be
rotated through an angle of 90, 180 or 270 degrees. This rotation code can be
included in every picture file IP1, ..., Ipn. It is also possible to record these
rotation codes in the control file BB or to store these rotation codes in a non-- volatile memory arranged in the read unit or connected to this unit.
- , - . ~ . . ..
wo 92/os6~8 PCr/NL91/00177
3 ~.,9
1 7
- During reproduction it is then possible to determine on the basis
of the rotation code whether the representation to be displayed should be rotated
and, if this is the case, a rotation througlh the desired angle can be performedprior to reproduction. A drawback of including the rotation codes in the pictureS files IP is that these rotation codes have to be determined akeady during
scanning of the pictures. In practice this means that an operator of the picture storage system should determine for each scanned picture whether the stored
picture is to be rotated during reproduction, because the known auxiliary devices
are not always capable of detecting whether a scanned picture is of landscape or10 portrait format and whether the picture is presented to the scanning unit with the
- correct orientation. This is undesirable in particular because it implies that an
operator must be present during recording, which makes it difficult to realis a
fully automated picture storage system 12.
O If the rotation codes are already available during recording of the
lS coded picture information it will be advantageous to record these codes on the
record carrier. In the case of the file organisation shown in Fig. 2 a suitable
position for recording the rotation codes is the subfile FPS of the control file BB. -
For rcasons of user convenience it is desirable to specify, apart from the required
rotation, wether instead of a representation of stored coded pictures a
20 representation which is slightly shifted (to the left, right, top or bottom) should
be displayed. This is certainly desirable if the display area within which the
representation is to be displayed in a display unit is smaller than the dimensions
of the representations, because it is possible that an important detail of the
picture falls outside the display area. The desired shift can be specified by
25 assigning a translation code to every coded picture. In Figure 9 a suitable
translation coding for a picture 90 is defined by means of the coordinates xp and
yp of a vertex 9l of th~ picture 91 to be displayed after translation. By means of
- . a translation code and a magnification code it is possible to specify the
magnification factor with which a certain part of the original picture is to be
30 displayed. The reference numeral 93 indicates an enlarged representation of apart of the picture 90, defined by a translation xp, yp and a magnification factor
of 2. In addition to the above data it is also possible to include other picture
.
.
- WO 92/056~8 PC~/NL91/00177
18
display data in the subfile FPS of the control file BB, such as for example
parameters specifying a color or luminance adaptation to be applied before a
representation of the coded picture is displayed. Moreover, it is advantageous to
. store the desired se~guence in which the pictures must be reproduced in the
- 5 subfile FPS within the control file BB.
The afore-mentioned infurrnation about the display sequence,
rotation, translation, magnification, brightness and color adaptations and other- picture processing operations to be perfo1~ned prior to reproduction of the
representation of the coded picture will be referred to hereinafter as preferential
reproduction settings. A collection of preferential reproduction settings defining
the preferred sequence as well as all the desired picture processing operations for
all the coded pictures on a record carrier will be referred to hereinafter as a set
of preferential reproduction settings. It may be advantageous to record more than
one set of preferential reproduction settings in the file FPS. This enables a
different display sequence and other picture processing operations to be selected
by different persons, for example persons within a family. It also allows a userto make a choice from different sets of preferential reproduction settings. It is to
. be noted that when a record carrier of the write-once type is used the sets of
preferential reproduction settings can be recorded on the record carrier only if, 20 they are available during recording. This requires human intervention during
recording. During reading of the record carrier a set of preferential reproduction
settings is selected and the representations of the coded pictures can be displayed
in conformity with the selected set of preferential reproduction settings. Figure
- 10 is a block diagram of an embodiment of a picture retrieval and display system
2~ by means of which representations of coded pictures can be displayed in
conformity with a selected set of preferential reproduction settings. In this
diagram the reference numeral 100 refers to a read unit for reading the record
carrier. For the purpose of applying the information being read the read unit 100
- is coupled to a control and signal processing unit 101. From the information
. 30 received from the read device 100 the unit 101 selects the file FPS containing the
set(s) of preferential reproduction setting(s) and stores this (these) set(s) in a
- control memory 102. By means of a data entry unit 103, for example a remote
;. .
.. . .
.. . . , ... ,, , , ~ , :
-~ WO 92/05658 pcr/NL9l/oo177
.. ~3r,~
`....................................... 19
control device, a user can select a set from the control memory 102 and can
subsequently activate the unit 101 to start the read cycle, in which the coded
- picture information is read in the sequence specified by the selected set of
preferential reproduction settings under control of the unit 101. After the coded
5 picture information has been read out this inforrnation is processed in accordance
with the selected set of preferential reproduction settings and is applied to a -'
display unit 104.
It may occur that after some time the preferential reproduction
;` settings stored on the record carrier are no longer entirely in compliance with the
- 10 user's wishes or that no or wrong preferential reproduction settings have been
recorded on the record carrier. This is problematic in particular if the record
carrier is of a type which cannot be overwritten, because the recorded
preferential reproduction settings then cannot be adapted. This problem can be
mitigated by providing the retrieval and display system in Fig. 10 with a non-
15 volatile memory 105 in which together with a record carrier identification code a
new set of preferential reproduction settings or inforrnation about the desired
changes of the preferential reproduction settings relative to the set of preferential
reproduction settings recorded on the record carrier is stored for the record
carrier specified by means of the record carrier identification code. In view of20 ~he limited storage capacity of the non-volatile memory 105 it is desirable to
record the information necessary for the preferential reproduction settings in amost compact form, for which reason it is preferred to record the information
about the changes of the preferential reproduction settings.
. Fig. 11 shows-by way of exarnple a suitable format 110 of the
- 25 preferential reproduction settings included in the file FPS on the record carrier.
The format 110 comprises a section DID in which the unique record car~ier
identification code is stored. Such a code may comprise a large random number
generated by means of a random-number generator and recorded on the record
carrier. The code may comprise a time code indicating the time in years, months,days, hours, minutes, seconds and fractions of seconds. Alternatively, the record
carrier identification code may comprise a combination of a time code and a
random number. In the format 110 the section DId is followed by sections FPSl,
.
-- . . - . .--- .. - - . - - - :, . - ,. : .. , .-- . :, - - . :: .: .... . . .
WO 92/0S6S8 pcr/NL9l/oo177
-- FPS2, .. , FPSn in which a number of different sets of preferential reproduction
settings are stored. Each of the preferential reproduction setting sections FPS1,
..., FPSn contains a portion SEL in whic:h a set identification number for each of
the different sets of preferential reproduction settings to be selected by different
5 users are specified, and a portion specifying the sequence SEQ in which the
representations of the stored pictures are to be reproduced. This portion is
followed by the coded sections FIM#1, ..., FIM#n storing for the pictures 1, ....
n preferential processing operations to be perfonned before the representation of
the relevant picture are displayed.
Fig. 12 shows by way of exarnple a suitable format 120 in which
the information about the desired adaptations of the set of preferential
- ~ reproduction settings can be stored in the non-volatile memory 105. The format
120 comprises a section 121 specifying combinations of record carrier
identifications and set identification numbers for which information about
15 preferential reproduction settings has been stored. To each of these combinations
a pointer is assigned, which pointer is included in the section DID-POINT and
specifies the address of the sections DFPSl, ..., DFPSn in the non-volatile
memory 105.
- Every section DFPS comprises a portion LSEQ with a code
20 indicating the space (for example in numbers of bytes) required to specify the
new sequence. If the portion LSEQ indicates a length not equal to zero LSEQ
will be followed by a portion NSEQ with the data specifying the new display
; sequence. After NSEQ the new preferential processing operations are specified
- for every picture with modified preferential processing operations. ROT indicates
25 the section with the rotation code. The sections LTELE and LPAN specify the
length available lFor the storage of the new data relating to picture magnification
(in a section NTELE) and picture translation (in a section NPAN). In this way itis possible to select the accuracy with which the picture processing information is
to be stored. Thus, it is possible, for exarnple, to define three different lengths
30 indicating three clifferent accuracies. LTELE and LPAN are followed by the
portions NI~LE and NPAN. If the information about the picture magnification
and picture translation need not be changed this is indicated by the length zero in
' ~
.. :
WO 92/0s6~8 ~ 3. ~ ~3 PCr/NL91/0û177
21
: LTELE and LPAN. By storing only the preferential processing operations for
pictures with modified preferential processing operations the space required forthe storage of the new preferential reproduction setting is reduced considerably. :
- Apart from the reduction of the required storage space by said recording of the
5 differences it is possible to obtain an additional reduction by specifying the length
required for the storage of modified datal. When the record carrier is read an
adapted set of prefèrential reproduction settings is derived from the preferential
reproduction settings recorded on the record carrier and the differences stored in
the memory 105, and this adapted set is stored in the memory 102.
: 10 Instead of, or in addition to, the non-volatile memory 105 a
changeable memory 106, for example in the form of a magnetic card, EPROM,
EEPROM or NVRAM, can be employed for the storage of preferential
reproduction settings in the retrieval and display system shown in Fig. 10.
: This has the advantage that a user can display the picture
15 information on a record carrier in accordance with the same preferential
- reproduction setting on different picture retrieval and display systems to which a
changeable memory 106 can be connected. When one of the two or both
~; memories 105 and 106 are used for the storage of preferential reproduction
settings it is desirable that a selection is made from the different sets of
20 preferential reproduction settings defined by the sets of preferential reproduction
- settings on the record carrier and by the modifications of the preferential
reproduction settings stored in the memories `.05 and 106. For this purpose the
unit 101 should comprise selection means. These selection means may be of a
type which are operated by the user to make a choice from the various sets of
25 preferential reproduction settings defined for one specific record carrier and
selection number by the preferential reproduction setting information stored on
the record carrier and in the memories 105 and 106. However, alternatively these- selection means may be of a type which, prior to reproduction on the basis of the
contents of the memories 105 and 106 and the sets of preferential reproduc~ion
30 settings recorded on the record carrier, determine the sets of preferential
reproduction settings available for the relevant record carriers and store them, for
example, in the memory 102. Subsequently, one of the available sets of
''
. .. , ,.. .. . . ,. . ~ . ` . ; ............. . -
.
- wo 92/056~8 , ~ PCl/NLg1/00177
22
preferential reproduction seetings in the Imemory 102 is selected in accordance
with a predetermined selection criterion. Preferably, the selection criterion issuch that the highest priority is assigned to the preferential reproduction setting
information in the changeable memory 106, medium priority to the preferential
S reproduction setting information in the non-volatile memory, and the lowest
priority to the preferential reproduction settings on the record carrier. If the unit
- 101 comprises a computer, automatic selection can be realised by loading the
. computer with a suitable selection program.
Now reference is made again to the file OV in Fig. 2, which for
10 all the picture files IPl, ..., IPn comprises a subfile TV/16 containing an
- absolutely coded low-resolution picture. Recording a file OV has the advantage
that an overview of the coded picture information recorded on the record carriercan be obtained with a minimal access time. This is possible, for example, by
- successively by displaying the coded pictures in the subfile TV/16 as
15 representations which wholly or partly fill the display screen, preferably in the
sequence defined by the selected set of preferential reproduction settings.
- However, it is also possible to compose a representation in the forrn of a s~
called mosaic picture from the subfiles, in which mosaic picture a large number
of representations of the coded low-resolution pictures contained in the subfiles
20 TV/16 are arranged in the form of a matrix, preferably in an order dictated by
` the selected set of preferential reproduction settings. By way of illustration Fig.
: 13 shows a mosaic picture 130 made up of the representations (IM#l, IM#3,
- IM#26) of sixteen low-resolution subfile pictures.
Fig. 14 shows an embodiment of the picture retrieval and display
25 system of Fig. lc in more detail. In the present system the picture retrieval and
read unit 11 comprises the read unit 6, a control unit 140 and a picture
processing unit 141. The read unit 6 supplies the information read from the
record carrier to the control unit 140 and to the picture processing unit 141 via a
signal path 142. The control unit 140 then selects specific information contained
30 in the control files BB and IIDB from the information read. The picture
- processing unit 141 selects picture information from the information read and
converts this picture information into a form suitable for the display unit 10. The
,, .
, . - . - - , . . - .. - .. .. . , .. .. . ~ . . . .,.. , . , ~. . . . .. ~
WO 92/OS658 ~ 3 PCI~/NL91/00177
23
read unit 6 and the picture processing lmit 141 are controlled by the control unit
140 on the basis of the data entered by a user, for example via a data entry unit
143, and on the basis of the control data in the control files BB and IIDB.
- In view of the large arnount of inforrnation for every recorded
picture it is preferred to read files cont~ing picture information with a high ~ ~ -
speed, i.e. with a high bit rate, in order to rn~nimi7e the read time per picture
- read. However, this means that the data in the control file is also read with a
high bit rate. The control task is performed by the control unit 140. This control
task requirPs only a limited data processing rate, enabling a simple slow low-cost
. - 10 microcomputer having a low data processing rate to be used for this purpose.
However, in general such a low-cost microcomputer is not capable of processing
the control data which is supplied at a high rate during read-out of the control `
-~ files BB and IIDB. This is because the rate at which the control data is presented
twhich rate is substantially equal to the picture information rate) is too high to
15 enable it to be processed by the slow low-cost computer. This problem can be
mitigated in that every bit group containing control data is recorded n times (nbeing an integer greater than or equal to 2) in succession on the record carrier. A
group of n ~imes repeatedly recorded bit groups will be referred to hereinafter as
a packet. Packets of n identical bit groups are then supplied when the control
- 20 data is read. Fig. 15 by way of example illustrates the manner in which the
control data in the control files BB and IIDB can be supplied by the read unit 6- in the case that n is equal to 2 and the number of bits per bit group is 8.
In Fig. 15 the bit groups bear the reference numeral 150 and the
packets bear the reference numeral lSl. The number of bits per bit group is eight
25 and the number of bit groups per packet is two.
By repeating identical bit groups n times it is achieved that the
rate at which the control data is supplied by the read unit is reduced by a factor
of n without the use of additional auxiliary functions. By a suitable choice of the
value of n it is thus possible to reduce the rate at which the control data is
30 applied to the slow microcomputer system of the control unit 141 to such an
extent that it can be handled by the slow microcomputer system 144. Between the
signal path 145 and the microcomputer system 144 a data extraction circuit 145
,
:~ .
-- -- ~ . .. . . . . . .. . . . . .
, , . ~: . . . ; , : . i.
. ' '.. '' ~ , ' ' ' '
WO 92/0s6~8 ~f~.r ~!6; .~ ~ PCI/NL91/00177
24
can be arranged to supply each of the packets 151 of control data to the
- ~ microcomputer system 144 as one bit group at a rate equal to the bit group
- repetition rate divided by n.
Such a data extraction circuit 145 may comprise, for example, a
5 register 160 (see Fig. 16a) which is loade~ with a cloclc frequency equal to the
bit group repetition rate divided by n. This clock signal can be obtained very
simply by using one bit within each bit gI'OUp 150 as a synchronization bit 152.To the synchronization bits 152 of successive bit groups 150 a logic value may
be assigned which alternates with a frequency related to the repetition rate of the
10 packets 151 of bit groups 150. The alternation frequency may be equal to halfthe repetition rate of the packets (as shown in Fig. 15) or a multiple thereof. This
has the advantage that a clock signal can be used which is derived directly from: the synchronization bits.
The data extraction circuit 145 comprises a clock extraction
15 circuit 161 which supplies an alternating clock signal corresponding to the
alternating logic values of the synchronization bits to a load control input of the
register 160. The register 160 is of a customary type which is loaded with a bitgroup of each packet 151 under control of the clock signal. The clock extractioncircuit 161 also transfers the clock signal to the microcomputer system 144 via
20 the signal line 162. Preferably, the bit groups in the control file are arranged in
so-called frames, which bear the reference numeral 154 in Fig. 15. ln that case it
is desirable that the beginning of each frame 154 can be detected simply. A verysimple detection can be achieved by inserting at the beginning of the frames 154; a plurality of frame synchronization groups 153 with synchronization bits 152
25 which exhibit a predetermined pattern of logic values 150 which differs distinctly
from the possible patterns of logic values of the synchroniziation bits 152 which
can occur in the other packets.
Each frame 154 has a portion 155 containing redundant
information for the purpose of detecting whether the frame has been read-in
30 correctly by the microcomputer 144. An incorrect read-in may be caused, for
example, by a program interrupt, in which the process of reading in the control
data is interrupted in order to carry out another control program. Such a control
wo 92/05658 PCr/NL91/00177
program can be called, for exarnple as a result of the input of data in the dataentry unit 143, in order to fetch the entered data from the data entry unit 143.Since an incorrect read-in of data from the control files BB and IIDB is generally
caused by a prograrn interrupt this requires that the error correction performedS on the basis of the portion 155 is carried out by the microcomputer 144 itself.
The data extraction circuit 145 comprises a frame synchronization detector 163
which detects the beginning of each frame on the basis of the synchronization
:; bits 152 in the frame synchronization bit groups 153. After detection of the
beginning of the frame the frame synchronization detector 163 supplies a
synchronization signal to the microcomputer 144 via a signal line 164. Under
control of the signals received via the signal lines 164 and 165 the
microcomputer 144 reads in the control data available in the register 160 in an,in principle, customary manner. It is to be noted that, in principle, the functions
of the frame synchronization detector 163 and/or the register 160 and/or the
clock extraction circuit 161 can also be performed by the rnicrocomputer lM
itself.
In the above described process of reading in the control data
from the control files BB and lIDB the clock signal for the register 160 is
derived from the synchronization bits 152. However, it is also possible to derive
the clock signal for loading the register 160 from a picture inforrnation clock
~` signal which is usually generated in the picture processing unit 141 for the
purpose of reading in the coded picture inforrnation. This picture information
clock signal has a fixed relationship with the bit group repetition rate in the read-
out picture files and, consequently, with the bit group repetition rate in the
control files BB and IIDB. This is because the control files and picture files have -
been forrnatted- and coded in the same way. Therefore, the clock signal for
loading the register 160 can be derived simply from the picture information clock
signal by means of a frequency dividing circuit.
Fig. 16b shows an example of the data extraction circuit 145,
which employs a frequency divider 165 for deriving the clock signal for the
register 160, which divider derives the clock signal from the picture information
clock signal, which is applied to the frequency divider 165 by the signal
wo 92J0~6S8 ~r~ ,~r~3A J~t3 PCI/NL9l/00177
26
processing unit 141 via a signal line 166. The clock signal for loading the
register 160 must be synchronized with the beginning of the frarnes 154. lllis
can be realized simply by employing a resettable counting circuit for the
frequency divider 165, which counting circuit is reset each time by a reset signal
- S generated upon detection of the beginning of the frames. The reset signal can be
the signal supplied by the frarne synchronization detector 163 via the signal line
164 in response to every detection of the frame synchronization bit groups 153.
In the case that the information in the control files is arranged in
blocks, for example in a manner which is customary with CD-ROM and CD-
10 ROM XA and which will be described hereinafter with reference to Fig. 19, thereset signal for the counter can be derived on the basis of the block
synchronization sections (SYNC) situated at the beginning of each block (BLCK~. -
- However, this requires that the begiMing of each frarne 154 is always situated at
a fixed position relative to the block synchronization section (SYNC). This can
15 be achieved simply by selecting the beginning of each frame IS4 at the beginning
of a block. In the last described method of synchronizing the clock signal for the
register 160 no use is made of the frarne synchron~zation bit groups 153 situated
. at the beginning of each frarne ~54. However, in that case it is also desirable that
the beginning of each frame 154 comprises a number of bit groups not containing
' 20 any control data. Indeed, upon detection of the beginning of each frame the
microcomputer calls a read-in prograrn for controlling the read-in of the applied
control data. However, at this instant the microcomputer may be busy
performing another control task. Such a control task must be interrupted before
the read-in progra n can be called. This interruption of an active control task and
25 the subsequent call for the read-in progra n requires some time. Arranging a
number of bit groups without any control data at the beginning of each frarne 154
ensures with a high reliability that during read-out of the first packet 151 of
`- useful control data in each frame 154 the microcomputer 144 is ready to read in
the control data under control of the read-in program. From the above it is
evident that the synchronization bit groups 153 at the beginning of every frame
may serve a dual purpose, i.e. providing synchronization and realizing a waitingtime until the first useful control data is presented.
WO 92/0~6~8 ~ pcr/NL~l/oo177
; 27
In the case that the bit groups 1S3 are used only for realizing the
waiting time the logic values of the bits ,in these bit groups 153 may assume anarbitrary value.
If the bit groups 153 are also used for synchronization purposes it
S is important that the bit groups 153 exhibit a bit pattern which does not occur in
the other bit groups of the frame 154. For this purpose numerous different
methods are possible, such as for example the use of non-idendcal bit groups in a
packet or the insertion of additional packets without useful control informationbetween the packets of control data. The last-mendoned method may be, for
10 example, to insert packets comprising only bits of the logic value 10" after
every ten packets. When a group of, for example, thirty-two frame
- synchronization bit groups 153 comprising only bits of the logic value "1" is
used, this will guarantee that the pattern formed by the frame synchronization bit
groups 153 does not occur in the other packets of the frame 154.
- 15 Fig. 17 shows an embodiment of the picture storage system 12 in
greater detail. The scanning ùnit 1 in Fig. 17 comprises a scanning element 170
for scanning the image carrier 3 and for conver~ng the scanned picture
information into customary information signals, for example RGB picture signals,- representing the scanned picture. The picture signals at the output of the scanning
20 element define the highest attainable resolution in number of pixels per picture.
The information signals supplied by the scanning element 170 are converted into
- a luminance signal Y and two color-difference signals U and V by means of a
customary matrix circuit 171. A coding circuit 172 converts the signals Y, U andV in a customary manner into absolutely coded signals (for the lower-resolution
25 pictures) and residually coded pictures (for ~he higher-resolution pictures) in
accordance with the coding schemes described hereinbefore. The scanning
element 170, the matrix circuit 171 and the coding circuit 172 are controlled bymeans of a customary control circuit 174 on the basis of control commands
- applied to the control circuit 174 by the control unit 4 via an interface circuit
30 175. The absolutely and residually coded picture information generated by thecoding circuit 172 is applied to the control unit 4 via the interface circuit 175.
The control unit 4 may comprise a computer system comprising a display unit
.:, - . . : , .: ~. - ~ :
wO 92/0~65B PCr/NL91/00177
--9
. 28
176, a computing and storage unit 177 and a data entry unit 178, for example a
- keyboard, for data input by the user. In a customary manner the display unit 176
and the data entry unit 178 are coupled to the computing and storage unit 177.
The computing and storage unit 177 is further coupled to the picture scanning
unit 1 and the recording unit 5 via an interface circuit 179 and 180 respectively.
The recording unit 5 comprises a forrnatting and coding unit 181 which converts
- the information to be recorded, which inforrnation i5 received from the control
- unit via an interface circuit 182, into codes which are suitable for recording and
which are arranged in a format suitable for recording. The data which has thus
been coded and formatted is applied to a write head 183, which records a -
`f' corresponding information pattern on the record carrier 184. The recording
process is controlled by a control circuit 185 on the basis of the control
commands received from the control unit 4 and, if applicable, address
information indicating the position of the write head 183 relative to the record carrier 184.
The storage and control unit 177 is loaded with suitable software
to arrange the residually coded picture information supplied by the scanning unit
- 1 in a customary manner in accordance with the afore-mentioned formatting rules
and to compose the picture files IP and OV. Moreover, the computing and
storage unit 177 has been loaded with software for inserting in the control file, in ~
a customary manner and in accordance with the afore-mentioned formatting ~ ~ -
rules, the preferential reproduction settings input by an operator together withother automatically generated control data, such as for example a list of
addresses at which the various files have been recorded on the record carrier
2~ 184.
The computing and storage unit 177 may further have picture
processing software enabling the scanned picture information to be processed, for
- example for the purpose of error correction, such as for exarnple out-of-focus
correction and grain removal, or for the purpose of color adaptation or brightness
adaptation of the picture.
WO 92/05658 PCI~/NI,9~/00177
. 29
. - The files composed by means of the computing and storage unit
- 177 are applied to the recording unit S in the desired sequence in order to be
. recorded.
-- Very suitable combinations of a record carrier 184 and a
- S recording unit ~ have been described in detail inter alia in European Patent
Applications no. 88203019.0 (PHQ 88.001), 90201309~3 (PHQ 89.016),
. 8900092.8 (PHN 12.398), 8802233.8 (l?HN 12.299~, 8901206.3 (PHN I2.571),
.. -' 90201094.1 (PHN 12.925), 90201582.5 (PHN 12.994), 90200687.3 (PHN
13.148), 90201579.1 (PHN 13.243), and Dutch Patent Applications no. 8902358
10 (PHN 13.088) and 9000327 (PHN 13.242). The record carrier described therein
. is eminently suited for recording information in accordance with a CD format. A
. recording device for recording the files on such record carrier is shown
- diagrammatically in Fig. 18. The shown recording device comprises a formatting
O circuit 186, which composes the information to be recorded, which has been
- 15 applied via the interface circuit 182, in accordance with a formatting scheme,
for example as customary in the so-called CD-ROM or CD-ROM XA system.
By way of illustration this format is shown broadly in Fig. 19. In
accordance with this forrnat the data is arranged in blocks BLCK of a length
corresponding to the length of a subcode frame in the CD signal. Each block
20 BLCK comprises a block synchronizing section SYNC, a header section HEAD -
containing an address in the form of an absolute time code corresponding to the
absolute time code in the subcade portion recorded with the block, and if the
CD-ROM XA format is used the block BLCK further comprises a subheader
section SUBHEAD containing inter alia a file number and a channel number. In
25 addition, each block BLCK comprises a DATA section containing the
information to be recorded. Each block BI,CK may also comprise a section
EDC&ECC containing redundant information for the purpose of error detection
and error corrections. The recording unit 5 shown in Fig. 18 further comprises a :
CIRC coding circuit 187 for interleaving the information and for adding parity
30 codes for the purpose of error detection and error correction (hereinafter also
referred to as error correction codes). The CIRC encoding circuit 187 performs
- the above-mentioned operations upon the formatted information supplied by the
., - .
WO 92/056~8 , r~ c. ~ ~ pcr/NL9l/ool77
:- 30
- formatting circuit 186. After these operations have been performed the
inforrnation is applied to an EFM modulator 188, in which the information is
given a form which lends itself better for recording on the record carrier.
Moreover, the EFM modulator 188 adds subcode information, which includes
- 5 inter alia an absolute time code as address information in the so called subcode Q
channel.
~- Fig. 20 shows an organization of the record carrier in the case
~ that the inforrnation has been recorded in the tracX 20 in accordance with the CD
-~ ` format described above. Parts corresponding to the organization shown in Fig. 2
10 bear the same reference numerals.
The recorded information is preceded by a lead-in section LI
- (also referred to lead-in tracX), as customary in the recording of CD signals, and
is terminated with a customary lead-out section LO (also referred to as lead-outtrack).
When the information is recorded in CD format it is preferred to
include in the control file BB a section recorded in accordance with the CD-I
- standard. These sections are the "Disk Label & Directory", referenced DL, and
the so-called application programs, referenced AF. This enables the recorded
picture information to be displayed by means of a standard CD-I system.
20 Preferably, a subfile FPS with the sets of preferential reprodu~tion settings is
also included in the application program section AF. In addition to the sectionsDL and AT the control file BB comprises a subfile IT comprising a section
CNTR with control data and a section FPS with the sets of preferential
reproduction settings in the format already described with reference to Fig. 15.25 Preferably, the section IT is recorded in a predetermined area on the record
carrier in a section of predeterrnined length. This is in order to simplify retrieval
of the required inforrnation by the microcomputer. If the section IT is no~ large
. enough to accommodate all the control data a part of the control data can be
recorded in a section ITC after the file OV. In that case it is preferred to include
30 a pointer in the section IT' to specify the star~ing address of ITC.
For the case that the information has been recorded in CD format
Fig. 21 shows for the absolutely coded subfile TV such an arrangement of the
:
-; W0 ~2/056~8 PCI/NL91/00177
3 1
picture lines Y01, Y02, ..., Y16 with absolutely coded luminance information
and the picture lines C01, C03, ..., Cl~ with absolutely coded color information,
that successive lines do not adjoin each other in the track direction (also referred
to as tangential direction) and in a direction transverse to the track (also referred
- 5 to as radial direction).
O Fig. 22 shows the positions of the picture lines for the associated
; picture representation. As is shown in Figs. 21 and 22, a number of odd coded
luminance picture lines (Y01, Y03, ..., Y15) with coded luminance information
are recorded in a section comprising the blocks BLCK #1, #2 and #3,
10 subsequently a number of even coded color picture lines (C01, C05, .. , C13)
with coded color information are recorded in a section comprising the blocks
BLCK #4 and #5, then the even coded luminance picture lines (Y02, ..., Y16)
with coded luminance information are recorded in a section comprising the
. ~ blocks ~3LC~C #5, .. , #8, and finally the coded even color picture lines (C03,
15 C07, ..., C15) with coded color information are recorded in a section comprising
the blocks BLCK #8 and #9. The coded picture lines in the bloclcs BLCK#1, ....
BLCK~9 define a conti~uous part of the picture representation shown in Fig. 22.
A group of sections defining a contiguous part of the representation will be
referred to hereinafter as a section group. In a manner similar to that described
20 above, section groups define other contiguous parts of the representation in the
subfile TV. The coded picture lines with picture inforrnation for the subfiles
TV/4 and TV/16 can be arranged in a similar way, as is shown in Figs. 23 and
24.
- This arrangement prevents two or more adjacent picture lines in
25 the representation of the read coded picture from being read incorrectly as aresult of disc defects. The restoration of representations of pictures in which
-; incorrectly read picture lines adjoin each other is very difficult to realize. This is
in contradistinction to the restoration of an incorrectiy read picture line situated
between two properly read picture lines in the representation. In the last-
30 mentioned case restoration is simple by replacing incorrectly read picture lines by
pixels derived *om adjacent picture lines.
. :
: -: : ... ~ .
- WO 92/0s6~8 PCI/N~91/OOi77
32
-. Fig. 25 shows the picture processing unit 141 in greater detail.
.~ The picture processing unit 141 comprises a first detection circuit 250 for
. detecting the synchronization codes LD and the picture line numbers LN
indicating the beginning of each residually coded picture line. A second detection
S circuit 251 serves for detecting the begir~ing of each subfile in each picture file `~
; with a residually coded picture to indicate the beginning of the section IIDB
- conkuning the addresses of a nwnber of coded picture lines. It is to be noted that
. the detection circuits 250 and 251 are needed only for processing the residually
- coded pictures ~nd not for processing absolutely coded pictures. For the purpose
- 10 of these detections inputs of the first and the second detection circuit 250 and 251
are coMected to the signal path 142. A decoding circuit 252 for decoding the
residually coded picture inforrnation and a control circuit 253 for controlling the
-` picture processing operation are connected to the signal path 142. The signal path
142 and outputs of the decoding circuit 252 are connected to data inputs of a
lS picture memory 255 via a multiplex circuit 254, to store the read and decoded
picture information. Data outputs of the picture memory 255 are connected to theinputs of the decoding circuit 252 and to the inputs of the multiplex circuit 254.
The control circuit 253 comprises a memory loading circuit for the generation ofcontrol signals and for controlling the process of loading the memory 255. The
memory loading circuit comprises an address generator for addressing the
memory locations in the picture memory 255. The picture processing unit 141
; further comprises a read control circuit 257 for generating control signals for
controlling the process of fetching information from the memory 255. The read
control circuit 257 further comprises an address generator for addressing the
- 25 memory locations in order to output the content of the picture memory 2SS to a
signal converter 258. The signal converter 258 is of a customary type which
converts the picture information read from the picture memory 255 into a form
suitable for application to the picture display unit 10. The decoding circuit 252
may comprise, for example, a Huffman decoding circuit 261a controlled by the
control unit 253 and an adder circuit 259. The Huffman decoding circuit 261a
decodes the information received via the signal path 142 and subsequently
supplies this decoded information to one of the inputs of the adder circuit 259.
. .
: ~
.- . , .. . , ~ ~ . . . - .' .
: - . ...... . ... . . . ..
- - wo 92/0~658 PCT/1~91/00177
- - .
- 33
- Another input of the adder circuit 259 is connected to the data outputs of the
.- picture memory 255. The result of the adding operation performed by the adder
circuit 259 is applied to the multiplex circuit 2S4. The control circuit 253 is
coupled to the control unit 140 via a control signal path 260. The control circuit
5 253 may comprise, for example, a programmable control and computing unit.
Such a control and computing unit may compAse, for example, a dedicated
hardware unit or a microprocessor system loaded with suitable control software,
by means of which on the basis of control commands received via the control
signal path 260 the load control circuit 256 and the multiplex circuit 254 are
10 controlled in such a way that a selected portion of the picture inforrnation applied
via the signal path 142 is loaded into the picture memory. The information thus
stored in the picture memory 255 is read with the aid of the read control circuit
.. 257 and is subsequently applied to the display unit 10 via the signal converter
258 in order to be displayed.
In Fig. 26 the reference numerals 261, 262, 263 denote picture
representations of the same picture but with different resolutions. The
representation 261 comprises 256 picture lines of 384 pixels each. The
representation 262 comprises 512 picture lines of 768 pixels each and the
representation 263 comprises 1~24 picture lines of 1536 pixels each. The coded
pictures corsesponding to the representations 261, 262 and 263 are included in
consecutive subfiles TV/4, TV and 41V of a picture file IP. The capacity of the
picture memory 255 shown in Fig. 26 is 512 rows of 768 memory locations (also
called memory elements). If a representation should represent the entire coded
picture that subfile is selected from the picture file IP, whose number of pixels
corresponds to the capacity of the picture memory, which in the present case is
the subfile defining the representation 262. This selection can be made on the
basis of the setting data, such as picture numbers and resolution osder (this is the
identification of the subfile resolution), which ase stored at the beginning of each
subfile in, for exarnple, the header HE~AD and the subheader SUBHEAD of the
blocks BLCK. For each subfile this data is read in by the control circuit 253 inresponse to a signal supplied by a block synchronization detector 262a upon
detection of the beginning of each block BLCK.
. . .
-. , : "~ . .. . .
. - WO 92/056~8~ PCT/NL91/00177
. ` . --.
:- 34
In the case that a representation of an absolutely coded picture is
to be reproduced, upon detection of the beginning of the subfile to be selected,- the control circuit sets the multiplex circuit 254 to a state in which the signal
path 142 is connected to the data inputs of the picture memory 255. Moreover,
5 the load control circuit 2S6 is set to a state in which the memory locations are
addressed in synchronism with the re~eption of the information of the successivepixels, in such a way that the inforrnation for the picture lines l1, ..., 1512 is
stored in the respective rows rl, ..., r512 of the memory 255. The picture
information thus loaded into the memory 255 is read out and is converted into a
- 10 form suitable for the display unit 10 by means of the signal converter 258. The
read-out sequence is determined by the sequence in which the read control circuit
257 generates the successive addresses. During normal reproduction this
sequence is such that the memory is read in a row-by-row fashion, starting with
the row rl and starting with column cl within a row. This is possible both in
15 accordance with the interlaced-scan principle and the progressive-scan principle.
In the case of read-out according to the interlaced-scan principle all the odd rows
- of the picture memory 255 are read first and subsequently all the even rows of
the picture memory 255 are read. In the case of read-out in accordance with the
progressive-scan principle all the rows are read in sequence.
A very attractive alternative for the method of storing the picture
information in the picture memory 255 is that in which the picture memory 255
is first filled with picture information from a picture file defining a lower~
- resolution representation of a picture and subsequently the content of the memory
is overwritten with a coded picture defining a higher-resolution representation of
the same picture. In the above example this is possible in that during read-out of : .
each coded pixel from the subfile IV/4 each of a group of 2x2 memory elements
is each time filled with the signal value defined by this coded pixel. This method
is known as the "spatial replica" method. A better picture quality is obtained by
filling only one of the memory elements of the 2x2 matrix with the signal value
defined by a read-out pixe~, and by deriving the other pixels of the 2x2 matrix
from adjacent pixels by means of known inte~polation techniques. This method is
known as the "spatial interpolation" method. After detection of the next subfile
wo 92/056~8 PCI/NL91/00177
J9
; (in the present case TV) the content of ~e picture memory is each time
overwritten with the picture inforrnation of this subfile in the methods described
above. The amount of information in the subfile TV/4 is only a quarter of that in
- the subfile TV. This results in a substantial reduction of the time after which a
- 5 first provisional picture is displayed on the display unit. After read-out of the
picture file TVt4 this low-resolution picture is overwritten vith a representation
of the sarne picture having the desired resolution. As the picture files with coded
- pictures of successive resolutions succeecl one another directly no time is lost in
searching for the subfile TV after read-out of the subfile TV/4.
In the case that a picture is to be rotated the load control circuit
256 is set to a state in which the address sequence of the memory locations is
adapted to suit the cesired rotation angle. Figs. 27b, 27c and 27d illustrate how
- the picture information is stored in the memory for a rotation through an angle of
- 270, 180 and 90 degrees respectively. For the sake of ~larity these Figures only
show the positions of the information of the first two picture lines ll and 12 of
the picture.
In the case that a representation of a small picture is to be
displayed within the outline of a full-scan representation of another picture or, if
desired, the same picture (PIP function), this can be achieved simply by fillingthe desired location of the picture memory 255 with the low-resolution picture of
the subfile TV/4 without magnification. When the picture memory 255 is filled
the load control circuit 2S6 is then set to a state in which the infonnation forthose memory locations in which the small picture is to be stored is addressed.
To illustrate this these memory locations are represented as a frarne 264 in Fig.
26. During the picture processing described above the presence of the low-
resolution picture in the subfile TV/4 again has the advantage that the picture
- information required to perform this function is directly available in the picture
- file IP, so that additional processing is not necessary.
When an enlarged representation of a part of the absolutely coded
: 30 picture is to be clisplayed the information of a part of the pic~ure, for example the
part corresponding to a frame 265, is selected. The information of each pixel ofthe selected part is loaded into every memory location of a group of 2x2 memory
.... .. . . . . . . . .
- ,, - : , : ~ ~
wo 92/~6~8 PCT/NL91/00177
36
locations, so that a magnified full-scan representation of low resolution is
- displayed on the display unit. Instead of repeating each pixel ~x2 times in the
memory the memory may be filled in accordance with the spatial-interpolation
principle mentioned in the foregoing.
S In order to magnify the residually coded pictures the a~ove step
is performed first. Subsequently, the part represented by the frarne 266 is
selected in the subfile 4TV. The part in the frame 266 corresponds to the part
within the frame 265 in the representation 262. The control circuit 253 sets themultiplex circuit 254 to a state in which the output of the residual decoding
10 circuit 252 is connected to the data inputs of the memory 255. The load control
circuit 256 is set to a state in which it addresses the picture memory 255 in
synchronism with the received coded pixels in the sequence in which the
- residually coded picture information from the subfile 4TV becomes available.
The picture inforrnation in the addressed memory locations is applied to the
15 decoding circuit 252 and by means of the adder circuit 259 it is added to theresidual value, after which the information thus adapted is loaded into the
addressed memory location. The part of the picture information recorded on the
.: record carrier corresponding to the f~me 266 is preferably read on the basis of
the information in the control file IIDB. The information in the section nDB is
20 read in by the control circuit 253 in response to a signal from the detector 250.
- Subsequently, the address of that coded picture line is selected from this
information which is situated shortly before the first coded picture line
. corresponding to the picture line in the frame 266. After this, the control circuit -
supplies a command to the control unit 140 via the control signal path 260,
25 which control unit in response to this command initiates a search process in
which the part with the selected coded picture line is located. When this part is ` ~ -
fourid the read-out of the picture information is started and the adaptation of the
content of the memory 255 is started as soon as the part of the first coded picture
Iine which corresponds to the part of the picture within the frame 266 is reached.
30 The detection of this coded picture line is effected on the basis of the line- numbers which together with the line synchronization codes LD have been
; inserted at the beginning of each coded picture line. The control circuit reads in
. ~ .
- ,.,
~ ~ wo 92/0s6s8 PCI/NL9l/00177
3 ~.
37
these line numbers LN in response to a signal from the detector circuit 251. Thestorage of the address information at the beginning of the subfile 4TV enables arapid access to the desired information to be obtained. The detection of the read-
out of the desired residually coded picture lines is simplified by the presence of
5 the line synchronization codes and line numbers in the subfile 4T~.
Fig. 28 shows an embodiment of the read unit 6 by means of
which it is possible to read out the codecl picture information recorded on the
record carrier by means of the recording unit shown in Fig. 18. The shown read -~
unit 6 comprises a customary read head 280 which reads the inforrnation patternson the record carrier 184 by scanning the track 20 and converts the resulting
; ~ information into corresponding signals. The read unit further comprises a
- customary positioning unit 284 for moving the read head 280 in a direction
transverse to the tracks to a portion of the track 20 specified by a selected
- address. The movement of the read head 283 is controlled by a control unit 285.
The signals converted by the read head 280 are decoded by ~n EFM decoding
circuit 281 and are subsequently applied to a CIRC decoding circuit 282. The
CIRC decoding circuit 282 is of a customary type, which restores the original
-. structure of the information which has been interleaved prior to recording and ,
~ which detects and, if possible, corrects incorrectly read codes. Upon detection of
incorrigible errors the CIRC decoding unit supplies a new error flag signal. Theinformation which has been restored and corrected by the CIRC decoding circuit
282 is applied to a deformatting circuit 283 which removes the additional
information added by the formatting circuit 186 prior to recording. The EFM
demodulating circuit 281, the CIRC decoding circuit 282, and the defonnatting
`. 25 circuit 283 are controlled in a customary manner by the control unit 285. The
information supplied by the deformatting circuit 283 is applied via an interface
circuit 286. The deformatting circuit may comprise an error correction circuit by
means of which errors which cannot be corrected by the CIRC decoding circuit
can be detected and corrected. This is effected by means of redundant
information EDC & ECC added by the ~ormatting circuit 166. The error
~- correction circuit, which is comparatively complex and therefore comparatively
expensive, is not necessary. This is because the effects of erroneously read codes
. .
WO 92/0s658 PCI~/NL91/ûû177
. '. .
3 8
in the absolutely coded picture information can be masked simply by replacin~
the incorrectly read coded pixels and/or a complete coded picture line by picture
information derived from one or more adjacent coded pixels or adjacent c~ded
picture lines. Such a correction can be effected simply by means of the signal
processing unit 141 shown in Fig. 25, by programming the control circuit 253 so
as to be responsive to the error flag signal supplied by the CIRC decoding circuit
282 to control the load control circuit 256 in such a way that the information of
an adjacent pixel is read and, at the same time, the multiplex circuit 254 is set to
: a state in which the data outputs of the picture memory 255 are connected to the
data inputs. Subsequently, the load control circuit 256 is reset to its previousstate and instead of the incorrectly read coded pixel the information read from
: the picture memory 255 is stored at the addressed memory location.
- In the case that a residu~lly coded picture is read the value in the
memory 255 is not adapted upon detection of an incorrectly read residual value
but remains unchanged. This can be achieved, for example, by causing the
- control circuit to generate a signal which inhibits writing into the memory 255 .
when the erroneous residual value is applied.
The capacity of the picture memory 255 is large, so that the cost
price of such a memory is comparatively high. The memory capacity may be
reduced by arranging between the multiplexer 254 and the picture memory 255 a
sample rate converter 290 of a customa~y type, which reduces the number of
pixels per line from, for example, 786 to 512.
Fig. 31 shows an example of the sample rate converter 290. The
present example comprises a series arrangement of an upsampling and
- 25 interpolation circuit 310 and a low-pass filter 311, and a downsampling and
decimating circuit 312.
The use of the sample rate converter 290 enables a memory of,
for example, 512 by 512 memory locations to be employed. Since for practical
reasons the number of rows and the number of columns of memory locations in a
memory are preferably powers of two, this yields a memory of particularly
salisfactory dimensions. Moreover, as a result of the reduction of the number of ~
memory locations to 512 per row the required memory read-out frequency is - -;
.;
.. .
WO 92/OS658 ~ 3 PC~/NL91J00177
39
reduced by 2/3, so that less stringent requirements have to be imposed on the
read-out specd of the memories and the read control circuit which are used.
The usually employed picture tubes have a maximum resolution
. corresponding to approximately 5 MHz, which corresponds to approximately 500
.- 5 pixels per line, so that the reduction of the number of memory locations per row
- -: has no visible effects on the reproduced picture.
- The use of the sarnple rate converter is also advantageous when
portrait-format representations of pictures are to be displayed on a display
screen, which will be explained hereinafter with reference to Figs. 30a, 30b, 30c
- 10 and 30d.
In Fig. 30a the reference numeral 300 refers to the dimensions of
a picture in accordance with the PAL TV standard. Such a picture in accordance
with the PAL TV standard comprises 575 useful picture lines. During,
reproduction of the information in the picture memory 255 of 512x512 mernory
15 elements 512 of these 575 useful picture lines are utilized. This means that a
- representation 301 of the coded picture in the picture memory fits completely
within the aspect ratio of the frame 300 as defined by the PAL TV standard, onlya small part of the available display screen area being left unused.
.;~ ln Fig. 30b the reference numeral 302 denotes a frame having the
20 dimensions of a picture in accordance with the NTSC TV standard. Such a
picture in conformity with the NTSC TV standard comprises 485 useful lines.
This means that only a limited part of a representation 303 of the coded picturepresent in the picture memory 255 falls outside the outline of a picture in
accordance with the NTSC standard.
Figs. 30a and 30b concern landscape-format reproductions of
representations of coded pictures. However, if portrait-format representations of
coded pictures are required the problem arises that the heighl of the picture
corresponds to 768 memory locations, the number of useful picture lines being
- 575 in accordance with the PAL TV standard and 485 in accordance with the
30 NTSC TV standard. When a picture memory of 512 rows of 768 (or 1024)
memory locations is employed without the use of the sample rate converter 290
this would mean that the picture is not filled correctly (see Fig. 5). However, by
: . - ~ .... .
WO 92/05~58 PCI/NL91Jo0177
r. ~ - 9
-` 40 -
the use of the sarnple rate converter 290 it is achieved that the coded picture lines
of 768 coded pixels are converted into coded picture lines of 512 coded pixels,
so that the picture is filled correctly (se~e Figs. 30c and 30d). This means that
during reproduction the height of the representation of the picture stored in the
S memory 255 substantially corresponds to the height of the picture frames defined
in the PAL and NTSC TV standards.
In order to ensure that the ratio between the length and width of
- the representation of the coded picture stored in the picture memory 255 ~. .
corresponds to the original ratio it is required to fill only 256 of the 512 columns
of the picture memory with picture information. This is possible, for example, by ~ -.
stonng only the even or only the odd coded picture lines in the memory 255.
However, other methods utilizing interpolation techniques may al~o be used.
The method of reducing the number of columns in the picture
memory employing interpolation techniques yields a picture representation of
15 satisfactory qu~lity. This is in contradistinction to the method in which only a
part of the coded picture lines is stored in the columns of the picture memory. ~ -
- An example of a method which in a simple way yields representations of
satisfactory quality will be described below for a picture memory comprise
512x512 memory locations.
This method uses the subfile TV/4 with 384x256 coded pixels,
- instead of the subfile TV with 768x512 coded pixels, for loading the picture
- memory. -:
The use of a sample rate converter 20, by means of which the
number of pixels per read coded picture line can be reduced and increased,
: 25 enables the number of pixels per read coded picture line of the subfile TV/4 to
be increased from 384 to 512. The 256 available adapted picture lines of 512 -
coded pixels each are loaded into the memory 255. Thus, 256 columns of 512
memory locations each are filled with picture information. Reading out this
information yields an undistorted portrait-format representation, whose height
substantially corresponds' to the height of the display screen of a PAL or NTSC -
TV system, and whose quality is substantially better than that of a portrait-format
representation obtained on the basis of a coded picture of 768x512 coded pixels
- w0 92/~s6s8 ~ 3 PCI/NL91/00177
41
. whose width is adapted by using only half (256) the available number of 512
` coded picture lines.
By way of illustration Fig. 30c shows a portrait-fDrrnat
representation 304 of the stored coded picture (of 256x512 coded pixels) thus
S obtained within the frame 300 defined by the PAL TV standard. The entire
representation falls within the frame defined by the PAL standard. Fig. 30d by
way of illustration shows a portrait-formlat representation 305 of the coded
picture thus stored. The representation falls largely within the frarne 302 defined
by the NTSC TV standard.
As will be apparent from the foregoing the use of a sample rate
` converter 290 enables the use of a picture memory having equal numbers of rows
and columns and corresponding substantially to the number of useful picture lines
in accordance with the NTSC or PAL standard. This means that both in the case
of portrait-format and landscape-format representations of coded pictures the
height of the representation substantially corresponds to the number of useful
.- picture lines, so that the display screen will be filled correctly for representations
of both types.
An advantageous use of the sam~ple rate converter 290 will now
be described in detail with reference to Figs. 32 to 40. Fig. 32 shows
diagrammatically an embodiment of a read device 320 in accordance with the
invention, in which all the parts which are irrelevant are not shown or are shown
in broken lines, such as the decoding circuit 25~. In this Figure parts
corresponding to those already described bear the sarne reference numerals. In
` the Figure the combination of the load control circuit 256, the memory 255 and
the read control circuit 257 bears the reference numeral 231. Hereinafter, this
combination will be referred to as the picture memory unit 321.
The operation of the read device will be described in detail with
`- - reference to Fig. 33 for the case that a landscape-format representation of the
coded picture IV is to be reproduced on the display unit 10. The coded picture
TV represents 512 lines of 768 pixels each. In Fig. 33 a picture of such
dimensions bears the reference numeral 330. The ratio between the dimension of
` the pixel in the horizontal direction and the dimension of the pixel in the vertical
.
' ' '' ': :~: ', ~
, ~ : ' '~:: ` ',', ' . ::,
WO 92/056~8 pcr/NL9l/ool77
~ t? ~? `^ ~
; 42
direction is unity. Hereinafter, pixels with such a ratio between the hori ontal: and vertical dimensions will bear the reference letter p. A picture made up of
512 lines of 768 pixels is referred to as ,a picture having the dimensions 512.1 x
768.p. By means of the sample rate converter 290 the read~ut coded picture TV
is converted into a coded picture representirlg a picture 331 of 512 lines of 512
pixels p' each. The sample rate converter is then set to adapt the sample rate by
a factor of 2/3. The ratio between the dimension in the horizontal direction andthe dimension in the vertical direction for the pixel p' is 3t20 This means that a
group 332 of three horizontally adjacent pixels represented by the coded picture. 10 TV corresponds to a group 333 of two horizontaIly adjacent pixels p' in ehe
' picture represented by the coded picture supplied by the sarnple rate converter.
- The coded picture on the output of the sarnple rate converter thus represents a
picture made up of 512 lines of 512 pixels p' each. For the storage of such a
coded picture in the memory 255 a storage capacity of 512 x 512 memory
15 locations is adequate. The coded picture lines are stored in rows in the memory
255. The picture memory 255 is also read in a row-by-row fashion. It will be ~
obvious that instead of a row-by-row read-in and read-out of the memory the '
memory can also be read in and read out in a column-by-column fashion. An
essential feature is, however, that the read-out direction of the memory
; ~ 20 corresponds to the read-in direction. The read device shown in Fig. 33 has the
advantage that a so-called square memory may be used, i.e. the memory ;
locations can be addressed as mat~ix elements of a matrix having a number of
rows equal to the number of columns. Such square memories can be realised
simply by means of commercially available integrated memory circuits. The read
25 device further has the advantage that the speed with which the coded pixels are
fetched from the memory 255 is only 2/3 of the fetching speed in the case of a
memory storing a coded picture of 512 x 768 coded pixels. The reduction of the
number of pixels per picture line from 768 to 512 does not adversely affect the
picture quality during the reproduction of a representation of the picture on a
3Q display unit in the form of a standard TV set. This is because exisbng TV sets
are adapted to display video signals with a bandwidth of approximately 4 to 5
MHz. Such a bandwidth corresponds to approximately 4Q0 to 50Q pixels per line.
.. , :
wo 92J0~658 PCr/NL91/OQ177
-- 43
: The use of picture lines comprising 768 pixels per picture line would result in a
bandwidth of approximately 7.5 MHz. 1[he high-frequency content in the part of
the band from 4-5 MHz to 7.5 MHz is not reproduced on a standard TV set.
How an enlarged representation of a part of the recorded coded
5 picture can be obtained will be explained with reference to Fig. 34. In Fig. 34
the picture represented by the recorded c oded picture bears the reference numeral
340. A part 341 theret~f having a height equal to half the height of the entire
- picture and having a width e~ual to half the width of the entire picture is to be
displayed on ~he display unit 10 enlarged linearly by a factor of two. During
10 read-out of the coded picture TV only those 256 picture lines are read which
contain inforrnation of the picture to be displayed. The part of the picture
- represented by these 256 coded picture lines bears the reference numeral 342.
The part 342is a picture comprising 256 lines of768 pixels p each. The read-out
coded picture lines at the output of the read unit are applied to the sample-rate
15 converter 290, which has been set to a state in which it converts a picture line of
768 coded pixels into a coded picture line of1024 coded pixels. These coded
pixels represent pixels (p" in Fig. 34) whose ratio between the dimension in thehorizontal direction and the dimension in the vertical direction is 3/4. The 256coded picture lines thus obtained are stored in the picture memory, every line
being stored two times. In this way a coded picture is stored which comprises
512 picture lines, of which two lines are identical each time. It is to be noted that
the number of picture lines can also be increased by means of interpolation
techniques instead of picture line duplication. The coded pixels in the coded
picture lines again represent a pixel (p') whose ratio between the horizontal
25 dimension and the vertical dimension is 3/2. A representation of the coded
picture thus stored can be displayed on the display unit 10 in the same way as
described with reference to Fig. 33. The above method of displaying an enlarged
picture employs the absolutely coded picture IV instead of the residually coded
- picture 41V as the basis for the coded picture stored in the picture memory 255.
30 This has the advantage that no additional de~ding circuit is required for
decoding the residually coded picture. Moreover, the time required for reading-
- out the picture information is shorter. During reproduction of the enlarged
.
WO 92/056~8 pcr/NL9l/ool77
, - .
3 '--~3
; . representation of a part of the coded picture on a standard TV set the horizontal
- resolution of the TV set corresponds to the number of pixels per picture line.
The vertical resolution of the coded picture is not adapted to an optimum extent'. to the TV set. This is because the TV set displays pairs of identical picture lines.
- 5 In practice this is found to be hardly perceptible. In the foregoing the coded
picture lines being read are wholly applied to the sample rate converter and only
half the adapted picture line is stored in the memory. However, it is alternatively
possible to apply only the desired half of the picture line being read to the
- sample rate converter.
lD The reproduction of an enlarged representation of a part of a
coded picture will now be described with reference to Fig. 35 in the case that
- this is based on the residually coded picture 4TV. The read unit 6 reads 512 of
the 1024 available residually coded lines (1') of 41'V from the record carrier 184.
The coded picture lines thus read represent a part of the picture comprising thelS enlarged representation to be displayed. The decoding circuit converts the
- residually coded picture lines being read into series of adapted coded pixels each
representing a pixel p. The sample rate converter converts said series of pixelsinto series of coded pixels each representing a pixel p'. The memory 255 is
loaded with 512 lines each comprising 512 adapted coded pixels. A
; 20 representation of the coded picture stored in the picture memory can be displayed
on the display unit 10. The method described above has the advantage that the
resolution (in a vertical direction) of the coded picture is higher than that
obtained with the method of producing an enlarged representation on the basis ofthe ccded picture TV. In the foregoing the read-out coded picture lines are
wholly applied to the sarnple rate converter and only half the adapted picture line
is stored in the memory. However, it is alternatively possible to apply only thedesired half of the picture line being read to the sample rate converter.
- The reproduction of a representation of a coded picture TV/4
rotated through 90 degrees on a standard TV set will now be described with
- 30 reference to Fig. 36. The read unit now reads the coded picture TV/4. This
coded picture represents a picture of 256 lines (1) of 384 pixels (p) each. The
sample rate converter 290 converts the series of read-out coded pixels into series
.
, ,:. .. .~. ... . .. .. .. ~ . . , .. .. , ,;
wo 92/0565B ~_ ~ ~f~f~ PCT/r~lLgl/00177
of coded pixels forming a picture of Sl~ lines of 576 pixels p; For this purpofse
the sample rate converter is set to fadapt the sarnple rate by a factor of 2/3. The
- pixels p exhibit a rado of 2/3 between the horizontal dimension and the vertical
dimension. The 256 coded picture lines of 576 coded pixels each at the output ofthe sarnple rate converter 290 are stored in 256 columns of the picture memory
. . 255 instead of in rows as in the reproduction of non-rotated representations.
Since the number of memory locations in each column is only 512 it is not
possible to store all the coded pixels. The coded pixels at the beg~nning and/orthe end of the coded picture line are not stored. When the memory is now read
row by row in the same way as described with reference to Figs. 33, 34 and 35 a
coded picture is read which represents a picture rotated through 90 degrees and
comprising 512 lines of 512 pixels p'. As only 256 of the available S12 columns
in the memory contain coded pixels (preferably those columns in the central part
' t of the memory) the representation will exhibit two areas not containing any
lS essential picture information. In Fig. 36 these areas on the display unit 10 bear
the reference numerals 361 and 362. The method described above has the
advantage that a rotated representation of a coded picture is obtained in a simple
manner, without loss of picture quality and with a correctly filled display screen.
Another method of reproducing a rotated representation on the
basis of the coded picture TV will now be described with reference to Fig. 37.
The coded picture TV (512.1 x 768.p) is read by the read unit 6. The sample
rate converter 290 converts the read-out coded picture lines of 768 coded pixelseach into coded picture lines of 576 adapted coded pixels each. For this purposethe sample rate converter 290 is set to adapt the sample rate by a factor of 3/4.
- 25 The adapted coded pixels represent a pixel p whose ratio between the
horizontal dimension and the vertical dimension is 4/3. Thfe coded picture thus
obtained is converted into a coded picture of 256 lines of 576 coded pixels each.
This can be achieved by deriving one new coded picture line from pairs of
consecutive lines. This is possible, for example, by suppressing one picture line
30 of every pair. The coded picture thus obtained represents a picture comprisin~
256 picture lines of 576 pixels (p ) each. The coded picture lines of the coded
picture thus obtained are stored in columns of the picture memory 255.
WO 92/056~8 PCTtNL91/00177
?~-~
,
46
~ Subsequently, this memory is read in a row by row fashion, yielding a coded
; picture whose coded pixels represent a pixel p'.
The reproduction of an enlarged and rotated representation of a
part of a coded picture will now be described with reference to Fig. 38. The
S coded picture is the coded picture TV recorded on the record carrier 184. Theread unit 6 reads 256 of the available coded picture lines of TV. The read-out
coded picture lines of 384 coded pixels each are converted into a coded picture
- of 256 lines of 576 coded pixels ~ach by means of the sarnple rate converter.
` These pixels represent pixels p . Of this coded picture 512 coded pixels of each
- lO line are loaded into the picture memory 255 in a column by column fashion. The
memory is loaded column by column to obtain a coded picture representing a
:- picture which has been rotated through an angle of 90 degrees. Each coded pixel
represents a pixel p'. In the foregoing the read-out coded pictut~e lines are wholly
~ applied to the sample rate converter and only half the adapted picture line is
- 15 stored in the memory. However, it is also possible to apply only the desired half
- the read-out picture line to the sarnple rate converter.
A method of obtaining an enlarged and rotated representation on
the basis of the coded picture 4TV will be described with reference to Fig~ 39~
The read unit 6 now reads out 512 of the available 1024 residually coded picture20 lines. By means of the decoding circuit 252 these read-out residually coded
picture lines are converted into series of coded pixels each representing a pixel
p'. By means of the sample rate converter 290 these series of coded pixels are
converted into series of adapted coded pixels each representing a pixel p . For
- this purpose the sarnple rate converter 290 is set to adapt the sample rate by a
- 25 factor of 3/4. Subsequently, pairs of consecutive coded picture lines are
converted into one coded picture line, for exarnple by suppressing one picture
^ line of every pair. This operation results in a coded picture in which each coded
pixel represents a pixel p . Of every coded picture line of this coded picture 512
- coded pixels, corresponding to the part to be displayed, are loaded into the
30 columns of the picture memory 255. Subsequently, the picture memory is read
row by row, yielding a coded picture of which each coded pixel corresponds to a
pixel p'.
'
:,, ~, .. .. , . - ~ -
WO 92/05658 ~ 3 ~J~ pcr/NL9l/oo177
47
- . The above shows that it is advantageous to use the sample rate
conYerter 290 in combination with the read-out of absolutely coded pictures and
also with residually coded pictures. However, its use is particularly attractive in
.-- combination with the read-out of the absolutely coded pictures TV and TV/4. By
- 5 means of a memory having 512 x 512 memory locations and a sarnple rate
- converter it is then simply possible to display a normal, an enlarged, a rotated,
and an enlarged and rotated representation with satisfactory picture quality on a
standard TV set. Since the method is based on absolutely coded pictures TV and
F TV/4 a decoding circuit for decoding the residually coded picture is not needed.
10 The sample rate converter should then merely be capable of adapting the sarnple
rate by factors of 2/3, 3/2, 4/3 and 3/4.
In the above embodiments for the rotated reproduction of a
representation of a coded picture the rotation is obtained by storing coded picture
lines column by column in the picture memory and subsequently reading the
15 stored picture out of the memory in a row by row fashion. However, this
- rotation can also be obtained by storing the pictures row by row and extracting
them from the memory in a column by column fashion. An essential feature is
that the ratio between the vertical and the horizontal dimension of the pLl~el (p')
representing the coded pixel supplied to the signal converter 258 by the memory
20 unit 321 is the same both for rotated and non-rotated pictures.
:' .
:, :
