3D Interface Specifications White Paper
Contents
1. crc32 h header file for crc32 checksum function prototypes unsigned long CalcCRC32 unsigned char p unsigned long len ifdef _ cplusplus 3OROS2H CrO32 Philips 3D Solutions Eindhoven CRC32 implementation adapted from public domain code written by Eric Durbin Original header follows START OF ORIGINAL HEADER Qre32 6 C implementation of CRC 32 checksums for NAACCR records Code is based upon and utilizes algorithm published by Ross Williams This file contains CRC lookup table function CalcCRC32 for calculating CRC 32 checksum function AssignCRC32 for assigning CRC 32 in NAACCR record 15 February 2008 26 of 29 02005 2008 Philips Electronics Nederland B V include 3D Interface Specifications function CheckCRC32 for checking CRC 32 in NAACCR record Provided by Eric Durbin Kentucky Cancer Registry University of Kentucky October 14 1998 Status Public Domain END OF ORIGINAL HEADER 3D Solutions cro32 h CRC LOOKUP TABLE The following CRC lookup table was generated automagically by the Rocksoft tm Model CRC Algorithm Table Generation Program V12. 2D plus Depth format is very flexible Among the advantages of using 2D and Depth images compared to for example 9 separate views are e user control over the amount of perceived depth controlled via the header and or I2C e future compatibility with displays which have a different number of views than the current 9 or have a different lens design which requires different interleaving see Section 7 2 allowing the display to render in a way optimized for its optical components e suitability for compression for content storage and distribution e advantages in content editing because the disparity is explicit More information can be found in 3DTVC Note that whenever 2D and Depth is mentioned the Depth map describes the disparity and not the depth The picture below shows how the header and the 2D and Depth sub images are organized in a frame The format contains the following data e Header see sections 5 and 6 e 2D sub image left e Depth sub image right A Depth map in grey scale picture This depth map belongs to the 2D sub image Figure 1 Overview of 3D frame in the WOWvx 2D plus Depth format The 2D and Depth sub images are positioned besides each other which result in a total line width equal to the native panel line width Between each video line a blank line is added This doubles the vertical resolution to equal the native panel vertical resolution Sections 4 3 and 4 4 explain the 2D and Depth sub image
3. 3D Interface Specifications White Paper Philips 3D Solutions Document Information Title 3D Interface Specifications White Paper Date 15 February 2008 Security The attached material and the information contained herein are proprietary to Philips 3D Solutions Copying reproduction adaptation modification or dissemination in whole or part is not permitted without written permission from Philips 3D Solutions Contact address Philips 3D Solutions High Tech Campus 27 5656 AE Eindhoven The Netherlands E mail 3DSolutions philips com Website www philips com 3dsolutions 3D Interface Specifications 3D Solutions Table of Contents 1 litre DECIDES rapere arena EH NEUES ro topi icd peti tetto tav 5 MEE autor NEMORE 5 E SEE i a EROR 6 E B co E E E EE paninenenuadeno et 6 2 Physical Interface eheu E dM Mute dau ee 7 3 PES Play MOO E E 8 4 DE MUN ES 9 41 WOWyx 2D plus Depth format cesessecssessessssneessessesnscnseseesscsssesecsessssuseneessesueessscuscuceasesuscuecueeasenuscuecaeeasensecueese 9 42 Declipse 9 me E 4A E EE T E Er E 5 Video data iriterface s 5 esse iiare esas aeo ariela aserat ear aa aaa ia a ae e eaa A beoe 12 6 Header eee 15 6 1 Header locat
4. vx 2D plus Depth format Section 4 explains the 2D and Depth sub images in more detail From now on we continue only with the explanation of 42 1920x540 2 images of 960x540 The reader interested in 20 display has to read 1600x600 2 images of 800x600 The rest stays the same More information can be found in 3DTVC Note that whenever Depth is mentioned the Depth Z map describes the disparity and not the depth Figure 5 shows how the header and the 2D and Depth sub images are organized in a frame of 1920x1080 pixels in case of the WOWvx Declipse format 1920 NO oc N 0 959 960 1919 960 960 Figure 5 3D frame layout for the WOWvx Declipse format The header is located in the upper left corner i e the first part of the first video data line The function of the header is twofold When the display detects the header it switches to 3D mode and the header contains settings for rendering processing The header is further explained in section 6 Furthermore a 3D frame contains 2D and Depth sub images both with a resolution of half the native panel resolution in both horizontal and vertical direction The 2D sub images are positioned in the left half of the The video data interface format towards the Philips 3D displays is different from the storage format cf section 4 The main reason to store content in a format with four spatially separated quadrants is to have a higher coding efficiency 15 February
5. 0 using the following model parameters Width 4 bytes Poly Ox04C11DB7L Reverse FALSE For more information on the Rocksoft tm Model CRC Algorithm see the document titled A Painless Guide to CRC Error Detection Algorithms by Ross Williams ross guest adelaid d ats s This document is likely to be in the FTP archive ftp adelaide edu au pub rocksoft ay v br ur oy ay Ef g unsigned long crctable 256 0x00000000L Ox04C11DB7L 0x09823B6EL 0 004326091 0x130476DCL Ox17C56B6BL 0x1A864DB2L Ox1E475005L 0x2608bEDB8L 0x22C9F00FL Ox2F8AD6D6L Ox2B4BCB61L 0x350C9B64L Ox31CD86D3L 0x384FBDBDL Ox4C11DB70L Ox48DOC6C7L 0x4593EO1EL 0x4152FDA9L Ox5F15ADACL Ox5BDA4BOIBL 0x569796C2L 0x52568B75L 0x6A1936C8L Ox6ED82B7FL 0x639BODA6L 0x675A1011L 0x791D4014L Ox7DDC5DA3L 0x709F7B7AL 0x745E66CDL 0x9823B6E0L Ox9CE2AB57L Ox91A18D8EL 0x95609039L 0x8B27CO3CL Ox8FE6DD8BL 0x82A5FB52L 0x8664E6E5L OxBE2B5B58L OxBAEA46EFL 0xB7A96036L 0xB3687D81L OxAD2F2D84L OxA9EE3033L OxA4AD16EAL OxAO6COBSDL 0xD4326D90L OxDOF370
6. 2008 12 of 29 2005 2008 Philips Electronics Nederland B V 3D Interface Specifications 3D Solutions frame and Depth sub images in the right The 2D and Depth sub images of 960x540 are positioned besides each other which result in a total line width of 1920 pixels For the WOWYNvx Declipse format the 2D and Depth sub images of the top and bottom background data half of the format are line interleaved i e every second line is filled with the corresponding data from the bottom half of the format For the WOWvx 2D plus Depth format a blank or rather don t care line is inserted below each line of the 2D and Depth This doubles the vertical resolution from 540 to 1080 This is illustrated in Figure 6 1920 o LLL 2 4 SE 1080 1079 0 959 960 1919 960 960 Figure 6 3D frame layout for the WOWvx 2D plus Depth format To illustrate the difference between the storage format and the video data interface format some concrete examples are given in Figure 7 for the different WOW vx formats See section 6 2 9 for more information on the two different Declipse modes 15 February 2008 2005 2008 Philips Electronics Nederland B V 13 of 29 3D Interface Specifications 3D Solutions 2D plus Depth Declipse removed redundant data Declipse full background data Figure 7 Difference between the storage format and the video data interface format 15 Februa
7. image example 4 4 Depth sub images A Depth sub image has a resolution of half the native panel resolution in both horizontal and vertical direction It contains disparity values with a range of 0 to 255 where a value of 0 corresponds with objects located with a maximum disparity behind the screen and 255 corresponds with objects located closest to the observer An implementation is that the Depth image is a black and white image This means that the R G and B sub pixels have the same value per pixel However the display only uses the red sub pixels The green and blue sub pixels are discarded Note that a Depth sub image actually contains disparity values Figure 4 Depth sub image example 15 February 2008 2005 2008 Philips Electronics Nederland B V of 29 3D Interface Specifications 3D Solutions 5 Video data interface The resolution of a frame that is sent to the display via the DVI cable has a resolution of 1920x1080 for 42 or 1600x1200 for 20 A frame contains the following data e Header e 2D sub image with a resolution 960x540 for 42 or 800x600 for 20 e Depth sub image with a resolution 960x540 for 42 or 800x600 for 20 e Background 2D sub image with a resolution 960x540 for 42 or 800x600 for 20 e Background Depth sub image with a resolution 960x540 for 42 or 800x600 for 20 The latter two sub images contain data in the case of the WOWvx Declipse format while they are blank in the case of the WOW
8. 110 Reserved etc 6 2 3 Hdr_Factor Table 6 Header factor value Percentage of the display recommended depth value AA actor Factor 64 64 0 255 Each 3D Display has a Display recommended depth value which corresponds to an acceptable maximum depth factor value for that specific type of display This value strongly depends on the lens design The factor field in the header contains the percentage to be used from the display recommended depth value The value of 64 corresponds with the 10076 of the display recommended depth value It is allowed to use values higher than 64 The factor works on a linear scale and is multiplied with the factor controlled by the user in the Display Control Tool 6 24 Offset Table 7 Header offset value Amount of range behind the screen 0 Rangeis shifted in the direction of the viewer 128 Range is equally divided in front and behind the screen 255 Range is shifted away from the viewer H 3 Offset 128 0 255 Values in the Depth map equal to the header offset value will be located on the plane of the display All values in the disparity map with a higher value will de displayed in front of the display Offset CC is the offset controlled by the Content Creator In the system there is also an Offset user present which is controlled by the user using the Display Control Tool 6 25 Select Table 8 Header select H 4 Hdr Facto
9. 27L OxDDBO56FEL 0 097148491 0xC7361BA4CL OxC3F706FBL OxCEB42022L OxCA753D95L OxF23A8028L OxF6FB9D9FL OxFBB8BB46L OxFF79AG6FI1L OxEl3EF6FA4L OxE5FFEBA43L OxE8BCCD9AL OxEC7DDO2DL 0x34867077L 0x30476DCOL 0Ox3D044B19L 0x39C556AEL 02005 2008 Philips Electronics Nederland B V 15 February 2008 27 of 29 3D Interface Specifications 3D Solutions 0x278206ABL 0x23431B1CL 0x2E003DC5L 0x2AC12072L 0Ox128bE9DCFL 0x164F8078L 11 Ox1FCDBB16L OxO18AEB13L Ox054BF6A4L 0x0808D07DL OxOCCO9CDCAL Ox7897AB07L Ox7C56B6BOL 0x71159069L Ox75D48DDEL 0x6B93DDDBL O0x6F52CO06CL 0x6211E6B5L 0x66DOFBO2L Ox5E9FA6BFL 0x5A5E5BO8L 0x571D7DD1L Ox53DC6066L 0x4D9B3063L 0x495A2DDA4L 0x44190BODL 0x40D816BAL OxACA5C697L OxA864DB20L 0xA527FDF9L OxAIEG6EO04EL OxBFA1BOABL OxBB60ADFCL 0xB6238B25L 0 2 296921 Ox8AAD2B2FL 0x8E6C3698L 0x832F1041L Ox87EEODF6L 0x99A95DF3L 0x9D684044L 0x902B669DL O0Ox94EA7B2AL OxEOBAIDE7L 0xE4750050L 0
10. L 0xC8EAO00AO0L OxD6AD50A5L OxD26C4D12L OxDF2F6BCBL OxDBEE767CL OxE3A1CBCI1L OxE760D676L OxEA23FOAFL OxEEE2ED18L OxFOA5BDIDL OxF464A0AAL OxF9278673L OxFDE69BC4L Ox89B8FD09L 0x8D79EOBEL 0x803AC667L Ox84FBDBDOL Ox9ABC8BD5L 0 9 7096621 0x933EBOBBL Ox97FFADOCL OxAFBO10B1L OxAB710D06L 0xA6322BDFL 0xA2F33668L OxBCB4666DL 0xB8757BDAL 0xB5365D03L OxB1F740B4L KOR KKK KKK IK RK I KI I I A A A OR KK RK KK End of CRC Lookup Table Calculate CRC 32 Checksum oy Uses reflected table driven method documented by Ross Williams 15 February 2008 28 of 29 2005 2008 Philips Electronics Nederland B V unsigned long CalcCRC32 unsigned char p unsigned long len 3D Interface Specifications 3D Solutions PARAMETERS unsigned char p Buffer containing data over which to calculate the CRC unsigned long len Length of data in the buffer RETURNS checksum value Author Eric Durbin 1998 10 14 Modified by Philips 3D Solutions 2007 10 11 Status Public Domain unsigned long i unsigned long crc 0 process each byte prior to checksum field for i 0 i lt len i crc crcetable crce gt gt 24 cre lt lt 8 return cro 15 February 2008 2005 2008 Philips Electronics Nederland B V 29 of 29
11. agement Signaling DPMS standrad version 1 0 revision 1 0 Augustus 20 1993 Digital Visual Interface DVI Digital Display Working Group Revision 1 0 02 april 1999 VESA enhanced extended display identification data standard Release A Revision 1 February 9 2000 VESA Generalized timing formula GTF standard version 1 0 revision 1 0 December 18 1996 The I2C bus specification Version 2 1 January 2000 Philips Semiconductors 3D throughout the video chain by B Barenbrug proceedings of ICIS 2006 pp 366 369 OxNN Hexadecimal numbers are represented by using C language notation ObNN Binary numbers are represented by using C language notation NN Decimal numbers have no prefix 15 February 2008 6 of 29 02005 2008 Philips Electronics Nederland B V 3D Interface Specifications 2 Physical interface 3D Solutions The connection between the host and the 3D Display makes use of a DVI based interface see DVI Table 1 DVI specifications for the 42 displays Resolution 1920 x 1080 Refresh rate 60 fps Gamma No gamma correction Reference white Wx 0 280 Wy 0 285 at 10500 K Aspect ratio 16 9 Order of sample scanning Left to right Top to bottom Progressive Coded signal colour space RGB 8 bit coding Supported standards DDC DDC CI Not supported standards GTF DPMS Table 2 DVI specifications for the 20 displays Res
12. art 10 bytes the content is interpreted as being in the 2D plus Depth format The table below lists the content of the 32 header bytes The right column contains the actual value or the parameter which is further defined in the following sections H 0 Header IDI H 1 Content type H 2 Factor H 3 Hdr Offset CC H 4 Select 4 5 Reserved H 6 H 9 EDC I H 10 Header ID2 H 11 H 12 Data types H 13 27 Reserved Fixed H 28 H 31 EDC 2 Unused and reserved bits must be set to zero The display interprets the header each frame 60 times per second Changed header values are effectuated directly 6 2 1 Header IDI Indicates the format of the remainder of the header see Table 4 Table 4 Header IDs This header ID indicates the generic header type with a length of 10 AAO bytes as described 6 2 11110001 6 2 2 Content type This value defines the kind of content Based on this the various visualization parameters like factor and offset are chosen that influence the rendering process Table 5 Content type H 1 Hdr Content type 00000000 No depth 00000001 Signage 15 February 2008 02005 2008 Philips Electronics Nederland B V 17 of 29 3D Interface Specifications 3D Solutions 00000010 Movie 0000001 I Game 00000100 CGI 00000101 Still 00000
13. at for the use of the 2D plus Depth format the use of the first header part is sufficient but both header parts can be used as well For the use of Declipse content however both header parts are required Table 11 Standard headers first header part F10140800000C42DD3AF Used for signage content s3d F102408000001F3A7B38 Used for video content v3d F10540800000E4D9502C Used for stills b3d Table 12 Standard headers second header part F2140000000000000000000000000000000036958221 2D plus Depth F2149A0000000000000000000000000000006BF6C689 removed redundant data F214EF0000000000000000000000000000002FF0C45F full background data Section 6 1 contains more information on the location of the header and how to code it into the image 15 February 2008 02005 2008 Philips Electronics Nederland B V 21 of 29 3D Interface Specifications 3D Solutions 7 Operation modes The display has two operational modes i e 2D mode and 3D mode which are described in sections 7 1 and 7 2 respectively 7 1 2D Mode At the moment that the display detects no header it switches back to 2D mode The 2D mode is achieved by placing the 1920x1080 pixels on the display after a image processing step that lowers the visibility of the optical coupling between the LCD and the lens layer Thereby a high quality 2D image is visualized 7 2 3D Mode The image processing
14. gure 9 shows how the images A B C and D are arranged on the interface H 0 A B 1 C D 2 A B 3 C D 4 A B 5 C D 6 V l 1 V 1 0 H 2 1 H 2 H 2 H 2 Figure 9 Naming order of the 4 sub images 15 February 2008 2005 2008 Philips Electronics Nederland B V 19 of 29 3D Interface Specifications 3D Solutions Table 10 Data type of the supporting render methods 11 Data Traditional 2D plus Depth format as denoted below DE H1 Data type gt Declipse Removed redundant data format as denoted below ee H 12 10011010 H 11 ap Data type Declipse Full background data format as denoted below 00010100 H 12 Using the above information we show three supported image formats The images are displayed as present in the video files On the interface towards the 3D display they are interleaved as shown in figure 9 Location Traditional 2D plus Depth Data type 00010100 00000000 A 2D left B Depth right C Empty D Empty Declipse Removed redundant data This is de default version especially designed for content that is being stored in compressed form Data type 00010100 10011010 7 A 2D top left B Depth top right C 2D background bottom left removed redundant data D Depth background bottom righ
15. he blue sub pixels that contain the header are replaced by the blue component of their neighboring pixels This masks the header entirely while it hardly affects image quality The following formula shows which bit of which sub pixel byte forms which bit y and byte x of the header x 0 30 y 0 7 2 7 16 0 H x 1 The following translation table shows how the header is composed from bits the blue sub pixels The first header byte H 0 is composed by combining MSB bits of blue sub pixels 15 February 2008 2005 2008 Philips Electronics Nederland B V 15 of 29 3D Solutions 3D Interface Specifications Table 3 Translation table of blue sub pixel MSB bits into header bytes B 0 0 H 0 B 2 0 H 0 B 4 0 H 0 B 6 0 H 0 B 8 0 H 0 B 10 0 H 0 12 0 H 0 B 14 0 H 0 B 16 0 HA B 18 0 H B 20 0 Hy 22 0 24 0 HO B 26 0 28 0 B 30 0 HA 32 0 2 B 34 0 H 2 B 36 0 etc etc 15 February 2008 16 of 29 2005 2008 Philips Electronics Nederland B V 3D Interface Specifications 3D Solutions 6 2 Header data The header is 32 bytes long and consists of two main parts the first part of 10 bytes and the second part of 22 bytes Declipse formats can only be used by applying both header parts By only using the first header p
16. image is stretched slightly before rendering to prevent this problem Figure 11 shows how the image is stretched Stretching and rendering takes place inside the display Not Not 960 visible 800 visible Not area area visible area 540 600 Original data image Original data image Displayed on 3D Monitor Displayed on 3D Monitor Figure 11 A band at the left and right side of the image data are not displayed on the 3D Display The original 2D input data has a resolution of 960x540 The leftmost pixels and the rightmost pixels of each row are not visible on the display when the depth of the pixels is on the screen However when the depth is behind the screen the observer is able to see this information in these bands So while sending content to the display the designer must keep in mind that small bands at the left and right side will mostly not be visible This holds especially for text and logos that are closely located to left and right borders of the screen A side effect of this stretching is that the aspect ratio of the content is somewhat distorted Table 13 Sizes of the invisible borders 42 display firmware version gt 9 6 10 10 20 display firmware version gt 8 0 8 8 15 February 2008 24 of 29 2005 2008 Philips Electronics Nederland B V 3D Interface Specifications 3D Solutions 9 Converting depth to disparity In 3D application a 3 dimensional Cartesia
17. ion T the 3D d ORE agp rp tap aos paquete 15 6 2 DEO abd aD palin Rc pi Ne Bp ced De des beso Pu RD ue 17 6 9 Standard headers istisi ped nt pp ORBE b aD pO 21 7 Op rationiodes DU 22 c MIDI M EU LL M 22 P EE i crc ng pre ee EN eee Te eee a a 22 8 Stretching 24 9 Converting depth to dispallby coute 25 Appendix A CRC 32 implementation eese tentent tentent tentent n ttente tenen tento 26 15 February 2008 2005 2008 Philips Electronics Nederland B V 3 of 29 3D Interface Specifications 3D Solutions 15 February 2008 4 of 29 02005 2008 Philips Electronics Nederland B V 3D Interface Specifications 3D Solutions 1 Introduction 1 1 Scope and Purpose This document contains the interface specifications of the 3D Displays as developed by Philips 3D Solutions This document covers the physical specification as well as the data format All displays with type numbers 42 30 and 20 3D developed by 3D Solutions comply with the specification described herein The header described in section 6 is only supported by displays with firmware 10 7 or higher Please check your firmware version using the Display Control Tool provided with each Philips 3D display The intended audiences are 30 sys
18. n coordinate system is used to model objects Normally these models are used to generate 2D images of scenes like games or other applications The function below describes the translation from depth Z to disparity D Z Herein depth describes the depth extracted from the application and needs to be normalized between 0 and Further disparity refers to the difference in images from the left and right eye that the brain uses as a binocular cue to determine depth or distance of an object The function below will describe the correct translation between depth and disparity D Z M 1 _ _ c Z Zd t vz where D disparity 0 255 Z depth 0 1 M Linear function multiplier Zd depth of display plane vz View distance in coordinate units C Linear function constant Within this formula there are a number of constants present M Zd vz and C To obtain the best 3D performance for each of our displays use the correct values from the table below 0 467481 0 459813 vz 7 655192 6 180772 M 1960 37 1586 34 127 5 127 5 02005 2008 Philips Electronics Nederland B V 15 February 2008 25 of 29 3D Interface Specifications 3D Solutions Appendix A CRC 32 implementation crc32 h Philips 3D Solutions Eindhoven CRC32 implementation adapted from public domain code written by Eric Durbin f ifndef CRC32 define CRC32 ifdef _ cplusplus extern C ndif
19. olution 1600 x 1200 Refresh rate 60 fps Gamma No gamma correction Reference white Wx 0 313 Wy 0 329 at 6500 K Aspect ratio 4 3 Order of sample scanning Left to right Top to bottom Progressive Coded signal colour space RGB 8 bit coding Supported standards DDC DDC CI Not supported standards GTF DPMS 02005 2008 Philips Electronics Nederland B V 15 February 2008 7 of 29 3D Interface Specifications 3D Solutions 3 Display Control Tool The Display Control Tool is supplied with each 3D Display from Philips 3D Solutions and is running on Windows based computers It enables the viewer to control the perceived depth and color settings real time It is advised to always install the Display Control Tool on each computer The Display Control Tool communicates with the display via the DVI cable making use of the DDC CI protocol There is no interaction on the computer between the Display Control Tool and any other software So while developing software according to these interface specifications no consideration regarding the Display Control Tool must be taken The Display Control Tool requires an NVIDIA graphics card See DCT for more information about the Display Control Tool 15 February 2008 8 of 29 02005 2008 Philips Electronics Nederland B V 3D Interface Specifications 3D Solutions 4 WOWNvx formats 4 1 WOWYNx 2D plus Depth format The
20. r select I Hdr Factor is used 0 Hdr content type is used 4 Offset CC select I Offset CC is used 0 content type is used 15 February 2008 18 of 29 2005 2008 Philips Electronics Nederland B V 3D Interface Specifications 3D Solutions When all select signals are low the rendering settings are set to optimal settings for the content type denoted by Hdr content type By making select signals high the settings for Factor and Offset cc can be controlled individually by the header 6 2 6 Reserved The bits 4 5 are reserved for future use and should be set to 0 to maintain compatibility with future systems 62 7 EDCI The 4 byte EDC field H 6 H 9 contains an Error Detection Code computed over the first 6 header bytes This EDC uses the standard CRC 32 polynomial as defined in IEEE 802 3 and ITU T V 42 The initial value and final XOR value are both 0 An implementation of this algorithm can be found in Appendix A 6 2 8 Header ID2 Indicates the format of the remainder of the header see table 4 Table 9 Header IDs This header ID indicates the second header type with a length of 22 bytes as described 6 2 11110010 6 2 9 Data types One image can contain up to 4 sub images Data types describes the contents of the 4 sub pictures as described in Table 10 The sub images are named A B C and D Fi
21. ry 2008 14 of 29 2005 2008 Philips Electronics Nederland B V 3D Interface Specifications 3D Solutions 6 Header The header described in this section is only supported by displays with firmware 10 7 or higher Please check your firmware version using the Display Control Tool provided with each Philips 3D display The header is positioned in the upper left corner of a frame Each frame will contain a header otherwise the display will interpret it as a 2D frame The header instructs the display which viewing experience is required 6 1 Header location in the 3D frame B 12 0 1920 012345678 9101112 ARUN 1080 Figure 8 Numbering of rows and columns of screen The figure above shows how the columns and rows are numbered Numbering starts with zero in the upper left corner Each square pixel consists out of an R G and B sub pixel B 12 0 Denotes the blue byte in the upper row in the 13th column and bits 7 down to 0 With 7 the Most Significant Bit The header is located in the blue sub pixels It is required to give all the 8 bits in the blue sub pixel the value of the header bit In this way the system becomes more tolerant to changes in brightness contrast settings and noise The header starts at the first pixel in the first row B 0 0 and onwards B x 0 Only in the even blue sub pixels the header is located starting with O 2 4 etc During rendering t
22. s in more detail 4 2 WOW Declipse format The WOW Nvx Declipse format is an extension on the existing 2D plus Depth format described in section 4 1 Information on the background is added enabling the rendering algorithm filling in the occluded areas created by the foreground object Figure 2 shows the four quadrants that should be supplied to the screen 15 February 2008 2005 2008 Philips Electronics Nederland B V 9 of 29 3D Interface Specifications 3D Solutions Figure 2 Overview of 3D frame in the WOWvx Declipse format The format contains the following data e Header see sections 5 and 6 e 2D sub image top left e Depth sub image top right A Depth map in grey scale picture This depth map belongs to the top left 2D sub image 2D background sub image bottom left 2D sub image of the background At the places where the 2D information is equal to the quadrant above top left the 2D information is set to black to avoid encoding redundant data e Depth background sub image bottom right The Depth map on the areas occluded by the foreground objects this depth map relates to the bottom left 2D image At the places where the disparity is equal to the quadrant above top right the disparity is set to maximal white to avoid encoding redundant data Note that the removal of the redundant data in the 2D and Depth background sub images is not strictly required It is also feasible to have the background information f
23. t removed redundant data Declipse Full background data The following is specifically used for uncompressed data as in applications that generate content Data_type 00010100 I 1IOLITI Depth background A 2D top left B Depth top right C 2D background bottom left D bottom right 6 2 10 Fixed H 13 H 27 i u wN LUE I RR EE These bits are reserved for future use and should be set to 0 to maintain compatibility with future systems 15 February 2008 20 of 29 02005 2008 Philips Electronics Nederland B V 3D Interface Specifications 3D Solutions 6 2 11 EDC 2 The 4 byte EDC field 28 31 contains an Error Detection Code computed over the first 18 bytes of the second header This EDC uses the standard CRC 32 polynomial as defined in IEEE 802 3 and ITU T V 42 The initial value and final XOR value are both 0 An implementation of this algorithm can be found in appendix A 6 3 Standard headers In many applications a standard header is sufficient With a proper combination of both header parts a choice can be made for the content and data types cf sections 6 2 2 and 6 2 9 For example the 3DS Media Player inserts a header depending on the content type derived from the file extension s3d v3d or b3d and the data type Declipse or 2D plus Depth Standard headers for both header parts are denoted in Table and Table 12 Note th
24. tem integrators that do not make use of software from Philips 3D Solutions The intended applications are 3D CAD design gaming gambling interactive applications etc For these persons the sections 6 1 general header location and 6 2 header description are most interesting For the development of digital signage related software we refer to section 6 1 where the general header location is denoted and to section 6 3 where the standard headers are mentioned for the 3d content using the s3d v3d or b3d extensions e For the development of other software please first contact Philips 3D Solutions WOWNvx Declipse is only supported on Displays with firmware 10 3 or higher Please check your firmware version using the DisplayControlTool Philips is under no circumstances responsible for malfunctioning of third party software Furthermore Philips is not liable for the consequences resulting from changes of mistakes within this whitepaper 15 February 2008 2005 2008 Philips Electronics Nederland B V 5 of 29 3D Solutions 3D Interface Specifications 1 2 References The following references are only informative DCT DDC DDC CI DPMS DVI EDID GTF I2C 3DTVC 1 3 Notations Display Control Tool User manual Philips 3D Solutions VESA Display data channel standard Version 3 December 15 1997 VESA Display data channel command interface DDC CI standard Version 1 august 14 1998 VESA Display Power Man
25. ully available in the frame This is especially useful for uncompressed data e g applications that generate content In the case of content that is being stored or transmitted in compressed form it is more favorable to remove the redundant data in order to improve the coding efficiency See section 6 2 9 for more information on the two different Declipse modes When the Declipse format in which the redundant data is removed is used as depicted in Figure 2 it is required that at least 6 pixels at the edge of the non redundant data part in the bottom layer have to be present in every direction as in the foreground top layer i e at least 6 pixels of redundant data in every direction around background data has to be available in the 2D and Depth background sub images Furthermore it is required that the resulting background data after having the 6 pixel overlap with the top layer is rounded up to macroblock boundaries So the overlap in both layers may vary between 6 and 21 pixels given a macroblock size of 16 16 pixels 15 February 2008 10 of 29 2005 2008 Philips Electronics Nederland B V 3D Interface Specifications 3D Solutions 4 3 2D sub images A 2D sub image has a resolution of half the native panel resolution in both horizontal and vertical direction It is an R G and B image with 8 bits per sub pixel No gamma correction is performed in the display In Figure 3 an example of a 2D sub image is given Figure 3 2D sub
26. within the display takes place in 3 steps see Figure 10 First a frame with a header and a 2D and Depth sub image is applied to the DVI input connector A demultiplexing block decomposes the 3D frame into the header and the 2D and Depth sub images These 3 components are applied to the rendering block The rendering block generates 9 images which have all a slightly different camera positions The amount of perceived depth and other depth related parameters are controlled by the values in the header The 9 different images are fed to the interweaving block This block ensures that each sub pixel is exactly located under the right lens which ensures the best 3D experience The interweaving process is optimized for the optical behavior of the lens layer 15 February 2008 22 of 29 02005 2008 Philips Electronics Nederland B V 3D Interface Specifications 3D Solutions 3D Frame 3 with header DVI Input Header Figure 10 Overview of 3D image processing in the display 15 February 2008 02005 2008 Philips Electronics Nederland B V 23 of 29 3D Interface Specifications 3D Solutions 8 Stretching of images If an object positioned at the left edge of the image has a depth that puts it behind the screen a viewer looking at it from the right would expect to see more pixels to the left of the object where the frame of the display would no longer be concealing the object Because these pixels are not in the original image the
27. xE9362689L OxEDF73BS3EL OxF3B0O6B3BL OxF771768CL 0xFA325055L OxFEF34DE2L OxC6BCFO5FL OxC27DEDE8L OxCF3ECB31L OxCBFFD686L OxD5B88683L 0xD1799B34L OxDC3ABDEDL OxD8FBAO5AL 0x690CEOEEL Ox6DCDFD59L Ox608EDB80L 0x644FC637L 0x7A089632L Ox7EC98B85L 0x738AAD5CL Ox774BBOEBL OxA4F040D56L Ox4BC510E1L 0x46863638L 0x42472B8FL 0x5C007B8AL 0x58C1663DL 0 558240 41 0x51435D53L 0x251D3B9EL 0x21DC2629L Ox2C9FOOFOL 0x285E1D47L 0x36194D42L 0x32D850F5L Ox3F9B762CL Ox3B5A6B9BL 0x0315D626L 0x07D4CB91L 0x0A97ED48L OxOE5GFOFFL Ox1011AOFAL Ox14DOBDA4DL 0x19939B94L 0x1D528623L OxF12F560EL OxF5EE4BB9L OxF8AD6D60L OxFC6C70D7L OxE22B20D2L OxE6EA3D65L OxEBA91BBCL OxEF68060BL OxD727BBB6L OxD3E6A601L OxDEA580D8L OxDA649D6FL 0xC423CD6AL OxCOE2DODDL OxCDA1F604L 0xC960EBB3L OxBD3E8D7EL OxB9FF90C9L OxB4BCB610L OxBO7DABA7L OxAE3AFBA2L OxAAFBE615L OxA7B8COCCL 0xA379DD7BL 0x9B3660C6L 0Ox9FF77D71L 0x92B45BA8L 0x9675461FL 0x8832161AL Ox8CF30BADL 0x81B02D74L 0x857130C3L 0x5D8A9099L 0x594B8D2EL 0x5408ABF7L 0x50C9B640L Ox4E8EE645L Ox4A4FFBF2L 0x470CDD2BL Ox43CDCO9CL 0x7B827D21L 0x7F436096L 0x7200464FL Ox76C15BF8L 0x68860BFDL 0x6C47164AL 0x61043093L 0x65C52D24L Ox119B4BE9L 0x155A565EL 0x18197087L Ox1CD86D30L 0x029F3D35L 0x065E2082L OxOB1D065BL OxOFDCIBECL 0x3793A651L 0x3352BBE6L Ox3E119D3FL Ox3AD08088L 0x2497D08DL 0x2056CD3AL Ox2D15EBE3L 0x29D4F654L 0xC5A92679L OxC1683BCEL OxCC2B1D17
Download Pdf Manuals
Related Search