Українська   Русский
DonNTU   Masters' portal

Abstract

Content

Introduction

Nowadays we see changes from traditional methods of visualization to 3D for increasing quality and realizm degree of computer systems. Terms 3D-graphics and 3D-reflection are different. First of them describes traditional methods of getting 3D-image on similar display. The second is used with stereoscopic technologies of volume reflection, which are supported by the biggest part of modern devices of 3D-reflection.

This process is not used only with tradional computer systems. All expensive movies are in 3D-format today, stereo effect is also used in computer games, systems of virtual reality and augmented reality. Modern TV models are also support 3D technology. That's why modern soft developper must understand on elementary level modern 3D technologies.

1. Theme actuality

Because of using 3D-technologies in mostly all spheres of computer graphics, IT-specialists are forced to spend more time in learning this aspect.

Today the great number of works are sanctified searching effective methods and facilities of a volume (spatial) 3D-visualization [1-4]. The existent methods of stereo visualization are not perfect, formats are not standardized, calculations are labour intensive and devices are expensive, that's why continuation of researches in this sphere is seen an obvious and perspective process.

2. Goal and tasks of research, future results

The goal of master's degree work is an increasing of efficiency of organization realistic visualization of 3D-scenes on the parallel computer systems, namely research of efficiency of realization presentation, post-treatment and output of stereo videostream for standard 3D devices of stereo reflection on GPU.

The basic tasks of research are:

  1. Research of existent methods of 3D-visualization.
  2. Research of existent formats of 3D-video.
  3. Research of presentation and organization output of standard 3D-stereo videostreams for devices of 3D-stereo reflection.
  4. Analysis of standard videocodecs that is used for creation 3D videos.
  5. Research of Nvidia CUDA technology, as facility of calculations speed increasing.
  6. Development of model and prototype of the programm system for presentation, post-treatment and output stereo videostream for standard 3D devices of stereo reflection on GPU.
  7. Research of descriptions and efficiency of the system, development of recommendations about it's using and revision.

3. Overview of researches and developments

3.1 Review of existent methods of 3D-visualization

There are three basic methods of stereo visualization : anaglyph, active 3D technology (shutter) and passive 3D technology.

Anaglyph is one of the simplest and oldest methods of getting 3D-image. Colour filters (special glasses with the lenses of different color (mostly, it is a red and blue color)) are used for the selection of the left and right foreshortenings .

Basic lacks of method:

  1. Distortion of tints (because of colour principle of division of foreshortening) and beating of conflict colors (twinkling of colors consilient with colors of glasses in opposite glasses)
  2. G-hosting (doubling) (appears because of off-grade color filters, off-grade video data or distortion of colors at transmissions of video on a screen) [5].

The general schemeof work of this technology can be described as follows:

  1. Two images (for every eye) output on a screen together
  2. Both images have some different prospects
  3. Every light filter in glasses blocks the color in a channel and skips opposite for every eye
  4. As a result the brain of man combines pictures of each eye and perceives them as 3D.

A distinctive feature of active 3D is a transfer image by turn on every eye. The special glasses that is named a shutter or 3D-glasses with an active shutter are used for this purpose. This glasses close left or right eye with high speed by turns [6].

At passive 3D technology two images output on a screen at the same time. Images are skipped through polarization light filters (under different corners) and then parallel passed in glasses, without being laid on each other. The glasses which are used in passive technology also have polarization light filters and perceive the information (intended for a certain eye) only [7].

Glasses for shutter technology 3D are more expensive, than glasses for passive (polarized) technology. They require the source of energy and that's why they are heavier than polarized glasses. As a rule, glasses of one producer do not suit to TV's of other. It can be attributed to minuses of technology.

Crosstalks (halo round an image) or twinkling are among minuses of the technology. But increasing of frequency of shots, response time matrix of TV minimizes this defect and makes it factually invisible [8].

Because of high frequancy of shots changing and permanent switching of shutter in glasses a viewer can get a headache or feeling of tiredness [9].

Passive 3D technology works quite othergates. Image transmission comes the same time on the left and right eye. An image is sent under the different corners of polarization and, passing through the polarization lenses of glasses, acts for separately perception of a viewer. At such technology, every eye gets it's image, and other image is sifted. Passive glasses does not need source of energy for viewing 3D images [6].

Passive technology 3D uses cheap polarization glasses that does not cause the tiredness of eyes and headaches during watching. In passive technology loss of brightness (50%) to view of volume image some lower than at active (70%) technology of reproducing 3D. It's about plusses of technology. Basic minus is that passive technology 3D uses the interlased method. A viewer see worser quality of image because of that. Because of the inflicted tape on the TV screen the brightness of image is lower when a viewer sees 2D video [10]. In addition, a viewer must sit right in front of display, the rejection of head results in the loss of stereo effect.

Investigating a today's market of 3D-monitors and TV's, it is possible to see that LG, Vizio and Toshiba made a choice for passive 3D technology, and Samsung, Sony and Panasonic chose active 3D-technology [11].

3.2 Review of existent formats of 3D-video

Except methods of 3D visualization, there are a great number of formats of presentation of stereo 3D video. Such variety is related to that the different formats of 3D-video are used depending of a purview (television, cinema, computer games). Also different formats are supported by different producers. So, Samsung in their 3D-TV's prefers active technology which uses the format of anamorph stereopair, LG prefers passive 3D technology, which is better for interlace. Choosing a format of 3D-video it's important to observe the standards which were envisaged by the producer of 3D-device, because deviation from them conduces to the loss of quality at the best and error during the video playing at worst.

All existent formats of 3D-video use the concept of stereopair.

A Stereopair is a pair of flat images of the same object (plot), which has the distinctions between images, called to create the effect of volume. An effect arises up because parts of plot are on the different distances from viewer and have different angular displacement (parallax) during the viewing from different points (corresponding to the right and left eye). During watching of stereopair every eye perceived the image intended for him. This situation creates 3D effect [12].

Among the formats of 3D-video there are a complete format 3D HD video, which uses a full-size stereopair, anamorphs, which squeeze stereopair's size.

A full-size stereopair can be two types: a horizontal stereopair and a vertical stereopair.

Horizontal stereopair (or SideBySide).Two images are located alongside with each other. There are two methods here:

  1. parallel stereopair: here a view is directed in parallel;
  2. cross stereopair: image located on the right is intended for the left eye, and image on the left – for a right eye [13].

A vertical stereopair (OverUnder) is analogical to horizontal. A difference is that two images are situated here above each other.

Horizontal and vertical stereopairs are rarely enough used for reproducing on a 3D TV. It is related to that a full-size horizontal stereopair has a resolution 3840x1080, full-size vertical – 1920x2160, while most modern TV's have Full HD format (1920x1080).

Anamorph stereopair has a resolution in two times less than in a horizontal or vertical line, depending on the location of foreshortening [14]. Anamorph stereopair is used in an order to blend in permission 1920x1080 of HD-TV.

There are horizontal and vertical anamorph stereopairs, where an image is compressed, accordingly, on a horizontal or a vertical line.

Interlaced is a method of getting stereo effect by the lines mixing of both foreshortening in one shot. The image of one foreshortening (for example left) is written down in even scan lines and the image of other foreshortening is written in not even other (for example right). At such method the half of vertical permission disappears at every foreshortening, i.e. permission of film becomes 720х240 at complete 720х480 in a 2D version. It looks as the coloured image with doubling as a "comb" and works with a line output on a monitor (linescan) and foreshortening is mixed up and there are not division in glasses and 3D effect with the included filter the "deinterlace" (progressive) [15].

Another method to get 3D effect is Pulfrih's effect. It'd 2D image with permanent motion (linear or on a radius) of object or camera. It takes off by one camera. In fact this is 2D image, but with the help of dark filter a viewer get brain delay and because of that we get a stereopair. The brain understands one foreshortening immediately but another one with delay in 1-2 frames. During that time an object or a camera changes position or point of survey. If there is no motion in a shot it's 2D [16].

Format of video for viewing anaglyph 3D has the same name. Prepared anaglyph is 2d image with foreshortening which was coded with corresponding colors. The format is not suitable for translation in other 3D formats and does not require the presence of stereo player for viewing. The plus of this format is that anaglyph can be reproduced on any 2d device.

Another format is a separate stereopair (both videorows which are forming a stereopair are divided into independent streams [17]).

This format is divided into two subspecieses: Dualstream and Separatefiles. In the subspecies of Dualstream videostreams are united by a general container. In the subspecies of Separatefiles videostreams are written down in separate files. Undeniable advantage here is a comfort and simplicity of treatment, and the main defect is possibility of origin of serious problems during synchronization of separate streams, because it's control is unsuffices and provided only by software [18].

4. Creating videofile in 3D-format

The goal of master's degree work is an increasing of efficiency of organization realistic visualization of 3D-scenes on the parallel computer systems, namely research of efficiency of realization presentation, post-treatment and output of stereo videostream for standard 3D devices of stereo reflection on GPU.

Two videostreams (left and right foreshortening) are given on an entrance. The program in turn gets shots from each of two streams and carries out a conclusion on a screen with standard frequency 60 Hertzs.

Decision of organization of stereo output on the basis of active technology with the use of anamorphosing was made after the analysis of existent methods of stereo visualization and formats of 3D-video. The choise was made because of necessity of getting maximal quality and output speed. Anamorph horizontal stereopair will be used as a format.

The general structure of work of the programm system of getting standard 3D-videofile is at figure 1.

General structure of work of the programm system of getting standard 3D-videofile

Figure 1 – General structure of work of the programm system of getting standard 3D-videofile
(animation: 6 shots, 7 cycles, 49 kb)

The general order of functioning of the system is next:

  1. Breaking up left/right videostreams on shots
  2. Shots Compression(left/right foreshortening) with a horizontal
  3. Combination of the compressed shots in anamorph stereopair
  4. Adding horizontal anamorph stereopair to the videostream 3D
  5. Videostream 3D coding in a standard videofile (MKV)

One of tasks of work is transformation of two static images (stereopairs) to the videofile that will be reproduced on a 3D device of stereo reflection.

In result we have two Full HD images with permission 1920 on 1080. One image is the left foreshortening of stereopair, second is the right one. Both images have *.png format.

It is necessary to squeeze every image on a horizontal in 2 times, and to unite both foreshortening in anamorph stereopair.

For realization of the task we are going to use Visual C# language and Microsoft Visual Studio 2012.

Using System.Drawing, it is possible to unite two foreshortening in a horizontal anamorrph stereopair. A code of this action is presented on figure 2.

Anamorph stereopair realization

Figure 2 – Anamorph stereopair realization

From figure 2 we can see, that the designer of class of Bitmap is used for the compression of image. The designer gets an initial image and new size values as parameters.

We get a new image with permission 1920 on 1080 pixels after the compression of two initial images.

After that both compressed images are copied in the loop. The left foreshortening is coping in the left part of the formed image, the right foreshortening is in the right part. The result is saved on a disk.

To see a result in 3D format, it is necessary to transform the image (horizontal anamorph stereopair) in video.

Aforge.Net library is used for transformation image to video. This library give comfortable facilities for video treatment [19].

Video will be got in avi format.

To use of resources of Aforge.Net library you need to write in code the directive "using Aforge.Video.VFW".

Standard video contains 24 shots in a second. In the program video consists of one hundred shots, consequently it's duration will be 4,167 seconds.

Code of transformation anamorph stereopair to video is presented on figure 3.

Converting image to video

Figure 3 – Converting image to video

From figure 3 we can see, that for realization of task we need to create stream of AVIWriter class.With the help of gunction AddFrame () we get the frame-accurate adding of images to the stream.

As a result of implementation of the program we get the file of "video.avi". With the help of "KMPlayer" it is possible to see the information about video, presented on figure 4.

Information about videofile

Figure 4 – Information about videofile «video.avi»

The video will not be reproduced on 3D devices. It is related to that video formats which Aforge.Net uses do not match to standards of 3D video. It is necessary to use videocodecs for transformationt to get video in a correct format.

A video codec is a device or software that enables compression or decompression of digital video. The compression is usually lossy. Historically, video was stored as an analog signal on magnetic tape. Around the time when the compact disc entered the market as a digital-format replacement for analog audio, it became feasible to also begin storing and using video in digital form, and a variety of such technologies began to emerge [20].

MPEG-4 AVC/H.264 is a standard format among compression formats of 3D videos. Videocodec x264 is used for transformation of the videofile to MPEG-4 AVC/H.264 format [21].

x264 is a free software library for encoding video streams into the H.264/MPEG-4 AVC format [22].

Video will be got in MKV expansion after file transformation with videocodec x264. This expansion is correct for 3D video.

The result of experimental creation of stereovideostream is presented on figure 5.

Stereo shot of experimental videostream

Figure 5 – Stereo shot of experimental videostream

Conclusion

As a result of research next results were got:

  1. The analysis of existent methods of 3D stereo of visualization was conducted: anaglyph, active (shutter) technology, passive (polarization) technology
  2. Existent formats of 3D video were considered and analysed : anaglyph, horizontal and vertical stereopairs, anamorphs, interlaced
  3. The structure coming application of stereo visualization of the scenes was considered.
  4. The fragment of the program system which execute transformation of static image to 3D video with expansion of MKV was worked out

The further researches which are related to implementation of master's degree work will be sanctified to the next aspects:

  1. Development of model of the programm system for presentation, post-treatment and conclusion of stereo of videostream for default 3D devices of stereo reflection on GPU
  2. Realization of the programm system
  3. Research of descriptions and efficiency of the worked out system
  4. Development of recommendations about using and finishing off the programm system

Note

This master's work is not completed yet. The full text will be avaliable at February, 2015.

References

  1. Zone R. Stereoscopic cinema & the origins of 3-D film (University Press of Kentucky, 2007) ISBN 0-8131-2461-1, p. 110
  2. 3D Videocommunication – Algorithms, conceptsand real-time systems in human centred communication // Edited by Oliver Schreer, Peter Kauff, Thomas Sikora John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England. – 2005. – 340 p.
  3. Башков Е.А., Зори С.А., Ковальский С.В. Современное алгоритмическое и аппаратное обеспечение виртуальных систем трехмерного моделирования окружающей обстановки В кн.: Наукові праці Донецького національного технічного університету. Серiя "Інформатика, кібернетика і обчислювальна техніка", ИКВТ-2009: – Донецьк: ДонНТУ. – 2009 – 11 с.
  4. Башков Е.А., Зори С.А. Реалистичная визуализация трехмерных объектов и сцен с использованием технологий объемного отображения /Известия ЮФУ. Технические науки. Тематический выпуск «Компьютерные и информационные технологии в науке, инженерии и управлении», № 5(130) – 2012 г., изд-во Технологический институт Южного федерального университета в г. Таганроге, с. 133 – 137
  5. Анаглиф – один из самых старых и самый ругаемый вариант 3D. [Electronic resource]. – Access mode: http://rav-anaglyph.ru.
  6. 3D активное или пассивное — что лучше. [Electronic resource]. – Access mode: http://televizor-info.ru/eto-interesno....
  7. Порфиров П.А., Звягинцев И.А., Зори С.А. Организация реалистичной стерео визуализации сцен 3D-график.
  8. Затворная технология 3D. Что это такое? [Electronic resource]. – Access mode: http://www.vybortv.ru/voprosy....
  9. Active 3D vs. Passive 3D. [Electronic resource]. – Access mode: http://3D-tvbuyingguide.com/3Dtv....
  10. Как работает технология воспроизведения 3D в телевизорах. [Electronic resource]. – Access mode: http://hobbyits.com/cifrovye-texnologii/kak-rabotaet....
  11. Активное или пассивное 3D: преимущества и недостатки 3D очков. [Electronic resource]. – Access mode: http://3Dtv-obzor.ru/aktivnoe_passivnoe_3D.
  12. Википедия. Стереопара. [Electronic resource]. – Access mode: http://ru.wikipedia.org/wiki/Стереопара.
  13. Форматы 3D видео. [Electronic resource]. – Access mode: http://mylcd.info/model/3D_format.html.
  14. Преобразование стереопары в анаморфную. [Electronic resource]. – Access mode: http://vokrug3D.ru/obrabotka-video/preobrazovanie....
  15. О форматах и методах просмотра 3D-Video. [Electronic resource]. – Access mode: http://rutor.org/torrent/225870/.
  16. Форматы 3D. Какой формат качать для вашей техники? [Electronic resource]. – Access mode: http://torrents3D.ru/article/843/.
  17. О Форматах и методах просмотра 3D. [Electronic resource]. – Access mode: http://rutracker.org/forum/viewtopic.php?t=3312998.
  18. Виды форматов 3D изображений. [Electronic resource]. – Access mode: http://www.mirstereofoto.ru/articles/raznoe....
  19. AForge.NET Framework. [Electronic resource]. – Access mode: http://www.aforgenet.com/framework.
  20. Википедия. Видеокодек. [Electronic resource]. – Access mode: http://ru.wikipedia.org/wiki/Видеокодек.
  21. Кодек x264. Часто задаваемые вопросы. [Electronic resource]. – Access mode: http://compression.ru/video/x264/x264_faq_ru.html#Q1.
  22. Википедия. x264. [Electronic resource]. – Access mode: http://ru.wikipedia.org/wiki/X264.