HA-Buffer: Coherent Hashing for single-pass A-buffer

نویسندگان

  • Sylvain Lefebvre
  • Samuel Hornus
  • Anass Lasram
چکیده

Identifying all the surfaces projecting into a pixel has several important applications in Computer Graphics, such as transparency and CSG. These applications further require ordering, in each pixel, the surfaces by their distance to the viewer. In real-time rendering engines, this is often achieved by recording sorted lists of the fragments produced by the rasterization pipeline. The major challenge is that the number of fragments is not known in advance. This results in computational and memory overheads due to the necessary dynamic nature of the data-structure. Similarly, many fragments which are not useful for the final image—due to opacity accumulation for instance—have to be stored and sorted nonetheless, negatively impacting performance. This paper proposes a novel approach which records and simultaneously sorts all fragments in a single geometry pass. The storage overhead per fragment is typically lower than 8 bits per record, and no pointers are involved. Since fragments are progressively sorted in memory, it is possible to assess during rendering whether a new fragment is useful. Our approach combines advantages of previous approaches at similar levels of performance, and is implemented in a single fragment shader of 24 lines of GLSL. Key-words: A-buffer, real-time, transparency, spatial hashing ha l-0 08 11 58 5, v er si on 1 10 A pr 2 01 3 HA-Buffer: application du hachage cohérent à la création de A-buffer en une passe Résumé : Plusieurs applications en synthèse d’image nécessitent le calcul de l’ensemble des surfaces visibles au travers d’un pixel. Citons le dessin correct de surfaces transparentes ainsi que le dessin de modèles CSG. Ces applications nécessite également de trier les surfaces, pour chaque pixel, selon leur distance au point de vue. Pour les applications en temps-réel, ce sont les fragments produits par l’étape de rasterisation qui sont triés et stockés en mémoire vidéo. Le nombre de ces fragments n’étant pas connu à l’avance, il est nécessaire d’utiliser de coûteuses techniques de gestion de la mémoire. De plus, tous les fragments sont traités même si une fraction non négligeable d’entre eux peut être inutile au dessin de l’image finale (grâce, par exemple, à l’accumulation de l’opacité de plusieurs surfaces combinées). Nous proposons une technique simple pour trier les fragments d’un même pixel au moment de leur rasterisation, sans utiliser de liste chainée (et donc de pointeur). Puisque la liste des fragments pour un pixel est toujours triée, il est possible de déterminer, au moment de sa rasterisation, si un fragment contribuera ou pas à l’image finale, et de le rejetter le cas échéant. La technique combine les avantages de plusieurs approches existantes pour un niveau de performance similaire. Elle a l’unique avantage d’être très simple à coder : 24 lignes de GLSL. Mots-clés : A-buffer, temps-réel, transparence, hachage spatial ha l-0 08 11 58 5, v er si on 1 10 A pr 2 01 3

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

A Novel Hashing Method Suitable for Lookup Functions

The paper describes a novel hashing method suitable for lookup functions. Traditional methods are either of lower hash quality or inferior performance. In this paper, we show how to compute a high-quality hash digest at extremely high performance on Intel ® processors using the crc32 instruction. Our method also generates multiple 32-bit hashes of an input data buffer, which have low correlatio...

متن کامل

Multi-Hash: A Family of Cryptographic Hash Algorithm Extensions white paper

The paper proposes extensions to cryptographic hash algorithms, such as SHA, that add support for parallel processing of a single message. The goal is to take concepts from tree hashing and apply the parallel performance benefits to a single data buffer in a single threaded core of a modern microprocessor. Additionally, a method for applying the Multi-Hash concept to HMAC is suggested. The pape...

متن کامل

Realization of Tunable Universal Filter

This paper provides a voltage mode (VM) multi input single output (MISO) universal filter with the help of voltage differencing inverting buffer amplifier (VDIBA). All the filter functions i.e., Low pass, band pass, band reject, high pass, and all pass filter functions can be realized without modifying the circuit configuration. The pole frequency can be independently tunable by its transconduc...

متن کامل

A Single-File Version of Linear Hashing with Partial Expansions

Linear hashing with partial expansions is a file organization intended for dynamic files. In this paper an improved version of the method is presented. Instead of having a separate overflow area, the storage area for overflow records is incorporated into the main file. Hence, all the records are stored in one file which grows and shrinks according to the number of records stored. By using sever...

متن کامل

A Hardware F-Buffer Implementation

This paper describes the hardware F-Buffer implementation featured in the latest ATI graphics processors. We discuss the implementation choices made in each chip and the various implementation challenges faced like overflow handling. The F-Buffer was originally intended as a solution for multi-pass shading. We demonstrate this functionality, comparing it to traditional multi-pass rendering tech...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2013