Talk:PVRTC: Difference between revisions
m Added wikiprojects |
Tag: |
||
Line 1: | Line 1: | ||
{{Talk header}} |
{{Talk header}} |
||
{{ |
{{WikiProject banner shell|collapsed=yes|class=Stub| |
||
{{WikiProject Computer graphics| |
{{WikiProject Computer graphics|auto=yes |
||
|auto=yes |
|||
}} |
}} |
||
{{WikiProject Computing |
{{WikiProject Computing|importance=|software=yes|auto=yes}} |
||
{{WikiProject Computer science}} |
{{WikiProject Computer science}} |
||
{{WikiProject Technology}} |
{{WikiProject Technology}} |
Latest revision as of 01:15, 9 February 2024
This is the talk page for discussing improvements to the PVRTC article. This is not a forum for general discussion of the article's subject. |
Article policies
|
Find sources: Google (books · news · scholar · free images · WP refs) · FENS · JSTOR · TWL |
This article is rated Stub-class on Wikipedia's content assessment scale. It is of interest to multiple WikiProjects. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
how accessing 2x2 texels requires access to 4 blocks?
[edit]Article says now: "In both PVRTC and PVRTC2, texture data is stored in blocks (but note that the decoding of any 2x2 set of texels requires access to 4 of these blocks.)". I find this both premature in the discussion, but the statement also seems vague or incorrect. AFAIK accessing 2x2 set of texels MIGHT require access to 4 of these blocks. Depending if all 4 texels are in same block or not. They can be in 1, 2 or 4 different blocks. But more importantly, accessing 1 pixels will always access 1 block. I am not sure what the initial author wanted to write here. Accessing 2x2 set of texels obviously is pretty common occurence when using linear interpolation, but then again, on average it will access less than 4 different blocks (even ignoring reuse of cached data, which might reduce this further). Should instead say something like: "In both PVRTC and PVRTC2, texture data is stored in blocks representing a small rectangular set of texels. When reading back texels which are in the same block, memory bandwidth use and latency can be reduced. When reading multiple nearby texels, depending on exact placement of texels, 1 or more blocks will be required to reconstruct each texel. In practical applications, one can reduce amount of blocks and memory read. I.e. in linear interpolation filtering, that requires 2x2 square set of texels, in best case scenario only one block would be required, and at worst 4 blocks; on average somewhere in between, but pretty close to 1. The memory bandwidth reduction also comes from the compression itself, which uses less memory bits to represent a block of texles, that would be required to represent them individually. The drop in texel quality (compression error) is strongly dependent on the content of the image, and in some cases (similar texel values, flat colors, slowly varying colors, repeating structures) might be none, and in some others (less ordered patterns) might be much bigger." Just my idea. 81.6.34.246 (talk) 06:04, 14 September 2019 (UTC)
Patents / IP.
[edit]Are there any patents on PVRTC, that would prohibit using it for other purposes, chips or software, or developing and distributing custom sofware / hardware using it? 81.6.34.246 (talk) 06:09, 14 September 2019 (UTC)
- Stub-Class computer graphics articles
- Unknown-importance computer graphics articles
- WikiProject Computer graphics articles
- Stub-Class Computing articles
- Unknown-importance Computing articles
- Stub-Class software articles
- Unknown-importance software articles
- Stub-Class software articles of Unknown-importance
- All Software articles
- Automatically assessed Computing articles
- All Computing articles
- Stub-Class Computer science articles
- Unknown-importance Computer science articles
- WikiProject Computer science articles
- Stub-Class Technology articles
- WikiProject Technology articles