Anonymous user
Talk:Arithmetic coding/As a generalized change of radix: Difference between revisions
Talk:Arithmetic coding/As a generalized change of radix (view source)
Revision as of 19:43, 28 January 2016
, 8 years agono edit summary
No edit summary |
No edit summary |
||
Line 21:
:::: After this point, it doesn't matter which number <math>N</math> we return, as long as <math>L <= N < U</math>. --[[User:Trizen|Trizen]] ([[User talk:Trizen|talk]]) 17:28, 28 January 2016 (UTC)
::::: For some reason, math markup isn't working right now, so that's a bit hard to read. But basically you seem to be restating the fact that you are using cumulative frequencies? But my question is why are you using cumulative frequencies? -- the use of cumulative frequencies conflicts with the example and text at [[wp:Arithmetic_coding#Arithmetic_coding_as_a_generalized_change_of_radix|wikipedia]]. --[[User:Rdm|Rdm]] ([[User talk:Rdm|talk]]) 19:35, 28 January 2016 (UTC)
:::::: The Wikipedia section uses cumulative frequencies. In the example, "DABDDB", gets mapped to 3, 0, 1, 3, 3, 1, where 0, 1, 3 are the cumulative frequencies for A, B, D, respectively. You can follow how those are used throughout the calculations.
::::::As another example, consider a situation where each character has frequency 1. Then the frequency dictionary is just every key mapped to 1. What you need for encoding is the cumulative frequencies 0, 1, 2, ..., which tells you what value to assign to each character. --[[User:Spoon!|Spoon!]] ([[User talk:Spoon!|talk]]) 19:43, 28 January 2016 (UTC)
| |||