Timeline for When to consider double (64 bit) floating point for Audio
Current License: CC BY-SA 3.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 25, 2012 at 2:01 | vote | accept | user1849104 | ||
| Nov 25, 2012 at 1:27 | comment | added | hotpaw2 | An example was given, IIR with poles/zeros near the unit circle. If there's a cache, algorithms and working data sets that fit in this cache can be significantly faster than ones that don't. | |
| Nov 25, 2012 at 0:56 | comment | added | user1849104 | Also, what exactly do you mean by trashing the cache? Do you mean that having twice as much data going through it will make things horribly slow? | |
| Nov 25, 2012 at 0:54 | comment | added | user1849104 | This is the sort of info I was after. I'll accept this answer if you could kindly provide a concrete example of a case where double precision is required to make a filter work, i.e., it'll sound bad (or at least fairly ordinary) with single precision, but smooth as butter with double precision. | |
| Nov 24, 2012 at 18:54 | history | edited | hotpaw2 | CC BY-SA 3.0 |
added 15 characters in body
|
| Nov 24, 2012 at 18:48 | history | answered | hotpaw2 | CC BY-SA 3.0 |