Age | Commit message (Collapse) | Author |
|
svn-id: r55850
|
|
svn-id: r55845
|
|
We used to allow this, but the result is a timestamp with a framerate
potentially different from that of both original timestamps, which can
lead to completely unexpected behavior. For example, consider this code
snippet:
a = a + b;
a = a.addFrames(1); // frame rate changed!
svn-id: r55840
|
|
svn-id: r55814
|
|
svn-id: r47244
|
|
svn-id: r47212
|
|
svn-id: r47120
|
|
frames directly.
svn-id: r47091
|
|
getFramerate -> framerate
svn-id: r47083
|
|
in a Timestamp.
- Add tests for these
svn-id: r47081
|
|
svn-id: r47071
|
|
Instead of storing milliseconds and frames (which causes rounding errors,
and causes ambiguity in how a given time is stored), we now do things
differently: We store a number of seconds, and frames. To make sure that
we can still handle milliseconds accurately, though, we change the
framerate to the least common multiple of the original framerate and
1000. So 60 becomes 6000, and 44100 becomes 441000. There are no visible
changes for client code, except for the increased accuracy.
svn-id: r47070
|
|
* Add convertToFramerate() method
* Add framerate() method
* Add operator == and !=
* Improve frameDiff() to work for two timestamps with distinct framerates
* Improve Doxygen comments
svn-id: r46994
|
|
- Add an assert which prevents the _frameRate from being 0 in the Audio::Timestamp constructor.
svn-id: r46958
|
|
svn-id: r40957
|
|
svn-id: r40867
|
|
svn-id: r39262
|
|
svn-id: r39257
|
|
unit test for it, based on the old (and very outdated) timetest.cpp. To be used by Audio::Mixer one day...
svn-id: r39112
|