diff options
author | Johannes Schickel | 2010-02-09 21:52:27 +0000 |
---|---|---|
committer | Johannes Schickel | 2010-02-09 21:52:27 +0000 |
commit | 6a624df560aae04b990a64dc6d8b34f960365787 (patch) | |
tree | 2d7ae05fa98b7af75bc4d9d81d3de6b7e15f1265 /sound | |
parent | fa2a1cd22dd3a7a7192bfcdfccff3df80e2f3297 (diff) | |
download | scummvm-rg350-6a624df560aae04b990a64dc6d8b34f960365787.tar.gz scummvm-rg350-6a624df560aae04b990a64dc6d8b34f960365787.tar.bz2 scummvm-rg350-6a624df560aae04b990a64dc6d8b34f960365787.zip |
Make Audio::convertTimeToStreamPos return a frame-precision based result instead of a sub-frame-precision based result. This fixes the SubLoopingAudioStream tests.
svn-id: r48020
Diffstat (limited to 'sound')
-rw-r--r-- | sound/audiostream.cpp | 19 |
1 files changed, 16 insertions, 3 deletions
diff --git a/sound/audiostream.cpp b/sound/audiostream.cpp index 57db44de1f..16c3eaca1a 100644 --- a/sound/audiostream.cpp +++ b/sound/audiostream.cpp @@ -373,9 +373,22 @@ Timestamp convertTimeToStreamPos(const Timestamp &where, int rate, bool isStereo // When the Stream is a stereo stream, we have to assure // that the sample position is an even number. if (isStereo && (result.totalNumberOfFrames() & 1)) - return result.addFrames(-1); // We cut off one sample here. - else - return result; + result = result.addFrames(-1); // We cut off one sample here. + + // Since Timestamp allows sub-frame-precision it might lead to odd behaviors + // when we would just return result. + // + // An example is when converting the timestamp 500ms to a 11025 Hz based + // stream. It would have an internal frame counter of 5512.5. Now when + // doing calculations at frame precision, this might lead to unexpected + // results: The frame difference between a timestamp 1000ms and the above + // mentioned timestamp (both with 11025 as framerate) would be 5512, + // instead of 5513, which is what a frame-precision based code would expect. + // + // By creating a new Timestamp with the given parameters, we create a + // Timestamp with frame-precision, which just drops a sub-frame-precision + // information (i.e. rounds down). + return Timestamp(result.secs(), result.numberOfFrames(), result.framerate()); } } // End of namespace Audio |