Sound file process
To add a sound file, the easiest way is to drag’n’drop it either from the operating system, or from the library.
Timestretching
ossia score supports 4 modes of playback:
- No stretching at all (Raw).
- Re-pitching through resampling thanks to the libsamplerate library.
- Time-stretching (pitch does not change) thanks to the Rubberband library, in a mode optimized for drums and another for tonal instruments.
If a file has any tempo information (either through ACID tags or a BPM present in the filename), it is automatically assumed to be a loop and put in loop mode with the timestretch enabled.
The tempo of the soundfile can be adjusted through the inspector ; the ratio between its tempo and the current playback tempo will be used to timestretch / repitch accordingly.
RAM / disk playback
score chooses a decoder automatically:
| Source | Usual playback path |
|---|---|
| WAV / W64 supported by the WAV decoder | Memory-mapped disk access, without decoding the whole file into RAM |
| AIFF / AIF / AIFC / CAF up to 4 GiB, supported by libsndfile | Decoded into RAM |
| Other audio up to 4 GiB | Decoded into RAM through FFmpeg |
| Files above 4 GiB that cannot use the memory-mapped path | Streaming through FFmpeg |
| Audio taken from a supported video container | Streaming through FFmpeg |
The size threshold concerns the source file, not the size of decoded samples. Disk speed and decoder cost still matter for streaming, especially with seeking. The browser build uses a different policy: video and files above 48 MiB use streaming decode, while smaller audio files are decoded into RAM.
Sample-rate mismatches are normally converted in the audio graph rather than by resampling the entire file on import. Builds without graph resampling fall back to a suitable decoding path. This conversion is distinct from the musical time-stretch mode above.
Supported formats and video soundtracks
The audio file drop handler recognizes WAV, W64, MP3, M4A, OGG, FLAC, AIF, AIFF, AIFC, CAF, APE, WV, WMA, AAC, OPUS, AC3, DTS and DTSHD extensions. Actual decoding depends on the codec and the libraries included in the build; an extension alone does not guarantee that a particular file is readable.
The sound-file loader also accepts audio from supported video containers. Select the video file as the sound process’s source when only its soundtrack is needed; this does not create a video output. Route the process’s audio outlet to an effect or an Audio device bus as usual.