User Manual

Table Of Contents
WAVELAB
Off-line processing 9 – 199
Additional settings
Limitations
Time stretch is a very complicated Digital Signal Processing (DSP) oper-
ation. It requires complicated mathematical operations and always affects
the sound quality to some extent:
For speech, stretch factors within a ±30% limit will normally provide very good re-
sults.
For composite music, try to limit the range to ±10%.
For sensitive material, like solo piano, try not to exceed ±3%.
Note that if the DIRAC algorithm is used (see below), the above range rec-
ommendations can be extended and still produce very good results.
Setting Description
Quality There are three Quality settings, “Quick Process”, “Standard” and “High
Quality”. The High Quality mode provides very high quality pitch correc-
tion, but the process will take longer. For most uses, the Standard mode
is probably sufficient.
Use DIRAC
processor
This is described separately below.
Preserve pitch When this option is activated, the pitch of the audio material will not be
affected when you apply Time stretch. This is the normal mode of oper-
ation. If you turn this off, the pitch will change proportionally with the
Time stretch ratio, much like when speeding up or slowing down a tape
recorder.
Audio quantize When this option is activated, the length of the resulting file will be exactly
the one displayed in the “Desired result” section of the dialog. When it is
deactivated, the actual file length may differ by a few milliseconds.
If you don’t need this accuracy, you should leave the option deactivated,
for a slightly better audio quality.
Rhythm accuracy This is only available if “Standard” or “High Quality” mode is selected.
Set this parameter according to whether the rhythmic feel of the audio
material has a high priority or not. If you set this to a high value, the tim-
ing and rhythmic feel will be preserved as far as possible, which is im-
portant for drum tracks, etc. If you set it to a low value, the audio quality
can be slightly better, but sections of audio might be moved slightly (in
the milliseconds region) in the process.