Hi !
I´ve a question regarding wavetable samples.
Commonly, sampled audio data is stored in 16-bit integer values (PCM).
The standard accepts sample chunks either in float or integer type but
also says, that sample data represented in integer value should be
rescaled to floating point.
One problem comes up, when very little RAM space is available on the
host platform for running the SA decoder on SA files containing a lot of
sample data, like "elpelele", "cdl" or "inmood": The decoder runs out of
memory unable to set up the sythesis engine.
To prevent from cases like this, it could make sense to store them as
integer tables in the local RAM. This would half the memory
requirements. The casting to the SA normative format of 32bit-float has
to be done every time, the table values are read in the run-time phase
of the decoding process.
What do you think ? Is this possible, or is it a stupid idea of a
non-normative alien ?
Oh, and there is another question: Does any decoder implementation, so
far (SAOLC, SAINT, SFRONT...), guarantee, that all memory being
allocated during the initializing phase of the decoder and during the
run-time phase is freed at the end of the decoding process ?
best
mareK
--
----
--------
----------------
--------------------------------
---------------------------------------------------------
Internet: http://i.am/marek
e-mailto:m.claussen@tuhh.de
Tel.: +49 (0) 40 / 220 72 88
Fax : +49 (0) 89 / 2443 23 390
ICQ-#: 15375072
guestbook: http://www.guestbook.de/yasg.cgi?X=83297
This archive was generated by hypermail 2b29 : Mon Jan 28 2002 - 12:03:55 EST