::
:: Unfortunetly, the internal represenation inside LAME is 16bit,
:: so we are using sf_read_short :-(
::
:: It is on my TODO list to convert the internal representation
:: to floating point, so then we can handle any type of input
:: (and the resampling code wont have to convert back and
:: forth).
::
:: The API is frozen, and takes input as short int, but these should
:: be converted to wrappers around similar functions which take
:: floating point data as input. And it would be nice to also
:: add an interface to mpglib which outputs floating point.
:: If anyone is interested, let me know and I can give an overview
:: of what needs to be done.
::
I thought lame works internally with floats.
So all my programs are working. float are enough to represent *exactly*
integer up to 25 bit and IEEE-754 floats.
And there are no problems with scaling. Space requirements are only twice
the requirements of the not sufficient 16 bit integer format.
No alignment problems at all like with 24 bit integers.
--
Mit freundlichen Gr��en
Frank Klemm
eMail | [EMAIL PROTECTED] home: [EMAIL PROTECTED]
phone | +49 (3641) 64-2721 home: +49 (3641) 390545
sMail | R.-Breitscheid-Str. 43, 07747 Jena, Germany
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
