[FFmpeg-devel] [PATCH] Expose QCELP's floating-point LSP-to-LPC function

Robert Swain robert.swain
Mon Aug 3 10:34:37 CEST 2009


2009/8/1 Michael Niedermayer <michaelni at gmx.at>:
> On Fri, Jul 31, 2009 at 09:18:42PM +0100, Colin McQuillan wrote:
>> I'd like to use the code in qcelp_lpc.c function for AMR. First, the
>> naming should be made consistent.
>> rename_qcelp_lsp.patch description:
>> Make the LSP naming more consistent
>> Use the convention from lsp.c: an LSF is a frequency, an LSP is the
>> cosine of an LSF, and LSP functions should have an ff_acelp prefix.
>> Use a "d" suffix to specify doubles.
> iam not maintainer of qcelp but as iam not sure if reynaldo is busy,
> ive not seen a mail from him recently IIRC, and soc time is limited
> i approve this one, if reynaldo disagrees it can be reverted


>> move_qcelp_lsp.patch description:
>> Expose QCELP's floating-point LSP-to-LPC function
>> qcelp_lsp exported a single function, ff_acelp_lspd2lpc, which was not
>> specific to qcelp. It can be kept with its fixed-point version
>> ff_acelp_lsp2lpc in lpc.c.
> similarly ok


I will revert if Reynaldo has any complaints but I didn't want to hold up Colin.


More information about the ffmpeg-devel mailing list