encoder/vorbis: accept floating point input samples

Improves quality by not squeezing 32 bit samples down to 16 bit, and
then back to 32 bit floating point.  Reduces CPU usage by skipping a
conversion step.
This commit is contained in:
Max Kellermann
2012-10-02 08:19:45 +02:00
parent e166ddf46f
commit 9e3f843291
2 changed files with 10 additions and 8 deletions

1
NEWS
View File

@@ -5,6 +5,7 @@ ver 0.18 (2012/??/??)
- vorbis: skip 16 bit quantisation, provide float samples
* encoder:
- opus: new encoder plugin for the Opus codec
- vorbis: accept floating point input samples
* output:
- new option "tags" may be used to disable sending tags to output
* improved decoder/output error reporting