Need to adjust scale for floating input.
diff -NaurbB a/libspeex/resample.c b/libspeex/resample.c
--- a/libspeex/resample.c 2013-12-20 17:40:27.326576921 +0500
+++ b/libspeex/resample.c 2013-12-20 17:40:16.918577087 +0500
@@ -915,7 +915,7 @@
if (in) {
for(j=0;j<ichunk;++j)
#ifdef FIXED_POINT
- x[j+st->filt_len-1]=WORD2INT(in[j*istride_save]);
+ x[j+st->filt_len-1]=WORD2INT(32768.*in[j*istride_save]);
#else
x[j+st->filt_len-1]=in[j*istride_save];
#endif
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.xiph.org/pipermail/speex-dev/attachments/20131220/93160412/attachment.htm
On 20/12/13 08:06 AM, Arslan, Fahad wrote:> Need to adjust scale for floating input. > diff -NaurbB a/libspeex/resample.c b/libspeex/resample.c > --- a/libspeex/resample.c 2013-12-20 17:40:27.326576921 +0500 > +++ b/libspeex/resample.c 2013-12-20 17:40:16.918577087 +0500 > @@ -915,7 +915,7 @@ > if (in) { > for(j=0;j<ichunk;++j) > #ifdef FIXED_POINT > - x[j+st->filt_len-1]=WORD2INT(in[j*istride_save]); > + x[j+st->filt_len-1]=WORD2INT(32768.*in[j*istride_save]); > #else > x[j+st->filt_len-1]=in[j*istride_save]; > #endifYou realize that applying this patch would break every application currently using this code? Not using a range of +/-1 in the float API was a mistake, but breaking the API would be worse. Jean-Marc
So this means this should be fixed in Pulseaudio then, right? Thanks, Fahad ________________________________________ From: Jean-Marc Valin [jmvalin at jmvalin.ca] Sent: Saturday, December 21, 2013 8:04 AM To: Arslan, Fahad; speex-dev at xiph.org Subject: Re: [Speex-dev] Scaling fix for float input On 20/12/13 08:06 AM, Arslan, Fahad wrote:> Need to adjust scale for floating input. > diff -NaurbB a/libspeex/resample.c b/libspeex/resample.c > --- a/libspeex/resample.c 2013-12-20 17:40:27.326576921 +0500 > +++ b/libspeex/resample.c 2013-12-20 17:40:16.918577087 +0500 > @@ -915,7 +915,7 @@ > if (in) { > for(j=0;j<ichunk;++j) > #ifdef FIXED_POINT > - x[j+st->filt_len-1]=WORD2INT(in[j*istride_save]); > + x[j+st->filt_len-1]=WORD2INT(32768.*in[j*istride_save]); > #else > x[j+st->filt_len-1]=in[j*istride_save]; > #endifYou realize that applying this patch would break every application currently using this code? Not using a range of +/-1 in the float API was a mistake, but breaking the API would be worse. Jean-Marc