Re: Guidance
Re: Guidance
- Subject: Re: Guidance
- From: Michael Thornburgh <email@hidden>
- Date: Mon, 14 Apr 2003 02:06:33 -0700
there are two pieces to this problem: sample rate conversion, and 5-law
coding/decoding.
for sample rate conversion, you can use AudioConverter in the
AudioToolbox
framework. however, i personally find the pull-model of AudioConverter
to be
a PITA for this kind of application, and wrote my own, simple,
push-based
sample rate converter. you are (or anyone on the list is) welcome to
use it --
the framework of which it's a part will one day (when i'm less lazy) be
as
open-sourcey as MTCoreAudio. contact me off-list if you'd like a copy.
the 5-law coding/decoding problem is much easier. here are routines to
initialize conversion tables and to convert from linear to 5-law; to
convert
from 5-law to linear, just use the "ulaw2linear" table in the obvious
way. :)
---
static UInt8 linear2ulaw[8159];
static Float32 ulaw2linear[256];
static void init_ulaw()
{
UInt32 x;
UInt8 segment;
UInt32 tmp;
segment = 0x00;
for ( x = 33; x < 64; x++ )
linear2ulaw[x - 33] = ((( x >> 1 ) & 0xf ) + segment ) ^ 0x7f;
segment = 0x10;
for ( x = 64; x < 128; x++ )
linear2ulaw[x - 33] = ((( x >> 2 ) & 0xf ) + segment ) ^ 0x7f;
segment = 0x20;
for ( x = 128; x < 256; x++ )
linear2ulaw[x - 33] = ((( x >> 3 ) & 0xf ) + segment ) ^ 0x7f;
segment = 0x30;
for ( x = 256; x < 512; x++ )
linear2ulaw[x - 33] = ((( x >> 4 ) & 0xf ) + segment ) ^ 0x7f;
segment = 0x40;
for ( x = 512; x < 1024; x++ )
linear2ulaw[x - 33] = ((( x >> 5 ) & 0xf ) + segment ) ^ 0x7f;
segment = 0x50;
for ( x = 1024; x < 2048; x++ )
linear2ulaw[x - 33] = ((( x >> 6 ) & 0xf ) + segment ) ^ 0x7f;
segment = 0x60;
for ( x = 2048; x < 4096; x++ )
linear2ulaw[x - 33] = ((( x >> 7 ) & 0xf ) + segment ) ^ 0x7f;
segment = 0x70;
for ( x = 4096; x < 8192; x++ )
linear2ulaw[x - 33] = ((( x >> 8 ) & 0xf ) + segment ) ^ 0x7f;
for ( x = 0; x < 256; x++ )
{
segment = (( x & 0x70 ) >> 4) & 0x7;
tmp = ( 1 << ( segment + 5 )) + (( x & 0x0f ) << ( segment + 1 )) -
32;
if ( x < 128 )
ulaw2linear[x ^ 0x7f] = ((double)tmp) * ( 1.0 / 8158.0 );
else
ulaw2linear[x ^ 0x7f] = ((double)tmp) * ( -1.0 / 8158.0 );
}
}
static inline UInt8 convertLinearToulaw ( Float32 sample )
{
UInt8 rv = 0;
if ( sample < 0.0 )
{
rv = 0x80;
sample = -sample;
}
if ( sample > 1.0 )
{
sample = 1.0;
}
rv |= linear2ulaw[(unsigned)(sample * 8158.0 )];
return rv;
}
---
-mike
On Monday, April 14, 2003, at 12:35 AM, Matthew Johnson wrote:
Hello All,
I am working on a simple OSX program that takes audio from a
microphone and
writes it down a socket in real time. Reads audio and plays it through
the
speakers in real time. I have the playback and record working fine for
48khz
which is the hardware default for device I am using.
My problem is that the audio must be read and written in 8khz ulaw
format
and I have no idea how to go about this.
Can someone come up with a suggestions on how this is done on the mac
osx?
Thanks,
Matt Johnson
_______________________________________________
coreaudio-api mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/coreaudio-api
Do not post admin requests to the list. They will be ignored.
_______________________________________________
coreaudio-api mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/coreaudio-api
Do not post admin requests to the list. They will be ignored.
| References: | |
| | >Guidance (From: Matthew Johnson <email@hidden>) |