mirror of
https://gitlab.linphone.org/BC/public/linphone-iphone.git
synced 2026-01-23 06:08:07 +00:00
2b8200409c
git-svn-id: svn+ssh://svn.savannah.nongnu.org/linphone/trunk@1 3f6dc0c8-ddfe-455d-9043-3cd528dc4637
59 lines
2.6 KiB
Text
59 lines
2.6 KiB
Text
USING OTHER CODECS
|
|
|
|
To use the library with other ACM-supported codecs besides
|
|
TrueSpeech, the following things are required:
|
|
|
|
1. DLL file supporting the codec
|
|
2. correct format tag for the codec
|
|
3. correct WAVEFORMATEX structure for the codec
|
|
|
|
For codecs which are supported by mplayer, the easiest
|
|
way to get items 1 and 2 is by looking at the audio codec
|
|
section of mplayer's codec status table:
|
|
http://mplayerhq.hu/DOCS/codecs-status.html
|
|
If mplayer says it is using a .acm file to support the
|
|
audio codec, then most likely it can be used with this
|
|
library. If mplayer is using another method to support
|
|
that codec, you should probably consider using that
|
|
same method in your Linux program instead of this library,
|
|
because it is most likely a native way of supporting the
|
|
codec. Using a native codec should allow greater portability
|
|
of your code to other architectures besides x86
|
|
and does not require your users to have Windows DLLs available.
|
|
|
|
Determining item 3, the correct WAVEFORMATEX structure
|
|
for the codec, is the hardest part. If your code is
|
|
simply playing files encoded by the codec this is easy.
|
|
Your code must register the mapping between format tags
|
|
and DLL filenames (items 1 and 2 above) by making repeated
|
|
calls to MSACM_RegisterDriver(). Note that you will
|
|
often have to call MSACM_RegisterDriver() twice for each
|
|
codec, once to register it for the codec's format tag for
|
|
decoding, and once to register it with the PCM format tag
|
|
for encoding. Then to play a file you read the .wav header
|
|
from the beginning of the file and use that to create a
|
|
WAVEFORMATEX structure which you pass to the acmStreamOpen()
|
|
function. From there on out everything can be the same
|
|
as the example program.
|
|
|
|
On the other hand, if your user is choosing the codec to
|
|
use for some encoding tasks, you do not have a .wav header
|
|
to read. The easiest thing to do in this case is probably
|
|
to create a .wav file in the desired format under Windows using
|
|
the Windows sound recorder. Then examine this file with
|
|
a hex editor and determine the appropriate values of the
|
|
fields to use in the WAVEFORMATEX structure. Note that
|
|
most codecs use extra fields beyond the basic WAVEFORMATEX
|
|
structure. test_truespeech.c shows how to create an extended
|
|
length WAVEFORMATEX and how to initialize the extra parts
|
|
for the TrueSpeech codec. A thorough writeup of the .wav
|
|
format headers is at
|
|
http://www.borg.com/~jglatt/tech/wave.htm
|
|
|
|
|
|
Most codecs will support 16-bit signed linear PCM, but probably
|
|
do not support a direct conversion to another proprietary
|
|
codec. To convert from one proprietary codec format to
|
|
another, typically a conversion to an intermediate 16-bit PCM
|
|
form is required.
|
|
|