How to use SDL or ALSA to mix multiply audio stream?

Question

0.00/5 (No votes)

See more:

I have played .wav file, using SDL_mixer library. But i don't know how to mix multiply audio stream. Please suggest for me!

Posted 9-Jan-12 17:57pm

mot sach

Add a Solution

Comments

Amir Mahfoozi 10-Jan-12 0:44am

When I was younger I sent voice with ALSA over network in Linux :). If their two buffers are the same size try to compute each corresponding bytes average and send the result, I think it will mix two sounds together. BTW you can add weight to them to amplify one channel more than the other one.

mot sach 10-Jan-12 1:12am

thank you, i will try to do that, and another question, how to get raw data from .wav file? (I have gotten header by reading from fist byte (byte 0) to 43th byte, but i'm not sure raw data beginning from 44th byte.)

Amir Mahfoozi 10-Jan-12 1:21am

When you speak about WAV you speak about vast variety of formats! There are some of them which are compressed and to some extent are mp3! So do not get confused with them and just search for libraries that work with old and uncomplicated ones. There are a lot of loaders which can be found by Google.

mot sach 10-Jan-12 1:59am

oh yeah, i have a little confusion, i think .wav file only contains raw data. Can load raw data from local file by using SDL_mixer lib, but i don't know how to mix multiply stream

Amir Mahfoozi 10-Jan-12 2:47am

I speculate that it can be done by computing average of each corresponding two bytes. If the sample rates and buffer lengths for two streams are the same.

mot sach 10-Jan-12 3:36am

yes, thank you so much, i understand that you say. No problem if play local file (because SDL_mixer support load raw data from local file, particular SDL_LoadWAV() function). But assume have a server send to client an audio stream. Stream's data is not raw data (such as .wav file's data), client must decode data getted, and then play audio exactly in real time. I don't know that. Please suggest for me (code example is so great)!

Amir Mahfoozi 10-Jan-12 4:18am

As I remember I found a client server UDP chat sample and changed it to send a buffer then found a Wave Play sample program and split it from the point that it wanted to send data to sound driver I mean I transferred reader part into sever program and transferred playing part to the client part then mixed all together and suddenly it worked ! Remember to use UDP sending method because it doesn't need establishing connection or error checking ;) Good Luck

mot sach 10-Jan-12 22:00pm

yes, i have a little knowledge about UDP, and it isn't a important problem. I think UDP server will don't send raw data because consumption of resource. Server must decode data, send more data to client (raw data is greater than data compressed). So client will get data compressed, then decode -> need to get raw data from stream. I'm not sure, could you like answer for me?

Amir Mahfoozi 11-Jan-12 0:16am

No, UDP is connectionless and consume less resources than TCP which is connection oriented. And neither UDP nor TCP doesn't care what do you give them to send. Just give them a buffer and they send it for you to the target host whether it is RAW data or compressed data.

mot sach 11-Jan-12 3:00am

yes, thank so much.

3 solutions

Solution 3

so sorry, but i wanna ask you another question. If i use SDL_mixer lib to play sound, i need to allocate a large buffer of Mix_Chunk.

C++

void play (uint8_t* stream, uint32_t len)
{
    int channel;
    int latency = len;          // size of buffer
    //init chunk
    Mix_Chunk* chunk = (Mix_Chunk*) malloc (sizeof(Mix_Chunk));
    chunk-&gt;abuf = (uint8_t*) malloc(latency);    // len is very big
    chunk-&gt;alen = len;
    chunk-&gt;alloceted = true;
    //copy data from stream to chunk
    memcpy(chunk-&gt;abuf, stream, len);
    //play audio
    channel = Mix_PlayChannel(-1, chunk, 0);
    while (Mix_Playing(channel); //while finished playing
}

How to reduce "latency"?

Posted 11-Jan-12 22:29pm

mot sach

Solution 4

yes, thank so much, this is my solution:

C#

unsigned int    period_size = 4096;     // size of buffer
//init mix chunk
Mix_Chunk*  sound = (Mix_Chunk*) malloc(sizeof(Mix_Chunk);

sound-&gt;alen = 4096;
sound-&gt;abuf = (uint8_t*) malloc(period_size);
sound-&gt;allocated = true;

void play(data, len)
{
    while (len) {
        //copy data to chunk
        if (len &gt; period_size) {
            memcpy(sound-&gt;abuf, data + index, sound-&gt;alen);
            len -= sound-&gt;alen;
                index += sound-&gt;alen;
        }
        else {
            memcpy(sound-&gt;abuf, data + index, len);
            len -= len;
            index += len;
        }
        //Play our sound file, and capture the channel on which it is played
        channel2 = Mix_PlayChannel(-1, sound2, 0);
        if (channel2 == -1) {
            printf("Unable to play WAV file: %s\n", Mix_GetError());
            exit(-1);
        }
        //Wait until the sound has stopped playing
        while(Mix_Playing(channel2) != 0);
    }
}

Sequentially, I has copied data from stream by block (4096 bytes), then waiting finished play it . It work ok. But i think this is not good ideal. Do you think so

Posted 12-Jan-12 16:33pm

mot sach

Add a Solution

Add your solution here

Treat my content as plain text, not as HTML

Preview 0

…

Existing Members

Sign in to your account

...or Join us

Download, Vote, Comment, Publish.

Your Email
Password
Forgot your password?

Your Email
This email is in use. Do you need your password?
Optional Password

I have read and agree to the Terms of Service and Privacy Policy
Please subscribe me to the CodeProject newsletters

When answering a question please:

Read the question carefully.
Understand that English isn't everyone's first language so be lenient of bad spelling and grammar.
If a question is poorly phrased then either ask for clarification, ignore it, or edit the question and fix the problem. Insults are not welcome.
Don't tell someone to read the manual. Chances are they have and don't get it. Provide an answer or move on to the next question.

Let's work to help developers, not make them feel stupid.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

Amir Mahfoozi · Accepted Answer · 2012-01-10T18:21:00

Solution 2

for examples of establishing UDP and TCP connections read here :
http://www.tenouk.com/Module40c.html[^]

And just for UDP read here :
http://www.abc.se/~m6695/udp.html[^]

And here is a sample wave player with SDL :
http://content.gpwiki.org/index.php/SDL_mixer:Tutorials:Playing_a_WAV_Sound_File[^]

Just move its reader part to the server and player part to the client (or vice versa). It's important to make sure that each module is working properly and then mix them together.

Hope it helps.

Posted 10-Jan-12 18:21pm

Amir Mahfoozi

Updated 10-Jan-12 23:45pm

v2

Comments

RaisKazi 11-Jan-12 6:12am

My 5.

Amir Mahfoozi 11-Jan-12 6:17am

Thanks Rais

mot sach 12-Jan-12 4:27am

so sorry, but i wanna ask you another question. If i use SDL_mixer lib to play sound, i need to allocate a large buffer of Mix_Chunk.
<pre lang="c++">
void play (uint8_t* stream, uint32_t len)
{
int channel;
int latency = len; // size of buffer
//init chunk
Mix_Chunk* chunk = (Mix_Chunk*) malloc (sizeof(Mix_Chunk));
chunk->abuf = (uint8_t*) malloc(latency); // len is very big
chunk->alen = len;
chunk->alloceted = true;
//copy data from stream to chunk
memcpy(chunk->abuf, stream, len);
//play audio
channel = Mix_PlayChannel(-1, chunk, 0);
while (Mix_Playing(channel); //while finished playing
}
</pre>

How to reduce "latency"?

Amir Mahfoozi 12-Jan-12 4:41am

If I were you I was finishing the job with a small file and then would focus on read the file in small chunks. So please do not allow this problem stop you from doing the whole job. However I will do a try to find out how a file could be read in chunks in SDL. If there was no solution SDL could be replaced by another library. Just reach to a small working client server program and then solve those problems.

Amir Mahfoozi 12-Jan-12 4:46am

Have a look at this page :
http://sdl.beuc.net/sdl.wiki/Audio_Examples
It have mixed two buffers together and used callbacks to avoid to load the whole file.
But firstly try to solve the problem with simplified assumptions, for example you have just one stream and small file ...

mot sach 12-Jan-12 5:38am

yes, if small file is ok, but if large file, can't allocate large buffer, so i think must split stream into smaller block, then write data to hardware using SDL lib. But with this solution, i met synchronization problem. i can't solve this. Can you help me?

Amir Mahfoozi 12-Jan-12 6:40am

Dear Mot, I want to help you but I'm heavily engaged in a project and have no time and I come here in rest time to share my information but I told you as much as I have experience in this matter. So do not get disappointed and try to solve your problem by simplifying its requirement. For example do the whole job with small files then go for large files.

mot sach 12-Jan-12 21:32pm

yes, thank so much, this is my solution:
<pre lang="c++">
unsigned int period_size = 4096; // size of buffer
//init mix chunk
Mix_Chunk* sound = (Mix_Chunk*) malloc(sizeof(Mix_Chunk);

sound->alen = 4096;
sound->abuf = (uint8_t*) malloc(period_size);
sound->allocated = true;

void play(data, len)
{
while (len) {
//copy data to chunk
if (len > period_size) {
memcpy(sound->abuf, data + index, sound->alen);
len -= sound->alen;
index += sound->alen;
}
else {
memcpy(sound->abuf, data + index, len);
len -= len;
index += len;
}
//Play our sound file, and capture the channel on which it is played
channel2 = Mix_PlayChannel(-1, sound2, 0);
if (channel2 == -1) {
printf("Unable to play WAV file: %s\n", Mix_GetError());
exit(-1);
}
//Wait until the sound has stopped playing
while(Mix_Playing(channel2) != 0);
}
}
</pre>

Sequentially, I has copied data from stream by block (4096 bytes), then waiting finished play it . It work ok. But i think this is not good ideal. Do you think so