No Sound with SDL_mixer - c++

I read and tried all the other posts to that topic, but nothing helped. When I try to play music with Mix_PlayChannel() I don't get an error nor do I hear some sound! I tried for hours now and nothing helps. The program just finishes happily. But no sound! I am using Ubuntu 12.04 64bit.
Thanks!
[EDIT]
Here is the code I use:
#include <iostream>
#include <SDL/SDL.h>
#include <SDL/SDL_mixer.h>
int main(int argc, char** argv) {
Mix_Music *music = NULL;
Mix_Chunk *wave = NULL;
SDL_Init(SDL_INIT_AUDIO);
int audio_rate = 44100;
Uint16 audio_format = AUDIO_S16; /* 16-bit stereo */
int audio_channels = 1;
int audio_buffers = 4096;
if(Mix_OpenAudio(audio_rate, audio_format, audio_channels, audio_buffers) < 0) {
printf("Unable to open audio!\n");
exit(1);
}
if(Mix_Init(MIX_INIT_MOD) != MIX_INIT_MOD)
std::cout << "errer";
Mix_Volume(-1, MIX_MAX_VOLUME);
music = Mix_LoadMUS("1.wav");
wave = Mix_LoadWAV("1.wav");
if (music == NULL) {
std::cout << "Could not load 1.wav\n";
std::cout << Mix_GetError();
}
if (wave == NULL) {
std::cout << "Could not load 1.wav\n";
std::cout << Mix_GetError();
}
Mix_VolumeChunk(wave, MIX_MAX_VOLUME);
Mix_VolumeMusic(MIX_MAX_VOLUME);
Mix_PlayMusic(music, 0);
std::cout << Mix_GetError();
Mix_FadeInChannelTimed(-1, wave, 0, 100,1);
std::cout << Mix_GetError();
return 1;
}
I try both PlayMusic() and Mix_FadeInChannelTimed(). Both files are loaded correctly but not played. Sound is not muted, wav-file is playable with aplay or other tools. I check with alsamixer that all channels are open and not too low.

I now found out that the program needs to run until the sound has finished playing! I added a usleep() command after the play command and it plays nicely. So that was really mentioned nowhere that PlayMusic() does not keep running.

Related

Possible to allow key bindings with MPV C API when no video (GUI) is being shown?

I am creating a background music player and I wanted to use the MPV C Plugin to do so, but my problem arrives when I disable displaying the video (with check_error(mpv_set_option_string(ctx, "vid", "no"));, this does the job of disabling the video, but then I can't use keys (like q (quit) or > (skip)) anymore... How do I allow them to be used in the terminal without the video GUI?
My Code:
#include <iostream>
#include <mpv/client.h>
static inline void check_error(int status)
{
if (status < 0)
{
std::cout << "mpv API error: " << mpv_error_string(status) << std::endl;
exit(1);
}
}
int main(int argc, char *argv[])
{
if (argc != 2)
{
std::cout << "pass a single media file as argument" << std::endl;
return 1;
}
mpv_handle *ctx = mpv_create();
if (!ctx)
{
std::cout << "failed creating context" << std::endl;
return 1;
}
check_error(mpv_set_option_string(ctx, "input-default-bindings", "yes"));
mpv_set_option_string(ctx, "input-vo-keyboard", "yes");
int val = 1;
check_error(mpv_set_option(ctx, "osc", MPV_FORMAT_FLAG, &val));
check_error(mpv_initialize(ctx));
const char *cmd[] = {"loadfile", argv[1], NULL};
check_error(mpv_command(ctx, cmd));
// THIS IS WHAT I USE TO DISABLE THE VIDEO
// check_error(mpv_set_option_string(ctx, "vid", "no"));
// Let it play, and wait until the user quits.
while (1)
{
mpv_event *event = mpv_wait_event(ctx, 10000);
std::cout << "event: " << mpv_event_name(event->event_id) << std::endl;
if (event->event_id == MPV_EVENT_SHUTDOWN)
break;
}
mpv_terminate_destroy(ctx);
return 0;
}
As you can see with mpv_set_option_string(ctx, "input-default-bindings", "yes") I allow it to use keybinding, but how do I make the keybinding works with just the terminal, since it only works when the GUI is visible? If you ran: mpv path/to/video.mp3 --no-video then the key bindings would still work fine, even without the video GUI.

Loading sound with SDL2.0 causes exception in MS Visual Studio

I have the following sound loading function:
bool Sound::load(const char* fileName)
{
sound = Mix_LoadWAV(fileName);
if (sound == nullptr)
{
std::cout << "Failed to load sound at path: " << fileName << "SDL_mixer error: " << Mix_GetError();
return false;
}
return true;
}
In Dev-C++, this works, fine. I wanted to use another IDE, so I started using Visual Studio 2017, and configured SDL2 for it. However, when I run this code, from the moment Mix_LoadWAV is called, Visual studio gives the following:
https://imgur.com/a/STnXx
I have searched alot on the internet, but I could not find anything useful that worked for me.
EDIT: as per request, I created a minimal example that still produces the same error.
#include "SDL2\SDL_mixer.h"
#include "SDL2\SDL.h"
#include <iostream>
#undef main
class SoundTest
{
private:
Mix_Chunk* sound = nullptr;
public:
bool load(const char* fileName)
{
sound = Mix_LoadWAV(fileName);
if (sound == nullptr)
{
std::cout << "Failed to load sound at path: " << fileName << "SDL_mixer error: " << Mix_GetError();
return false;
}
return true;
}
};
SDL_Window *window = nullptr;
int main(int argc, char** argv)
{
SDL_Init(SDL_INIT_EVERYTHING);
window = SDL_CreateWindow("Sound bug? ", SDL_WINDOWPOS_UNDEFINED, SDL_WINDOWPOS_UNDEFINED, 300, 300, SDL_WINDOW_SHOWN);
Mix_OpenAudio(22050, MIX_DEFAULT_FORMAT, 2, 2048);
SoundTest sound;
sound.load("audio/soundEffect.wav");
while (true)
{
//Do fun stuff with the sound
}
return 0;
}

SDL_mixer no sound for certain mp3 files

So I'm having issues with SDL_mixer, and it seems it plays fine on a good amount of mp3s (although "fine". I noticed that at the beginning of some MP3s there's a small segment of sound corruption, and then it starts playing) but on some MP3 files, there's just absolute silence. It doesn't seem to play anything. And Mix_PlayMusic doesn't seem to return an error, nor does anything else, but it just seems to not make a single sound. Occasionally I'll get a random "Access Reading Violation" when trying to play those files, but every other time, it's just silence.
Here is the code that I'm using to play the MP3 file:
#include "SDL.h"
#include "SDL_mixer.h"
#include <iostream>
bool initLibs()
{
//Initialize all SDL subsystems
if( SDL_Init( SDL_INIT_EVERYTHING ) == -1 )
{
return false;
}
int flags = MIX_INIT_MP3 | MIX_INIT_OGG;
if(Mix_Init(flags)!=flags)
{
return false;
}
//Initialize SDL_mixer
if( Mix_OpenAudio(MIX_DEFAULT_FREQUENCY, AUDIO_S16, 2, 4096 ) == -1 )
{
return false;
}
//If everything initialized fine
return true;
}
int main(int argc, char* argv[])
{
if(!initLibs())
{
std::cout << "error loading libraries" << std::endl;
}
Mix_Music* music = Mix_LoadMUS("test.mp3");
if(music == nullptr)
{
std::cout << "error loading music: " << Mix_GetError() << std::endl;
}
else
{
if(Mix_PlayMusic(music, -1) == -1)
{
std::cout << "error playing music: " << Mix_GetError() << std::endl;
}
}
SDL_Delay(30000);
system("PAUSE");
Mix_HaltMusic();
Mix_FreeMusic(music);
system("PAUSE");
return 0;
}
I was examining the differences in the MP3 files, and it seems that the only real difference is that the ones with LAME3.99 encoding are the ones that don't seem to work. Is anybody able to explain this?
EDIT: After more testing, it also seems to happen on other ones

VideoStream::setVideoMode() function doesn't work

I want to change VideoStream setting in my program, but it doesn't work
#include <OpenNI.h>
int main()
{
OpenNI::initialize();
Device device;
device.open(ANY_DEVICE);
VideoStream depthStream;
depthStream.create(device, SENSOR_DEPTH);
depthStream.start();
VideoMode depthMode;
depthMode.setFps(20);
depthMode.setResolution(640, 480);
depthMode.setPixelFormat(PIXEL_FORMAT_DEPTH_100_UM);
depthStream.setVideoMode(depthMode);
...
}
Even I change depthStream.start() line after setVideoMode() function, but still doesn't work.
I changed Fps to 24, 20, 5, 1 but it doesn't change anything.
p.s. : This is my simple code, without error handling.
Edit:
Answer:
with the help of dear "api55" i found that my device (Kinect Xbox) support only one mode of videoMode. so I can't change it.
My only supported video is :
FPS:30
Width:640
Height:480
I change the VideoMode succesfully in a code a did before. After creating the VideoStream you should do something like:
rc = depth.create(device, openni::SENSOR_DEPTH);
if (rc != openni::STATUS_OK)
error_manager(3);
// set the new resolution and fps
openni::VideoMode depth_videoMode = depth.getVideoMode();
depth_videoMode.setResolution(frame_width,frame_height);
depth_videoMode.setFps(30);
depth.setVideoMode(depth_videoMode);
rc = depth.start();
if (rc != openni::STATUS_OK)
error_manager(4);
First I get the VideoMode that is inside the stream to conserve the other values and only change what I wanted. I think your code should work, but not all settings work in all cameras. To check the possible settings you can use the function openni::VideoStream::getSensorInfo. The code to check this should be something like:
#include <OpenNI.h>
int main()
{
OpenNI::initialize();
Device device;
device.open(ANY_DEVICE);
VideoStream depthStream;
depthStream.create(device, SENSOR_DEPTH);
depthStream.start();
SensorInfo& info = depthStream.getSensorInfo();
Array& videoModes = info.getSupportedVideoModes();
for (int i = 0; i < videoModes.getSize(); i++){
std::cout << "VideoMode " << i << std::endl;
std::cout << "FPS:" << videoModes[i].getFps() << std::endl;
std::cout << "Width:" << videoModes[i].getResolutionX() << std::endl;
std::cout << "Height:" << videoModes[i].getResolutionY() << std::endl;
}
...
}
I haven't test this last piece of code, so it may have errors, but you get the idea of it. The supported settings change with each camera, but I think the supported FPS in my camera were 15 and 30.
I hope this helps you

Using OpenCV, Boost threading and multiple cameras

I am trying to write a program that is able to capture images from two different cameras in two different threads. I want to do this because when I do this in the same thread I have to keep waiting for cvQueryFrame twice the amount of time and so i can not grab images at 30 fps (I get 15 FPS from each camera).
I have taken a look at this SO post, but this only works for one camera. Using cvQueryFrame and boost::thread together
My current program gives varying results, sometimes it gives memory leaks, usually I just don't see anything happening and sometimes it worls for a few seconds but the the image freezes again. The strange thing is that earlier when I didn't call cvShowImage, but had my imageProcessing function do something useful I could see that I was getting real time results from both cameras. I assume this means that it is possible to make this work, but that I made a stupid mistake somewhere. My OS is LINUX and I am using OpenCV 2.4
My code:
#include <iostream>
#include <cstdio>
#include <cv.h>
#include <ml.h>
#include <cvaux.h>
#include <highgui.h>
#include <vector>
#include <stdio.h>
#include "producer_consumer_queue.hpp"
//Camera settings
int cameraWidth = 1280;
int cameraHeight = 720;
int waitKeyValue = 5;
bool threads_should_exit = false;
CvCapture * capture;
CvCapture * capture2;
using namespace std;
using namespace cv;
void grabFrame(concurrent_queue<IplImage* > * frame_queue, int camNumber) {
try {
//Load first frames
cout << "grabFrame: " << camNumber << " init with " << cameraWidth << " x " << cameraHeight << endl;
IplImage* frame;
if (camNumber == 0)frame = cvQueryFrame(capture);
if (camNumber == 1)frame = cvQueryFrame(capture2);
while (frame && !threads_should_exit) {
if (camNumber == 0)frame = cvQueryFrame(capture);
if (camNumber == 1)frame = cvQueryFrame(capture2);
IplImage* frame_copy = NULL;
frame_copy = cvCloneImage(frame);
if (camNumber == 0)cvShowImage("NE", frame);
cout << "grabFrame: " << camNumber << " pushing back to queue" << endl;
frame_queue->push(frame_copy);
int k = cvWaitKey(waitKeyValue);
if (k == 1048603 || k == 27 || k == '\r') {
cout << "grabFrame: Process killed" << endl;
//release memory
threads_should_exit = true;
}
}
} catch (const concurrent_queue<IplImage* >::Canceled & e) {
cout << "grabFrame: Show thread is canceled" << endl;
return;
}
}
void processFrames(concurrent_queue<IplImage* > * frame_queue0, concurrent_queue<IplImage* > * frame_queue1) {
try {
do {
cout << "processFrames: Processing two frames" << endl;
IplImage* frm = NULL;
frame_queue0->wait_and_pop(frm);
IplImage * frm2 = NULL;
frame_queue1->wait_and_pop(frm2);
cvReleaseImage(&frm);
cvReleaseImage(&frm2);
} while (!threads_should_exit);
} catch (const concurrent_queue<IplImage* >::Canceled & e) {
cout << "processFrames: Processing thread is canceled" << endl;
return;
}
}
int main() {
capture = cvCreateCameraCapture(0);
capture2 = cvCreateCameraCapture(1);
cvSetCaptureProperty(capture, CV_CAP_PROP_FRAME_WIDTH, cameraWidth);
cvSetCaptureProperty(capture, CV_CAP_PROP_FRAME_HEIGHT, cameraHeight);
cvSetCaptureProperty(capture2, CV_CAP_PROP_FRAME_WIDTH, cameraWidth);
cvSetCaptureProperty(capture2, CV_CAP_PROP_FRAME_HEIGHT, cameraHeight);
boost::thread_group frame_workers;
boost::thread_group frame_workers2;
concurrent_queue<IplImage* > frame_queue(&frame_workers);
concurrent_queue<IplImage* > frame_queue2(&frame_workers2);
boost::thread * query_thread = new boost::thread(processFrames, &frame_queue, &frame_queue2);
boost::thread * cam0_thread = new boost::thread(grabFrame, &frame_queue, 0);
usleep(10000);
boost::thread * cam1_thread = new boost::thread(grabFrame, &frame_queue2, 1);
frame_workers.add_thread(query_thread);
frame_workers.add_thread(cam0_thread);
frame_workers2.add_thread(query_thread);
frame_workers2.add_thread(cam1_thread);
while (true) {
if (threads_should_exit) {
cout << "Main: threads should be killed" << endl;
while (!frame_queue.empty()) {
usleep(10000);
}
frame_workers.remove_thread(query_thread);
frame_workers2.remove_thread(query_thread);
frame_workers.remove_thread(cam0_thread);
frame_workers2.remove_thread(cam1_thread);
frame_workers.join_all();
break;
}
usleep(10000);
}
return 0;
}
EDIT:
I added a simple function to detect a piece of paper to see if everything is working fine when I don't call on cvShowImage(). My program can detect a piece of paper fine if I don't call cvShowImage(). If I do the program has strange behavior again and freezes etc.
There should only be one thread manipulating the GUI (this is true for just about any GUI framework). You should organize your code so that cvShowImage is only invoked from the main "GUI thread".
It seems that the work that's being done in query_thread could just as easily be done inside the main thread.