gmane.comp.audio.portaudio.devel

wasapi loopback devices detection?

Tony Miller — 2013-06-17T18:08:54

I applied the patch I found in this message:
http://music.columbia.edu/pipermail/portaudio/2013-January/014834.html

I tried running the example pa_devs.c to see my loopback devices but I
didn't see any listed. Should the pa_devs example list them or do I
need to change the code?

I am pretty certain I have some loopback devices on this machine, I've
seen them listed in other audio software.

Thanks.

Problem with Pa_WriteStream with MOTU 896HD

De Sena, Enzo — 2013-06-08T21:50:41

Dear all,

I am having problems running portaudio with my MOTU sound card.
I use the blocking call Pa_WriteStream with the code attached below. 
The code runs properly (i.e. output all 0) with all audio interfaces except for my MOTU 896HD, 
in which case it hangs at Pa_WriteStream without ever outputting "run".

I compile it with the latest Xcode (4.6.2), latest MOTU drivers, latest stable portaudio.
The cpp project in Xcode is left with all standard options.
The sound card (set at 44.1KHz with internal clock) is connected via firewire 
to my Mac Pro running OSX 10.8.4.

I tried to run Matlab's playrec (which is based on portaudio) and, interestingly, 
it works well with the MOTU sound card too.

Any suggestion/help is greatly appreciated! Thanks!
And many apologies in advance in case this is too much of a newbie question..

Best,


Enzo



------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------

#include "portaudio.h"
#include <iostream>

void PaErrorHandler(PaError, PaStream*);


int main(int argc, char *argv[]) {
  const double sampling_frequency = 44100;
  
  PaStream *stream;
  PaError err;
  
  std::cout<<"Starting portaudio...\n";std::cout.flush();
  
  err = Pa_Initialize();
  if( err != paNoError ) { PaErrorHandler(err, NULL); }
  
  const   PaDeviceInfo *deviceInfo;
  int numDevices = Pa_GetDeviceCount();
  for(int i=0; i<numDevices; ++i) {
    deviceInfo = Pa_GetDeviceInfo( i );
    std::cout<<"Device id:"<<i<<", name: "<<deviceInfo->name<<
    ", max output channels: "<<deviceInfo->maxOutputChannels<<"\n";
  }
  std::cout<<"Select the device id: ";
  int outDevNum;
  std::cin >> outDevNum;
  const int max_num_channels(Pa_GetDeviceInfo(outDevNum)->maxOutputChannels);
  
  
  
  PaStreamParameters outputParameters;
  bzero( &outputParameters, sizeof( outputParameters ) );
  outputParameters.channelCount = Pa_GetDeviceInfo(outDevNum)->maxOutputChannels;
  outputParameters.device = outDevNum;
  outputParameters.hostApiSpecificStreamInfo = NULL;
  outputParameters.sampleFormat = paFloat32;
  outputParameters.suggestedLatency = Pa_GetDeviceInfo(outDevNum)->defaultLowOutputLatency ;
  outputParameters.hostApiSpecificStreamInfo = NULL; //See you specific host's API docs for info on using this field
  
  const int frames_per_buffer = 200;
  
  /* Open an audio I/O stream. */
  err = Pa_OpenStream(&stream,
                      NULL,
                      &outputParameters,
                      sampling_frequency,
                      frames_per_buffer,
                      paNoFlag, //flags that can be used to define dither, clip settings and more
                      NULL, //your callback function
                      NULL); //data to be passed to callback. In C++, it is frequently (void *)this
  if(err != paNoError) { PaErrorHandler(err, stream); }
  
  err = Pa_StartStream(stream);
  if(err != paNoError) { PaErrorHandler(err, stream); }
  
  float sample_block[frames_per_buffer*max_num_channels];
  for (int i=0; i<frames_per_buffer*max_num_channels; ++i) {
    sample_block[i] = (float)0.0;
  }
  const int num_frames = 1000;
  
  for (int i=1; i<num_frames; ++i) {
    
    err = Pa_WriteStream( stream, sample_block, frames_per_buffer );
    if( err != paNoError ) { PaErrorHandler(err, stream); }
    std::cout<<"run"<<std::endl;
  }
  
  err = Pa_StopStream( stream );
  if( err != paNoError ) { PaErrorHandler(err, stream); }
  
  Pa_Terminate();
  return 0;
}




void PaErrorHandler(PaError err, PaStream* stream) {
  fprintf( stderr, "An error occured while using the portaudio stream\n" );
  fprintf( stderr, "Error number: %d\n", err );
  fprintf( stderr, "Error message: %s\n", Pa_GetErrorText( err ) );
  if( stream ) {
    Pa_AbortStream( stream );
    Pa_CloseStream( stream );
  }
  Pa_Terminate();
  exit(1);
}


---------------------------------------------------------
Enzo De Sena 
Ph.D. research student
Institute of Telecommunications
King's College London
Strand, London, WC2R 2LS, UK
+44 (0) 20 7848 1138
www.enzo.desena.name
enzo.desena< at >kcl.ac.uk

[PATCH] coreaudio: suppress deprecated API warnings

Stefan Hajnoczi — 2013-06-08T07:43:38

Deprecated APIs are use intentionally to preserve compatibility with older Mac
OS X releases.  Since -Werror is used this would normally result in a
compilation failure.  Tell gcc to ignore deprecated APIs for coreaudio hostapi
files that make use of them.

See also ticket #218.

Signed-off-by: Stefan Hajnoczi <stefanha< at >gmail.com>
---
 src/hostapi/coreaudio/pa_mac_core.c           | 2 ++
 src/hostapi/coreaudio/pa_mac_core_utilities.c | 3 +++
 2 files changed, 5 insertions(+)

diff --git a/src/hostapi/coreaudio/pa_mac_core.c b/src/hostapi/coreaudio/pa_mac_core.c
index 1cef082..847721e 100644
--- a/src/hostapi/coreaudio/pa_mac_core.c
+++ b/src/hostapi/coreaudio/pa_mac_core.c
< at >< at > -71,6 +71,8 < at >< at >
 #include "pa_mac_core_utilities.h"
 #include "pa_mac_core_blocking.h"
 
+/* Use deprecated APIs for compatibility with older OS X versions */
+#pragma GCC diagnostic ignored "-Wdeprecated-declarations"
 
 #ifdef __cplusplus
 extern "C"
diff --git a/src/hostapi/coreaudio/pa_mac_core_utilities.c b/src/hostapi/coreaudio/pa_mac_core_utilities.c
index 63e616f..4929b6d 100644
--- a/src/hostapi/coreaudio/pa_mac_core_utilities.c
+++ b/src/hostapi/coreaudio/pa_mac_core_utilities.c
< at >< at > -63,6 +63,9 < at >< at >
 #include <pthread.h>
 #include <sys/time.h>
 
+/* Use deprecated APIs for compatibility with older OS X versions */
+#pragma GCC diagnostic ignored "-Wdeprecated-declarations"
+
 PaError PaMacCore_SetUnixError( int err, int line )
 {
    PaError ret;

Opening and closing streams frequently

Tyler Martin — 2013-06-07T18:08:46

Hey there,

I'm using portaudio in an app that needs to play both stereo and mono
signals.  Their playback doesn't overlap, however.  To play the signals
should I open a stream in mono and close it, and then open a stream in
stereo and close it, etc?  Or is it more correct to convert the mono into a
stereo signal so that it can be output on a stereo stream and I can avoid
frequently opening and closing a stream?

Thanks,
Tyler
_______________________________________________
Portaudio mailing list
Portaudio< at >music.columbia.edu
http://music.columbia.edu/mailman/listinfo/portaudio

Callback Full Duplex

Wouter Janmaat — 2013-06-02T20:14:06

Hi all,

I have a question on the callback in a situation when the device handles both input and output. Does portaudio differentiate between when the device has recorded new input and when the device needs new output? And if so, how can I make the callback do the right thing? It seems that when device has new input, the callback would also refill the outputbuffer. Would this cause any problems?

Best

Wouter

Random beep tone when opening stream

Wouter Janmaat — 2013-05-31T16:56:21

Hi all,

I use portaudio when opening my application. It opens up a stream for every audiodevice it finds which has either input or output channels. Sometimes when starting the application the program sends a loud beeping tone to the output of the soundcard. My callbuck function is empty and I use the settings from the tutorial to open the streams.

Any ideas?

For Port audio player

Nitin Rajput — 2013-05-30T12:25:36

Dear Sir/Madam,

i am working on sound editing in asp.net using c#. i have completed c++
function successfully calling  in my c#  application.

please help me how to use port audio player api for playing wave file and
edit them programmatically.

Paqa_latency print typos & QA check

Alan Horstmann — 2013-05-29T17:02:33

Phil,

While reading output from paqa_latency I noticed the printouts say 'Output'
when reporting for input (there is also one rogue tab).

Also there seemed no reason not to enable the QA_ASSERT check of default high
latency to be greater than the low (now // out), except that it should allow
them to be equal.  However, there may be a good reason!

The diff looks like this:-

--- ../svn-working-copy/qa/paqa_latency.c2013-04-27 15:29:30.000000000 +0100
+++ qa/paqa_latency.c2013-05-29 17:55:36.000000000 +0100
< at >< at > -255,7 +255,7 < at >< at >
     {
         numLoops = 1;
     }
-
+
     for( i=0; i<numLoops; i++ )
     {   
         if( numLoops == 1 )
< at >< at > -361,15 +361,15 < at >< at >
             printf("  Output defaultHighOutputLatency = %f seconds\n", pdi->defaultHighOutputLatency);
             QA_ASSERT_TRUE( "defaultLowOutputLatency should be > 0", (pdi->defaultLowOutputLatency > 0.0) );
             QA_ASSERT_TRUE( "defaultHighOutputLatency should be > 0", (pdi->defaultHighOutputLatency > 0.0) );
-            //QA_ASSERT_TRUE( "defaultHighOutputLatency should be > Low", (pdi->defaultHighOutputLatency > pdi->defaultLowOutputLatency) );
+            QA_ASSERT_TRUE( "defaultHighOutputLatency should be >= Low", (pdi->defaultHighOutputLatency >= pdi->defaultLowOutputLatency) );
         }
         if( pdi->maxInputChannels > 0 )
         {
-            printf("  Input defaultLowOutputLatency  = %f seconds\n", pdi->defaultLowInputLatency);
-            printf("  Input defaultHighOutputLatency = %f seconds\n", pdi->defaultHighInputLatency);
-            QA_ASSERT_TRUE( "defaultLowOutputLatency should be > 0", (pdi->defaultLowInputLatency > 0.0) );
-            QA_ASSERT_TRUE( "defaultHighOutputLatency should be > 0", (pdi->defaultHighInputLatency > 0.0) );
-            //QA_ASSERT_TRUE( "defaultHighOutputLatency should be > Low", (pdi->defaultHighInputLatency > pdi->defaultLowInputLatency) );
+            printf("  Input defaultLowInputLatency  = %f seconds\n", pdi->defaultLowInputLatency);
+            printf("  Input defaultHighInputLatency = %f seconds\n", pdi->defaultHighInputLatency);
+            QA_ASSERT_TRUE( "defaultLowInputLatency should be > 0", (pdi->defaultLowInputLatency > 0.0) );
+            QA_ASSERT_TRUE( "defaultHighInputLatency should be > 0", (pdi->defaultHighInputLatency > 0.0) );
+            QA_ASSERT_TRUE( "defaultHighInputLatency should be >= Low", (pdi->defaultHighInputLatency >= pdi->defaultLowInputLatency) );
         }
     }
     return 0;

I can commit this change if you wish?

Regards

Alan
_______________________________________________
Portaudio mailing list
Portaudio< at >music.columbia.edu
http://music.columbia.edu/mailman/listinfo/portaudio

Functionality similar to LineIn passthrough?

Joshua Kordani — 2013-05-28T22:54:44

Greetings all.

I need to develop a replacement for the LineIn application for osx. It 
allows you to pipe line in into system output.  I'm wondering if 
portaudio is the place for me to make an application that does this, or 
not, as I am not entirely certain if this is the domain portaudio 
handles.  Thanks for any pointers.

Joshua Kordani
LSA Autonomy
_______________________________________________
Portaudio mailing list
Portaudio< at >music.columbia.edu
http://music.columbia.edu/mailman/listinfo/portaudio

Buffering Latency and Timing

Alan Horstmann — 2013-05-28T21:17:20

Ross,

The questions from Alexander have prompted me to pick-up some investigations I 
was doing some time ago on the Alsa and OSS latency values suggested and 
reported etc, and using paqa_latency.

I've been re-reading 'BufferingLatencyAndTimingImplementationGuidelines' and 
wanted to check my understanding.

The Alsa implementation is similar to a 'N-buffer ringbuffer' scheme; 2 common 
configurations arise, either with 4 buffers (=Alsa periods) in the 
ringbuffer, or with just 2.

By my reading the reported output latency should be the time of:
 (whole ringbuffer - one buffer(=period))  since the user should start filling 
when there is one buffer(=period) of space in the ringbuffer.

At present the Alsa code reports and uses the time delay of the whole 
ringbuffer, which seems to be a mistake?  With N=4 the latency I would expect 
to be 3/4 of the ringbuffer, and that is the figure OSS reports on a 
configured stream.  With N=2 then the latency is just one buffer(=period) 
long, or have I mis-understood?

Regards

Alan

PS

Wouter Janmaat — 2013-05-28T13:47:04

PS:

I experimented on this by trying to immediately fill the outputbuffer with whatever is in the inputbuffer by in the callback setting:

const float *in = (const float*)inputbuffer;
float *out = (float*)outputbuffer;

*out = *in;

This compiles and I see that the device is initiated but the callback is never called. In debug the errorcode is -9988

Simultaneous Recording/Playback Callback

Wouter Janmaat — 2013-05-28T13:10:01

Hi all,

I have a question about the way portaudio handles a callback when Pa_OpenStream is used with both input and output parameters.
Is the callback supplied in the Pa_OpenStream called when both audio is recorded and when the outputbuffer needs audio or do you need to seperate the recording and playing back of the file.

I would like to have one stream managing both the input and output.

Thanks,

WASAPI callback buffering mode

Matt McCallus — 2013-05-23T20:18:38

Hello!

I'm working on a music title that's using PortAudio (Win/WASAPI), and I had a few questions regarding I/O stream latency.  The title takes an audio stream from a USB audio device, processes the audio, and then sends it to the audio output device.

My goal is to have the lowest possible overall latency (input + output).  Therefore I'm wanting to use the WASAPI callback event interface, which supposedly can have 3ms latency on the input or output.

I've found that when I set up the input and output streams, if I pass any value of PaStreamParameters::suggestedLatency, then PortAudio will force the polling WASAPI callback interface.
I was under the impression that one was supposed to use suggestedLatency as a "hint" to Pa_OpenStream().  I got that idea from the document "Guidelines for implementing PortAudio buffering, latency, and timing".

So, if I set suggestedLatency to 0.0, then WASAPI will use the event interface, but if I try to specify any suggestedLatency other than 0.0, then it uses the polling interface (with supposedly much higher latency than the event interface).

Am I doing something wrong?  Why would I not want to use the WASAPI callback event interface?

What could I be doing differently to minimize latency?

Thank you very much for your help!

Matt McCallus
Audio Programmer
Ubisoft Red Storm

_______________________________________________
Portaudio mailing list
Portaudio< at >music.columbia.edu
http://music.columbia.edu/mailman/listinfo/portaudio

Controlling Recording Level with PortAudio

Neal — 2013-05-23T19:53:00

I am currently successfully running a Visual Studio Application that incorporates PortAudio to send and receive data from the soundcard. I want to be able to change the recording level via software. The only way I can currently change this level in Win7  is by right clicking the speaker icon->recording devices->select external mic->levels. 

Is there a way to change this level in software via Port Audio???

Neal

_______________________________________________
Portaudio mailing list
Portaudio< at >music.columbia.edu
http://music.columbia.edu/mailman/listinfo/portaudio

leak in pa_linux_alsa.c

Daniel Pinyol Laserna — 2013-05-22T16:09:53

Hi,
it looks like there's a memory leak in pa_linux_alsa.c.

In the past method PaAlsaStreamComponent_EndProcessing was the
responsible for freeing variable nonMmapBuffer, but now it's not freed
anymore because it's reused.

This is an issue if a portaudio stream is opened/closed many times, as
I did in the past. As a workaround, now I only open the stream once,
and I just start/stop it for any play, so that I have a single memory
leak. However, in case each sound that I play has a different
configuration, I would need to close/reopen the stream constantly, and
I'd have a leak for each sound play.

I guess it should be deleted at PaAlsaStream_Terminate.




--
salutacions

Dani Pinyol

Alfred Adler - "The only normal people are the one's you don't know very well."

Compiling Portaudio for OSX 10.8.3

Wouter Janmaat — 2013-05-13T14:39:02

Hi All,

it's my first time using PortAudio and I'm currently trying to compile for OSX 10.8.3 using the ./config && make commands from the tutorial but I keep getting errors. They look like this:

checking build system type... i386-apple-darwin12.3.0
checking host system type... i386-apple-darwin12.3.0
checking target system type... i386-apple-darwin12.3.0
checking for gcc... no
checking for cc... no
checking for cl.exe... no
configure: error: in `/Users/wouterjanmaat/Downloads/portaudio':
configure: error: no acceptable C compiler found in $PATH
See `config.log' for more details


I tried using the disable universal binaries flag but that didn't help.I tried both the latest nightly build and the latest stable build.

Please help me out

Thanks

Wouter

Linux Latency problem

Alexander Carôt — 2013-05-13T07:45:10

_______________________________________________
Portaudio mailing list
Portaudio< at >music.columbia.edu
http://music.columbia.edu/mailman/listinfo/portaudio

Portaudio and Remote Desktop Sessions

Geordi Sinclair — 2013-05-10T17:18:57

Hello,

I am using the latest version of Portaudio, from svn.  It is built into iaxclient.dll.  It works very well on Windows XP and Windows 7 , but I have having odd issues with the audio when using Remote Desktop Connection (from windows 7 64 bit) to a remote desktop session (Windows 2008 Server).  I allow the microphone and speaker of windows 7 to be available over the RDC.  I can see them on the RDS, as "Remote Audio", but when I use them, the audio is very broken up in both directions.

If I check the audio from the local mic to the RDS, via an audio recorder, the quality is fine.

Any ideas?

Thanks

Geordi
_______________________________________________
Portaudio mailing list
Portaudio< at >music.columbia.edu
http://music.columbia.edu/mailman/listinfo/portaudio

Buffer alignment requirement of various audio format

Pradeep Kumar — 2013-05-07T07:42:32

Hi All,

I've some question on requirement of alignment of data buffer used by
Pa_WriteStream or Pa_Callback.
I want to support for two audio format paInt16 and paFloat32.
So do I've to align my data buffer on 32byte boundary for paFloat32 and
16byte boundary of paInt16?

Also if I use Pa_RingBuffer in my application do I ever need to worry about
alignment or ring buffer internally take care of alignment?

Please share some thoughts.

Regards,
Pradeep
_______________________________________________
Portaudio mailing list
Portaudio< at >music.columbia.edu
http://music.columbia.edu/mailman/listinfo/portaudio

Pa_Initialize crash

Sergio Teran — 2013-05-06T10:32:43

Hello,

I am trying to use portaudio in my Qt desktop app to use the audio
IO functionality  but I am not being able to initialize the library.

My project will build fine but as soon as I execute it it will crash
automatically if I have the Pa_Initialize() call in it. If I comment it out
it continues with no problem.

I did a whole new blank project and I only have this in it:

MainWindow::MainWindow(QWidget *parent)

    : QMainWindow(parent)

{

    // Test UI with just one button

     button = new QPushButton(tr("OKK"), this);

    connect(button, SIGNAL(clicked()), this, SLOT(close()));


    PaError err;

    err = Pa_Initialize();


    if(err != paNoError){

        qDebug() << "Pa_Initialize passed ... ";

        return;

    }

    else

        qDebug() << "THERE WAS AN ERROR";

}


can anyone help me out to what could be the problem?

The error I get when the program stops is:


The program has unexpectedly finished.

C:\Qt\dev\Projects\build-PaTest_record-Desktop_Qt_5_0_2_MinGW_32bit-Debug\debug\PaTest_record
exited with code -1073741515


Any help greatly appreciated, Thanks.
_______________________________________________
Portaudio mailing list
Portaudio< at >music.columbia.edu
http://music.columbia.edu/mailman/listinfo/portaudio

Noise when writing silence in pa_callback

Pradeep Kumar — 2013-05-02T11:59:33

Hi,

I've an user application which does both audio and video processing. I'm
getting loud noise whenever the playback is interrupted in the middle i.e
if application is playing a video file and suddenly user initiate some
other task from application, then playback is interrupted to its current
state  till the new task finishes, Now during this interruption since I'm
not decoding any audio sample my ring buffer is empty and I'm writing
silence in callback. However I'm hearing loud noise instead.
Please suggest what could be going wrong.

Thanks and Regards,
Pradeep
_______________________________________________
Portaudio mailing list
Portaudio< at >music.columbia.edu
http://music.columbia.edu/mailman/listinfo/portaudio