I have an application which requires to use the microphone for recording user voice. I'm trying to make a speech to text.
I'm work with SpeechKit.framework and below is my code used:
-(void)starRecording{
self.voiceSearch = [[SKRecognizer alloc] initWithType:SKSearchRecognizerType
detection:SKShortEndOfSpeechDetection
language:[[USER_DEFAULT valueForKey:LANGUAGE_SPEECH_DIC] valueForKey:#"record"]
delegate:self];
}
- (void)recognizer:(SKRecognizer *)recognizer didFinishWithResults:(SKRecognition *)results {
long numOfResults = [results.results count];
if (numOfResults > 0) {
// update the text of text field with best result from SpeechKit
self.recordString = [results firstResult];
[self sendChatWithMediaType:#"messageCall" MediaUrl:#"" ContactDetail:#"{}" LocationDetail:#"{}"];
[self.voiceSearch stopRecording];
}
if (self.voiceSearch) {
[self.voiceSearch cancel];
}
[self starRecording];
}
That makes the SKRecognizer to be always open and that thing reduce the application performance.
I want to start the SKRecognizer when the microphone is detecting input audio.
I have a method for that? A method which is called when the microphone have input sound for me or a method which is always returning the level of audio detected?
Thank you!
You need to use the SpeechKit class to set up the audio.
Look here for details;
http://www.raywenderlich.com/60870/building-ios-app-like-siri
This project shows how to detect audio threshold;
github.com/picciano/iOS-Audio-Recoginzer
Related
This is the code I am using to play a sound with AVAudioPlayerNode. It is just playing the last sound of the beats array. On iPhone, all sounds from the beats array are playing simultaneously with the same function.
-(void)playMix{
for (int i = 0; i< mix.beatsArray.count;i++) {
beat = [[WatchBeatObject alloc] initWithFileName:mix.beatsArray[i].fileName];
[beat.audioPlayerNode play];
[beat.audioPlayerNode scheduleBuffer:beat.buffer atTime:nil options:AVAudioPlayerNodeBufferLoops completionHandler:^{
}];
}
}
N.B: Method initWithFileName: handles initializing and creating AVAudioPlayerNode and everything needed.
Thank you in advance.
Does that WatchBeatObject has any category option in AVAudioSession to set AVAudioSessionCategoryAmbient? If yes, set it, they would mix different sound.
I'm using UIImagePickerController to take pictures and videos from my app. Toggling between the two isn't too bad. If the user chooses to record a video, I first check this:
if (picker.cameraCaptureMode == UIImagePickerControllerCameraCaptureModeVideo)
{
[self captureVideo];
}
else
{
picker.cameraCaptureMode = UIImagePickerControllerCameraCaptureModeVideo;
[self captureVideo];
}
This usually works totally fine. Here's the catch. I'm also using OpenTok by Tokbox to do video calls, and it seems like the captureMode assignment doesn't work after a video call. It seems completely crazy, but I made this modification to do some debugging:
if (picker.cameraCaptureMode == UIImagePickerControllerCameraCaptureModeVideo)
{
[self captureVideo];
}
else
{
picker.cameraCaptureMode = UIImagePickerControllerCameraCaptureModeVideo;
if (picker.cameraCaptureMode != UIImagePickerControllerCameraCaptureModeVideo)
{
NSLog(#"Assignment unsuccessful???")
}
[self captureVideo];
}
And i get this "Assignment unsuccessful???" log every single time. UIImagePickerController must not be allowing the assignment or something. I really can't figure it out. I've also made a forum post on OpenTok's site to see if they're possibly not releasing some camera resources, but I don't think it's their problem.
Any insight here?
Use:
+ (NSArray *)availableCaptureModesForCameraDevice:(UIImagePickerControllerCameraDevice)cameraDevice
to check which source types are available. Also if you're using a simulator, it will never assign properly.
Solved with a solution on the TokBox forum. I needed to first change my audio session before trying to access the microphone.
AVAudioSession *mySession = [AVAudioSession sharedInstance];
[mySession setCategory:AVAudioSessionCategorySoloAmbient error:nil];
[self presentViewController:picker animated:YES completion:NULL];
I have following code which performs stop running (video) session, delete last segment of video, store video to directory and starts new video recording. There is (naturally) gap between these two video segments. Is there any way to optimize this code (maybe assync optimization if possible)? By optimization I mean elimination time gap as much as possible between these two video segments, thank you.
- (void) restartVideoRecording {
[captureSession removeOutput:captureMovieOutput];
[captureMovieOutput stopRecording];
if(lastPathWasOne){
captureMoviePath = [[URLPathProvider getUrlPathProvider]videoTwoPathString];
[URLPathProvider deleteFileAtStringPath:captureMoviePath];
lastPathWasOne = NO;
} else {
captureMoviePath = [[URLPathProvider getUrlPathProvider]videoOnePathString];
[URLPathProvider deleteFileAtStringPath:captureMoviePath];
lastPathWasOne = YES;
}
captureMovieURL = [[NSURL alloc] initFileURLWithPath:captureMoviePath];
[captureSession addOutput:captureMovieOutput];
[captureMovieOutput startRecordingToOutputFileURL:captureMovieURL recordingDelegate:self];
}
[NSTimer scheduledTimerWithTimeInterval:loopDuration target:self selector:#selector(restartVideoRecording) userInfo:nil repeats:NO];}
Thank you very much
Yes, you can. Store captured samples in NSTemporaryDirectory and use AVMutableComposition to merge assets at the end of recording session.
there is a Sample code please check this also
I have an iOS app that reads text using the OpenEars API. I am using the latest version (1.2.5). I can not figure out how to change the pitch while the words are being read ("on the fly"). I created a slider to control the pitch. A delegate is fired as the slider is changed. In the delegate function, the FliteController target_mean is changed. The intent was to have the pitch change as soon as the target_mean value was changed. My code is as follows:
-(void)sayTheMessage:(NSString *)message {
// if there is nothing there, don't try to say anything
if (message == nil)
return;
[self.oeeo setDelegate:self];
// we are going to say what is in the label...
#try {
// set the pitch, etc...
self.flite.target_mean = pitchValue; // Change the pitch
self.flite.target_stddev = varienceValue; // Change the variance
self.flite.duration_stretch = speedValue; // Change the speed
// finally say it!
[self.flite say:message withVoice:self.slt];
}
#catch (NSException *exception) {
if ([delegate respondsToSelector:#selector(messageError)])
[delegate messageError];
}
#finally {
}
}
-(void)changePitch:(float)pitch {
if ((pitch >= 0) && (pitch <= 2)) {
// save the new pitch internally
pitchValue = pitch;
// change the pitch of the current speaking....
self.flite.target_mean = pitchValue;
}
}
Any ideas?
OpenEars developer here. You can't change the pitch on the fly with FliteController since the pitch is set before speech is processed.
I'd like to crossfade from one track to the next in a Spotify enabled app. Both tracks are Spotify tracks, and since only one data stream at a time can come from Spotify, I suspect I need to buffer (I think I can read ahead 1.5 x playback speed) the last few seconds of the first track, start the stream for track two, fade out one and fade in two using an AudioUnit.
I've reviewed sample apps:
Viva - https://github.com/iKenndac/Viva SimplePlayer with EQ - https://github.com/iKenndac/SimplePlayer-with-EQ and tried to get my mind around the SPCircularBuffer, but I still need help. Could someone point me to another example or help bullet-point a track crossfade game plan?
Update: Thanks to iKenndac, I'm about 95% there. I'll post what I have so far:
in SPPlaybackManager.m: initWithPlaybackSession:(SPSession *)aSession {
added:
self.audioController2 = [[SPCoreAudioController alloc] init];
self.audioController2.delegate = self;
and in
- (void)observeValueForKeyPath:(NSString *)keyPath ofObject:(id)object change:(NSDictionary *)change context:(void *)context {
...
self.audioController.audioOutputEnabled = self.playbackSession.isPlaying;
// for crossfade, add
self.audioController2.audioOutputEnabled = self.playbackSession.isPlaying;
and added a new method based on playTrack
-(void)crossfadeTrack:(SPTrack *)aTrack callback:(SPErrorableOperationCallback)block {
// switch audiocontroller from current to other
if (self.playbackSession.audioDeliveryDelegate == self.audioController)
{
self.playbackSession.audioDeliveryDelegate = self.audioController2;
self.audioController2.delegate = self;
self.audioController.delegate = nil;
}
else
{
self.playbackSession.audioDeliveryDelegate = self.audioController;
self.audioController.delegate = self;
self.audioController2.delegate = nil;
}
if (aTrack.availability != SP_TRACK_AVAILABILITY_AVAILABLE) {
if (block) block([NSError spotifyErrorWithCode:SP_ERROR_TRACK_NOT_PLAYABLE]);
self.currentTrack = nil;
}
self.currentTrack = aTrack;
self.trackPosition = 0.0;
[self.playbackSession playTrack:self.currentTrack callback:^(NSError *error) {
if (!error)
self.playbackSession.playing = YES;
else
self.currentTrack = nil;
if (block) {
block(error);
}
}];
}
this starts a timer for crossfade
crossfadeTimer = [NSTimer scheduledTimerWithTimeInterval: 0.5
target: self
selector: #selector ( crossfadeCountdown)
userInfo: nil
repeats: YES];
And in order to keep the first track playing after its data has loaded in SPCoreAudioController.m I changed target buffer length:
static NSTimeInterval const kTargetBufferLength = 20;
and in SPSession.m : end_of_track(sp_session *session) {
I removed
// sess.playing = NO;
I call preloadTrackForPlayback: about 15 seconds before end of track, then crossfadeTrack: at 10 seconds before.
Then set crossfadeCountdownTime = [how many seconds you want the crossfade]*2;
I fade volume over the crosssfade with:
- (void) crossfadeCountdown
{
[UIAppDelegate.playbackSPManager setVolume:(1- (((float)crossfadeCountdownTime/ (thisCrossfadeSeconds*2.0)) *0.2) )];
crossfadeCountdownTime -= 0.5;
if (crossfadeCountdownTime == 1.0)
{
NSLog(#"Crossfade countdown done");
crossfadeCountdownTime = 0;
[crossfadeTimer invalidate];
crossfadeTimer = nil;
[UIAppDelegate.playbackSPManager setVolume:1.0];
}
}
I'll keep working on it, and update if I can make it better. Thanks again to iKenndac for his always spot-on help!
There isn't a pre-written crossfade example that I'm aware of that uses CocoaLibSpotify. However, a (perhaps not ideal) game plan would be:
Make two separate audio queues. SPCoreAudioController is an encapsulation of an audio queue, so you should just be able to instantiate two of them.
Play music as normal to one queue. When you're approaching the end of the track, call SPSession's preloadTrackForPlayback:callback: method with the next track to get it ready.
When all audio data for the playing track has been delivered, SPSession will fire the audio delegate method sessionDidEndPlayback:. This means that all audio data has been delivered. However, since CocoaLibSpotify buffers the audio from libspotify, there's still some time before audio stops.
At this point, start playing the new track but divert the audio data to the second audio queue. Start ramping down the volume of the first queue while ramping up the volume of the next one. This should give a pleasing crossfade.
A few pointers:
In SPCoreAudioController.m, you'll find the following line, which defines how much audio CocoaLibSpotify buffers, in seconds. If you want a bigger crossfade, you'll need to increase it.
static NSTimeInterval const kTargetBufferLength = 0.5;
Since you get audio data at a maximum of 1.5x actual playback speed, be careful not to do, for example, a 5 second crossfade when the user has just skipped near to the end of the track. You might not have enough audio data available to pull it off.
Take a good look at SPPlaybackManager.m. This class is the interface between CocoaLibSpotify and Core Audio. It's not too complicated, and understanding it will get you a long way. SPCoreAudioController and SPCircularBuffer are pretty much implementation details of getting the audio into Core Audio, and you shouldn't need to understand their implementations to achieve what you want.
Also, make sure you understand the various delegates SPSession has. The audio delivery delegate only has one job - to receive audio data. The playback delegate gets all other playback events - when audio has finished being delivered to the audio delivery delegate, etc. There's nothing stopping one class being both, but in the current implementation, SPPlaybackManager is the playback delegate, which creates an instance of SPCoreAudioController to be the audio delivery delegate. If you modify SPPlaybackManager to have two Core Audio controllers and alternate which one is the audio delivery delegate, you should be golden.