We need to record the device audio output from cable or bluetooth headset speakers. For example if user listens to some song when Music app runs in background we want to record this song. Is it possible to implement output recording with AudioUnit? We use PalayAndRecord category for shared audioSession with AVAudioSessionCategoryOptionMixWithOthers option.
No, in general you can not listen to the output from another app. There is inter app audio and audio unit technology that enables this, but the connection must be arranged by both apps. It is not possible to just eavesdrop on whatever the user outputs from the device.
Related
I have an application that needs to record voice from Bluetooth headset. It needs work like Siri or Google Assistant. When I press button in App, it records voice from Bluetooth MIC rather than build in Mic. Is that possible?
My understanding is
1) either the Bluetooth is an Apple certificated External Accessory
2) or the Headset can implement a HFP like protocol to active a HFP link ?
Any hints?
If all you need is recording, it's quite straight if your bluetooth device is paired with iOS. All bluetooth headset with microphone will work well.
When you start to record, I think you have already changed your audio session category to play and record.
Then you can check all your available input route with
-[AVAudioSession availableInputs]
and iterate to find the bluetooth, then set it with
-[AVAudioSession setPreferredInput:error:]
Sometimes it will fail, you need to check the output parameter error to find if there is something wrong.
Is there a way create a virtual audio output device that would make it show up in the Music app's or Spotify's output options? Alternatively, is there a way to intercept the audio stream and then force audio output to something unused (say, open headphone port)?
What I want to do is take the raw audio stream, encode/compress it via a codec, and then send over BLE (not BT Classic). Ideally my "device", or service, would show up in the output options of the music/Spotify apps
I know how to record audio using watch speaker. But is there any api available to record using bluetooth headset? So that recording is possible without presenting the audio input controller
Currently there is no such API available to record using Bluetooth headset. If a call is established from watch which has a headset paired does not route voice output to headset rather it uses its default speaker and mic as output and input
I am working to play some voice messages from iPhone to car bluetooth audio and are testing with Lexus RX350, which supports A2DP and HFP. In our case, A2DP is preferred. However, I can only hear the sound when car is using bluetooth as audio source. When listening to CD or radio, no sound from iPhone is actually played over car audio system.
So, what I want to do is:
Use A2DP when bluetooth is the active audio source
Otherwise use HFP
I use AVAudioPlayer to play audio with AVAudioSessionCategoryPlayback as audio session category. It will always succeed as long as the bluetooth connection is established, no matter whether the car audio is on radio, CD or bluetooth. There is no route change notification when switch between audio sources, therefore I don’t know when we shall switch to HFP. Any suggestions?
Btw, music players like Spotify, Pandora, will pause when switch away from bluetooth audio and resume when switch back, Google Maps will use A2DP when bluetooth audio is active and use HFP if not, so I think there should be a way to detect the change
My project involves a magnetic card reader device that plugs into the phono socket (ie only uses microphone)
Can I get my project to output sound through the inbuilt speaker while simultaneously listening for input from the device?
Research suggests this is not possible:
iPhone audio playback: force through internal speaker?
Force iPhone to output through the speaker, while recording from headphone mic
Audio Session Services: kAudioSessionProperty_OverrideAudioRoute with different routes for input & output
The only way round I can see is actually changing the audio session every time I wish to emit a sound.
Is this really the only option? And is it practical to do this? How long would it take for the audio session to reconfigure itself?