macOS speaker data doesnt flow

[Jerolivine]

I am using the code block to see if i am able to retrieve speaker data. It works fine with windows but the same code doesnt work for macOS.

Is there some arrangement can be done to make it work for macOS such as accessibility or any code changes that i share below?

use cpal::traits::{DeviceTrait, HostTrait, StreamTrait};
use cpal::{Host, Stream, StreamConfig, SupportedStreamConfig};
use std::error::Error;

fn main() -> Result<(), Box<dyn Error>> {
        let host = cpal::default_host();

     let device = host
         .default_output_device()
         .ok_or_else(|| anyhow::anyhow!("no output device available"))?;

     let mut supported_configs_range = device
         .supported_output_configs()
         .map_err(|err| anyhow::anyhow!("error while querying configs: {}", err))?;

     let supported_config: SupportedStreamConfig = supported_configs_range
         .next()
         .ok_or_else(|| anyhow::anyhow!("no supported output config"))?
         .with_max_sample_rate();
    
     let err_fn = move |err| {
         eprintln!("an error occurred on stream: {}", err);
     };

     println!("{}",supported_config.sample_format());

     let stream = match supported_config.sample_format() {
                cpal::SampleFormat::I8
         | cpal::SampleFormat::I16
         | cpal::SampleFormat::I32
         | cpal::SampleFormat::I64
         | cpal::SampleFormat::U8
         | cpal::SampleFormat::U16
         | cpal::SampleFormat::U32
         | cpal::SampleFormat::U64
                 | cpal::SampleFormat::F32 => device.build_input_stream(
                     &supported_config.into(),
                     move |data:&[f32], _: &_| {
                         println!("data flows.")
                     },
                     err_fn,
                     None
                 )?,
     _ => {panic!("sample format");}
     };

    
    // Start the stream
    stream.play()?;
    
    // Keep the application running to keep receiving data
    std::thread::sleep(std::time::Duration::from_secs(10));
    
    Ok(())
}

[MatiasHiltunen]

I had similar issues with MacOS, default devices were not recognized properly. Made some fixes for myself but have not had time yet to look into it too much to make a pr. You can check the changes I made here if that is for any help: https://github.com/MatiasHiltunen/cpal/blob/master/src/host/coreaudio/macos/enumerate.rs

[Jerolivine]

I tried to import your git repository to my cargo.toml, but looks like i still have the same problem @MatiasHiltunen . What is the situation on your side? Could i ask you to check on your machine ?

[MatiasHiltunen]

Can you run the spectrogram example from my repo?

[Jerolivine]

hmm. In the console i see some colors pink to white and somme blacks too. @MatiasHiltunen

[Jerolivine]

by thte way your spectrogram example is taking microphone device. What i am trying to achieve is to get the speaker. @MatiasHiltunen Is there anything i miss to spesify the device shoulud be speaker ?

[MatiasHiltunen]

Oh, right! I'll try to check this when I have access to my mac, currently have only windows device :)

[MatiasHiltunen]

I think the problem is that on windows wasapi allows loopback recording but CoreAudio's loopback is apparently not supported with cpal yet. I think it might be possible to add support for it though, fe. AudioCap has working implementation with it. Perhaps some virtual driver could be useful here?

[sarosh]

This issue was resolved for me moving from 0.15 -> 0.16

[Jerolivine]

@sarosh are you able to retrieve speaker audio with macOS ?

[MatiasHiltunen]

@Jerolivine Just to be clear, in your sample code you try to build a new input stream from output device, would a simple feedback without the actual loopback support work in your case? Like the feedback.rs from examples but instead of input to output you'd feed the output to input. Not sure if that would work but might be worth to try? I have drafted loopback support for coreaudio based on AudioCap's solution but have not had time to complete it yet.