Why Your AirPods Sound Terrible During Voice Calls (And How We Fixed It)

If you have ever been listening to music on your AirPods and then joined a video call or started dictating, you have probably noticed something frustrating. The moment your microphone activates, your audio quality drops dramatically. Your high fidelity music suddenly sounds like it is coming through an old telephone. This is not a bug in your AirPods or your Mac. It is a fundamental limitation built into the Bluetooth protocol itself.

Understanding the Bluetooth Audio Quality Problem

Bluetooth wireless technology was not originally designed for high quality bidirectional audio. When engineers created the Bluetooth audio profiles, they had to make compromises to work within the bandwidth limitations of wireless transmission.

The result is two distinct audio modes that your AirPods and other Bluetooth headphones switch between automatically.

The first mode is called A2DP, which stands for Advanced Audio Distribution Profile. This is what your AirPods use when you are just listening to music or watching videos. A2DP supports high quality stereo audio with sample rates up to 48kHz and uses advanced codecs like AAC or aptX to deliver near CD quality sound.

The second mode is called SCO, which stands for Synchronous Connection Oriented link. This is the mode your Bluetooth headphones switch to whenever an application needs to use the microphone. SCO was designed for voice calls back when Bluetooth was primarily used for mobile phone headsets. It supports only 8kHz or 16kHz mono audio, which is why everything sounds like a telephone call from the early 2000s.

Why Most Voice Apps Make Your Music Sound Terrible

When you activate voice input in most applications on your Mac, the operating system needs to access your Bluetooth microphone. The moment this happens, your AirPods automatically switch from A2DP mode to SCO mode. This switch happens at the hardware level and affects all audio output, not just the microphone.

So if you were listening to your favorite song while working and decided to dictate a quick note, suddenly your music quality would plummet. The song would continue playing, but through the low quality SCO audio channel. Most people find this so jarring that they simply stop using voice input when they have music playing.

This is the tradeoff that virtually every voice application has accepted as unavoidable. After all, it is a hardware and protocol level limitation. What can software possibly do about it?

How Air Solves the Bluetooth Audio Quality Problem

We refused to accept this limitation. When we were building Air, we knew that many of our users would want to use voice commands while listening to music or podcasts on their AirPods. We could not ask them to choose between audio quality and voice productivity.

Our solution relies on a simple observation: most Macs have excellent built-in microphones. The microphone array in modern MacBook Pro and MacBook Air laptops uses beamforming technology and noise cancellation that rivals many external microphones.

When you activate voice input in Air, we perform a series of checks in real time:

First, we detect whether audio is currently being output through a Bluetooth device. We check the active audio route and identify the specific device that is playing your audio.

Second, we determine whether that Bluetooth device supports high quality A2DP audio. Not all Bluetooth devices are the same, and we want to make sure we only intervene when there is actually a quality tradeoff at stake.

Third, we check whether your Mac has a viable alternative microphone available. In most cases, this is the built-in microphone array, but it could also be an external USB microphone or other wired input.

If all these conditions are met, we automatically route your voice input through the Mac's built-in microphone instead of the Bluetooth microphone. Your AirPods stay in high quality A2DP mode because we never ask them to activate their microphone. Your music keeps playing in full fidelity while you dictate.

The entire detection and switching process takes less than 10 milliseconds. You will never notice it happening. You just hold the voice key, speak your command, and everything works exactly as you would expect.

The Technical Details Behind Automatic Audio Routing

Implementing this feature required deep integration with macOS audio APIs. We use Core Audio to monitor the active audio devices and detect changes in real time. When you hold the voice key, we query the current audio session to understand exactly what is happening with your audio routing.

The challenging part was making the switch seamless. If we simply changed the audio input device mid-stream, you might hear a click or experience a brief interruption. We had to carefully time the switch to happen before audio capture begins, ensuring a completely transparent experience.

We also had to handle edge cases like external Bluetooth microphones that are separate from headphones, USB audio interfaces, and AirPods in one ear while the other is in the charging case. Each scenario required specific handling to ensure the best possible experience.

Why This Matters for Voice Productivity on Mac

This might seem like a small detail, but it represents our philosophy of building voice tools that work with your life rather than against it. We want using voice commands to feel natural and unintrusive.

If you had to stop your music every time you wanted to dictate a note or send a message, voice input would feel like an interruption rather than a convenience. By solving the Bluetooth audio quality problem transparently, we removed one more barrier to voice becoming a natural part of your workflow.

This is the kind of invisible engineering that we obsess over at Air. The best features are often the ones you never notice because they just work exactly as you expect them to.