I hope to be able to customize the output device instead of the default system audio. Currently, SCAP uses the fixed cpal default_output_device() for default capture, and I hope to add capturing by specifying a name.

[kingandkongmi]

I hope to be able to customize the output device instead of the default system audio. Currently, SCAP uses the fixed cpal default_output_device() for default capture, and I hope to add capturing by specifying a name.

[tembo]

Looking at the codebase, the audio output capture is handled in crates/scap-cpal/src/lib.rs. The create_capturer function (lines 40-82) currently uses host.default_output_device() without any device selection parameter.

There's already a good pattern to follow in the microphone input code at crates/recording/src/feeds/microphone.rs. The MicrophoneFeed struct provides:

• A list() method that enumerates available input devices using host.input_devices() (line 178-199)
• Device selection by name via SetInput message handler (line 614-767)
• A MicrophonesMap type (IndexMap<String, (Device, SupportedStreamConfig)>) for device lookup

Similarly, crates/cap-test/src/discovery/audio.rs already has a discover_audio_devices() function that lists both input and output devices using host.output_devices(), returning DiscoveredAudioOutput structs that include device names.

For scap-cpal specifically, the implementation would need to:

1. Add an optional device name/ID parameter to create_capturer
2. When provided, iterate host.output_devices() to find the matching device by name
3. Fall back to default_output_device() if no name is specified or match is found

cpal's HostTrait::output_devices() returns an iterator of all output devices, and DeviceTrait::name() gives the device name for matching.