Audio Quality

Kodio provides predefined quality presets optimized for common recording scenarios. Each preset balances sample rate, channels, and bit depth for its intended use case.

Presets overview

Preset	Sample Rate	Channels	Bit Depth	Best For
🎤 `Voice`	16 kHz	Mono	16-bit	Speech, voice memos, transcription
🎵 `Standard`	44.1 kHz	Mono	16-bit	General audio, podcasts (default)
🎧 `High`	48 kHz	Stereo	16-bit	Music, professional content
🎚️ `Lossless`	96 kHz	Stereo	24-bit	Studio recording, archival

Usage

Pass a preset to any recording function:

// Voice memo
val voiceNote = Kodio.record(
    duration = 30.seconds,
    quality = AudioQuality.Voice
)

// Music recording
val music = Kodio.record(
    duration = 3.minutes,
    quality = AudioQuality.High
)

// Studio quality
val master = Kodio.record(
    duration = 5.minutes,
    quality = AudioQuality.Lossless
)

With a recorder:

val recorder = Kodio.recorder(quality = AudioQuality.Standard)

In Compose:

val recorderState = rememberRecorderState(
    quality = AudioQuality.Voice
)

Choosing a preset

🎤 Voice

When to use: Voice memos, speech-to-text input, phone calls, podcasts with speech only.

Trade-off: Smallest files, optimized for speech frequencies (human voice range). Not suitable for music.

🎵 Standard

When to use: General-purpose recording, mixed content, compatibility with most playback systems.

Trade-off: CD-quality sample rate (44.1 kHz), good balance of quality and file size. Default if not specified.

🎧 High

When to use: Music recording, professional podcasts, content for distribution.

Trade-off: Stereo capture, higher sample rate. Larger files than Standard.

🎚️ Lossless

When to use: Studio recording, archival, post-processing workflows.

Trade-off: Maximum quality, largest files (~35 MB/minute). Overkill for most applications.

File sizes

Approximate sizes for uncompressed WAV files:

Preset	Per Minute	5 Minutes	30 Minutes
Voice	~2 MB	~10 MB	~60 MB
Standard	~5 MB	~25 MB	~150 MB
High	~11 MB	~55 MB	~330 MB
Lossless	~35 MB	~175 MB	~1 GB

Platform negotiation

Quality presets are forwarded to the platform's audio system as a request. If the exact format isn't supported, the platform gracefully falls back to the closest supported format:

Platform	Negotiation behavior
Android	Tries the requested format first, falls back to 48 kHz/mono/16-bit if `AudioRecord` rejects it
JVM (JavaSound)	Checks mixer support, falls back to the device's default format
macOS (CoreAudio)	Tries AudioQueue with the requested format, falls back through device default and platform default
iOS (AVAudioEngine)	Converts from hardware format; falls back to 48 kHz/mono/16-bit if the target format isn't representable in AVAudioFormat (e.g., 24-bit int)
Web (AudioWorklet)	Sample rate is a hint to `getUserMedia`; output is always mono 16-bit PCM due to worklet constraints

After recording starts, Recorder.format returns the actual negotiated format, which you can compare against the requested quality.format.

Custom formats

Need something different? Create a custom AudioFormat:

val customFormat = AudioFormat(
    sampleRate = 22050,
    channels = Channels.Mono,
    encoding = SampleEncoding.PcmInt(IntBitDepth.Sixteen)
)

See Custom Formats for details.

Last modified: 04 April 2026