audiomixer: dsAudio.h gap analysis — missing features and design questions

[Ulrond]

Context

During the audiomixer review, @BatthalaH provided a comprehensive gap analysis between the legacy dsAudio.h API and the current AudioMixer AIDL interface. This discussion captures the full gap table for team review and prioritisation.

Per-input volume and fader control are already tracked in issue #396 and PR #399. This discussion covers the remaining gaps.

---

Full dsAudio.h Gap Table

High Priority

#	dsAudio.h Feature	Missing in AIDL	Notes
1	Per-input mixer volume (dsSetAudioMixerLevels)	IAudioMixerController input attenuation	Tracked in #396 / PR #399
2	Fader control main/assoc balance (dsSetFaderControl)	Property.FADER_CONTROL (range -32 to +32)	Tracked in #396 as FADER_LEVEL (0-100 range). Need to decide: keep 0-100 or adopt dsAudio's -32 to +32 range?
3	DRC Mode Line/RF (dsSetDRCMode)	AQParameter.DRC_MODE or OutputPortProperty.DRC_MODE	Currently AQParameter.DRC (id=17) exists but only covers compression, not Line/RF mode selection
4	AC4 language selection	Property.AC4_PRIMARY_LANGUAGE, AC4_SECONDARY_LANGUAGE	Not present. Needed for AC4 content with multiple language tracks
5	Volume Leveller mode+level (dsSetVolumeLeveller)	AQParameter.VOLUME_LEVELLER_MODE + LEVEL	Currently AQParameter.VOLUME_LEVELLER (id=0) exists but lacks mode/level granularity
6	Per-input attenuation (mix weights)	setInputAttenuation() in IAudioMixerController or Property.INPUT_ATTENUATION[]	Tracked in #396 / PR #399 via setInputVolume() / getInputVolume()

Medium Priority

#	dsAudio.h Feature	Missing in AIDL	Notes
7	Audio Compression 0-10 (dsSetAudioCompression)	AQParameter.COMPRESSION	Not present as a distinct parameter
8	Dolby Volume Mode (dsSetDolbyVolumeMode)	AQParameter.DOLBY_VOLUME_MODE	Not present
9	Intelligent Equalizer Mode 0-6 (dsSetIntelligentEqualizerMode)	AQParameter.INTELLIGENT_EQUALIZER_MODE	AQParameter.INTELLIGENT_EQUALIZER (id=11) exists but lacks mode selection
10	Surround Decoder enable (dsEnableSurroundDecoder)	AQParameter.SURROUND_DECODER_ENABLED	AQParameter.SURROUND_DECODER (id=7) exists but lacks explicit enable/disable
11	Surround Virtualizer mode+boost (dsSetSurroundVirtualizer)	AQParameter.SURROUND_VIRTUALIZER_MODE + BOOST	AQParameter.SURROUND_VIRTUALIZER (id=8) exists but lacks mode/boost granularity
12	MI Steering (dsSetMISteering)	AQParameter.MI_STEERING_ENABLED	AQParameter.MI_STEERING (id=9) exists but lacks explicit enable/disable
13	MS11/MS12 decode capability flags	Capabilities.supportsMS11Decode, supportsMS12Decode	Not present in Capabilities parcelable
14	MS12 Capabilities bitfield (dsGetMS12Capabilities)	OutputPortCapabilities.supportedMS12Features	Not present
15	MS12 Profile list (dsGetMS12AudioProfileList)	Capabilities.supportedAQ_Profiles as String[]	Not present
16	Loudness Equivalence (dsEnableLEConfig)	AQParameter.LOUDNESS_EQUIVALENCE_ENABLED	Not present
17	Audio Gain -2080..+480 (dsSetAudioGain)	OutputPortProperty.GAIN	Not present. Currently only VOLUME (0-100) exists
18	ATMOS output mode enable (dsSetAudioAtmosOutputMode)	OutputPortProperty.ATMOS_OUTPUT_ENABLED (write)	Currently DOLBY_ATMOS_SUPPORT (id=6) is read-only capability only
19	Auto stereo mode (dsSetStereoAuto)	OutputPortProperty.AUTO_MODE	Not present

Low Priority

#	dsAudio.h Feature	Missing in AIDL	Notes
20	ARC/eARC management (SAD, ARC type, enable)	OutputPortProperty.ARC_*	May belong in HDMI HAL rather than AudioMixer

---

Questions for Discussion

1. AQParameter granularity: Many existing AQParameters (VOLUME_LEVELLER, SURROUND_DECODER, SURROUND_VIRTUALIZER, MI_STEERING, INTELLIGENT_EQUALIZER) exist as simple IDs but lack the mode/level/enable sub-parameters that dsAudio.h provides. Should we:

   • Add new AQParameter enum values for each sub-parameter (e.g., VOLUME_LEVELLER_MODE = 18, VOLUME_LEVELLER_LEVEL = 19)?
   • Use structured PropertyValue to encode mode+level in a single parameter?
   • Keep the current single-ID approach and let the vendor interpret the value?

2. DRC Mode: dsAudio.h distinguishes Line mode vs RF mode for DRC. Should this be a new AQParameter or an OutputPortProperty?

3. AC4 language selection: Is this a mixer concern or should it live on the AudioDecoder? AC4 language selection happens at the decode stage, not the mix stage.

4. MS12 capabilities: Should MS12/MS11 support flags be part of Capabilities (mixer-level) or OutputPortCapabilities (per-port)? Or both?

5. Audio Gain vs Volume: dsAudio.h has both dsSetAudioGain (-2080 to +480, presumably 0.1dB steps) and volume (0-100). Do we need both, or does 0-100 volume plus AQ post-gain cover the use cases?

6. ARC/eARC: Should ARC management stay in the AudioMixer HAL or be handled entirely by the HDMI Output HAL?

7. FADER range: Issue Task: audiomixer — add per-input volume/attenuation and fader control #396 uses 0-100. dsAudio.h uses -32 to +32. Which range should the AIDL adopt?

cc @BatthalaH @srinivasgtl @lkennedylamb @dougadler @deekshithdevadas