Files

mynameisdeleted d80fd810f5 feat: update tweeter thermal guard parameters to 100mW and adjust RMS threshold in graph configurations

2026-09-04 11:04:26 -04:00

12 KiB

Raw Permalink Blame History

MacBook Pro 15,1 — Warm, Natural Audio DSP for t2linux

based on https://github.com/lemmyg/t2-apple-audio-dsp/tree/master

A custom PipeWire filter-chain DSP graph engineered to deliver warm, natural audio to t2linux on the MacBook Pro 15,1 (2018/2019 Intel T2)—aimed at matching or beating macOS (OS X) audio quality both subjectively and objectively.

Note

Architecture: This is not a kernel driver. The Linux T2 kernel/ALSA stack exposes raw speaker PCM. WirePlumber hides the raw device and splices this graph in front of it (alsa_output.platform-sound.RawSpeakers), executing warm voicing EQ, psychoacoustic sub-bass, 8-band dynamic control, driver crossover, FIR correction, and hard driver protection limiters.

⚡ Key Improvements Over Upstream

Upstream graphs (asahi-audio / t2-apple-audio-dsp) target a measurement-flat response that can sound thin, treble-heavy, and distort at high volumes due to missing driver limiters.

Upstream Limitation	Solution in This Graph	Real-World Result
❄️ Thin / Cold Sound	Equal-energy warm voicing curve (+3 dB/octave tilt)	Rich, full, balanced audio across all genres
💥 Distortion at High Volume	Post-FIR driver limiters (`wlim` @ -2dB, `tlim` @ -1dB)	Crystal clean output at 100% volume with zero amp clipping
🔊 Woofer Over-Excursion	60 Hz high-pass + 8-band multiband compressor	Woofers don't bottom out or rattle on heavy bass beats
🔇 No Deep Sub-Bass	Psychoacoustic sub-bass (`virtualbass` via Bankstown)	Extended perceived low-end without physical cone strain
🛡️ Voice-Coil Overheating	Dual Woofer & Tweeter Virtual Thermal Guards	Continuous white noise automatically attenuated to 500mW / 75mW
🎚️ Fixed / Rigid EQ	Isolated 8-band `user_eq` preference node	Custom tone presets that survive git updates

🎛️ Signal Processing Chain

Audio flows through tone controls, dynamic management, ISO-226 equal loudness tracking, FIR driver correction, and physical driver protection limiters:

flowchart TD
    subgraph Stage1 ["1. Input & Voicing"]
        In["🔊 Audio Input"]:::input --> UserEQ["🎚️ User EQ (8-Band Tone Control)"]:::eq
        UserEQ --> EQ["🎼 Voicing EQ (+3dB/oct Warmth & 60Hz HPF)"]:::eq
    end

    subgraph Stage2 ["2. Dynamics & Headroom Management"]
        EQ --> VB["🔊 Virtual Bass (Bankstown Sub-Harmonics)"]:::dynamics
        VB --> MBComp["📊 Multiband Compressor (8-Band LSP)"]:::dynamics
        MBComp --> LoudComp["👂 Loudness Comp (ISO-226 2k FFT)"]:::dynamics
    end

    subgraph Stage3 ["3. Crossover & Driver FIR Correction"]
        LoudComp --> Copy["🔀 4-Channel Crossover Splitter"]:::input
        
        subgraph Tweeters ["Tweeter Channels"]
            Copy --> ConvLT["🔊 Tweeter L FIR (convLT)"]:::fir
            Copy --> ConvRT["🔊 Tweeter R FIR (convRT)"]:::fir
            ConvLT --> TGuard["🛡️ Tweeter Thermal Guard (75mW / 1.0s Tau)"]:::limiter
            ConvRT --> TGuard
            TGuard --> TLim["🛡️ Tweeter Limiter (-1 dB Ceiling)"]:::limiter
        end
        
        subgraph Woofers ["Woofer Channels"]
            Copy --> ConvLW["🔊 Woofer L FIR (convLW)"]:::fir
            Copy --> ConvRW["🔊 Woofer R FIR (convRW)"]:::fir
            ConvLW --> WLim["🛡️ Woofer Limiter (-2 dB Ceiling)"]:::limiter
            ConvRW --> WLim
            WLim --> WGuard["🛡️ Woofer Thermal Guard (500mW / 2.5s Tau)"]:::limiter
        end
    end

    subgraph Stage4 ["4. Driver Output"]
        TLim --> Out["🔈 RawSpeakers Sink"]:::input
        WGuard --> Out
    end

    classDef input fill:#2d3748,stroke:#4a5568,color:#fff;
    classDef eq fill:#2b6cb0,stroke:#3182ce,color:#fff;
    classDef dynamics fill:#d69e2e,stroke:#d69e2e,color:#000;
    classDef fir fill:#805ad5,stroke:#9f7aea,color:#fff;
    classDef limiter fill:#c53030,stroke:#e53e3e,color:#fff;

🛡️ Dual-Driver Virtual Thermal Guards & Power Caps

Because Linux cannot access the T2 hardware current-sensing ADCs, we implement a Virtual Voice-Coil Thermal Model that models continuous electrical power dissipation (P = V_{\text{rms}}^2 / R_{\text{vc}}) and exponential cooling into the air gap:

C_{\text{th}} \frac{d\Delta T(t)}{dt} = \frac{V_{\text{rms}}^2(t)}{R_{\text{vc}}} - \frac{\Delta T(t)}{R_{\text{th}}}

Thermal Guard Configurations

Woofer Thermal Guard (thermal_guard @ Post-wlim):
- Continuous Power Cap: 1.0\text{ W} (1,000\text{ mW}) per channel (16\% of max 6.25\text{ W} peak power).
- Linear RMS Threshold (al): 0.4000 (-7.96\text{ dBFS} RMS).
- Heating Tau (\tau_{\text{heat}}): 2,500\text{ ms} (2.5 seconds).
- Cooling Tau (\tau_{\text{cool}}): 5,000\text{ ms} (5.0 seconds).
Tweeter Thermal Guard (tweeter_thermal_guard @ Pre-tlim):
- Continuous Power Cap: 100\text{ mW} (0.100\text{ W}) per channel (1.6\% of max 6.25\text{ W} peak power).
- Linear RMS Threshold (al): 0.126491 (-17.96\text{ dBFS} RMS).
- Heating Tau (\tau_{\text{heat}}): 1,000\text{ ms} (1.0 second).
- Cooling Tau (\tau_{\text{cool}}): 2,000\text{ ms} (2.0 seconds, fast decay recovery from Zoom/WebRTC glitches).

Acoustic Result: Dynamic music (Blues, Classical, Rock) has high crest factor (12\text{ dB} - 18\text{ dB}) and operates at 0\text{ dB} gain reduction with full peak dynamics. Continuous high-power signals (full-volume white noise or sine sweeps) are automatically attenuated to harmless power levels after ~2 seconds.

🚀 Quick Start

1. Install Dependencies

Installs required LSP & SWH plugins via your package manager (dnf, pacman, apt, zypper) and builds Bankstown from source:

./install-deps.sh

2. Apply Graph

Preflights FIR paths and plugin URIs, merges user EQ overrides, copies the configuration to WirePlumber, and reloads:

./apply.sh

Tip

Ensure "MacBook Pro 15,1 DSP Speakers" is selected as the default output in your desktop sound settings.

🎚️ Sound Profiles & User EQ

Customize tone settings without touching calibrated internal DSP nodes. user_eq sits at the front of the chain, so even aggressive boosts are safely governed by downstream multiband limiters.

Quick Preset Setup

Create your override file:
```
cp user_eq.example.json user_eq.json
```
Edit user_eq.json with your preferred gain multipliers (g_0 to g_7) and apply:
```
./apply.sh
```

Recommended Tone Presets

Profile	70 Hz (`g_0`)	110 Hz (`g_1`)	220 Hz (`g_2`)	450 Hz (`g_3`)	1 kHz (`g_4`)	2.5 kHz (`g_5`)	6 kHz (`g_6`)	10 kHz (`g_7`)
Reference (Flat)	`1.00`	`1.00`	`1.00`	`1.00`	`1.00`	`1.00`	`1.00`	`1.00`
Rock / Pop	`1.00`	`1.26`	`1.00`	`0.94`	`1.00`	`1.12`	`1.19`	`1.12`
Classical / Acoustic	`1.00`	`1.00`	`1.06`	`1.00`	`1.00`	`1.00`	`1.12`	`1.12`
Electronic / Hip-Hop	`1.26`	`1.19`	`1.00`	`0.94`	`1.00`	`1.00`	`1.06`	`1.00`
Movie (Dialogue Focus)	`0.84`	`0.94`	`1.00`	`1.06`	`1.19`	`1.19`	`1.06`	`1.00`
Movie (Action / Bass)	`1.41`	`1.12`	`1.00`	`1.00`	`1.00`	`1.06`	`1.12`	`1.12`
Late-Night (Low Level)	`0.63`	`0.79`	`1.00`	`1.00`	`1.06`	`1.19`	`1.00`	`0.94`

(Note: Gain values are linear multipliers: 1.0 = 0 dB, 1.41 ≈ +3 dB boost, 0.71 ≈ -3 dB cut)

🛠️ Fine-Tuning Guide

If your specific physical unit requires custom acoustic tuning:

Woofer Ceiling: Edit wlim.control.limit in graph.json (Default: 0 dBFS into thermal_guard).
Thermal Recalibration: Edit al in thermal_guard (al = sqrt(P_watts / 6.25W)).
Sub-Bass Synthesis: Adjust virtualbass.control.amt (Default: 1.0).

⚠️ AT-YOUR-OWN-RISK DISCLAIMER & Replacement Resources

Caution

USE AT YOUR OWN RISK: Loud signals in general, as well as all continuous loud audio (e.g. uncompressed music, white/pink noise, full-volume sine sweeps, or software glitches), carry an inherent physical risk of voice-coil overheating, mechanical over-excursion, or permanent speaker driver damage.

While this DSP engine includes virtual thermal-emulation limiters and dynamic excursion limiters as a best-effort protective measure, this software is provided "AS IS", WITH NO IMPLIED WARRANTY OR GUARANTEE OF ANY KIND. You are solely responsible for protecting your own hardware and managing playback volume levels. The authors and maintainers assume no responsibility or liability for damaged speaker cones, blown voice coils, or hardware failures.

🔧 Speaker Replacement & Repair Resources for Power Users

For expert users who wish to customize thresholds, push thermal limits higher will run a risk of speaker blowout. Dont increase the limits unless you understand and are prepared for a speaker-driver replacement process at your own risk. If you wish to replace speaker drivers or to be able to do so yourself these instructions and part-links may help.

🛠️ iFixit Repair Guide: MacBook Pro 15" Touch Bar 2018 Speaker Replacement - iFixit
🛒 Replacement Speaker Modules: Cheap OEM speaker assemblies are readily available if you ever push limits and need a fresh set of drivers: search for MacBook Pro 15,1 / 15,2 (A1990 2018/2019) Left & Right Speaker Assemblies on iFixit Parts or eBay A1990 Speakers.

📚 Documentation & Repository Links

🔗 Repositories & Mirrors

🐙 GitHub Repository: github.com/mynameisdeleted/mbp15-1-audio-dsp
🏢 Fairfax Media Git Server: git.fairfaxmedia.net/t2linux/mbp15-1-audio-dsp

📖 Guides & Deep Dives

📖 INSTALL.md — Full prerequisites, manual plugin build steps, package manager lookup, and troubleshooting.
🎓 README.advanced.md — Comprehensive electroacoustic design rationale, magnitude-vs-power physics, gain staging equations, issue tracebacks, and complete parameter reference.
📊 mb-compressor-params.md — Detailed parameter guide for the 8-band LSP multiband compressor.

🔄 Reverting to Stock

To return to the stock PipeWire graph provided by t2-apple-audio-dsp:

sudo cp /path/to/t2-apple-audio-dsp/configs/15_1/graph.json \
        /usr/share/t2-linux-audio/15_1/graph.json
systemctl --user restart wireplumber

12 KiB Raw Permalink Blame History