Issue Replies

Look up the manufacturer's audio spec before theorising

Before forming any hypothesis about a device's speaker topology, read what the manufacturer publishes. Compare its physical driver count against the pins --speaker-info reports. With fewer pins than drivers, suspect a hidden speaker pin, as in #53. With equal counts, the topology is fine and the fault is elsewhere. This is a lookup you perform, not an ask you send the reporter.

Do it early: skipping it in #53's triage produced three wrong leads that the spec refuted in one pass.

• Lenovo: resolve the machine type to the global model name first. The name a reporter gives may be regional, such as China's XiaoXin / 小新 line, and PSREF carries only global names, so a search on the name they typed lands in regional retail listings and never reaches PSREF. The machine type is the region-independent key, and --speaker-info prints it as Product:, e.g. 83SG. Search that alone, restricted to psref.lenovo.com. Result titles read <Family>, <Model name>, Model:<MTM>. Verified both ways: 83SG gives IdeaPad Pro 5 14AGP11 for #67, whose DMI says "XiaoXinPro 14GT AGP11", and 21CD gives ThinkPad X1 Yoga Gen 7, the development machine. PSREF's own APIs return 403/404 and its pages are JS shells, so run this search by hand.
• Never identify a Lenovo by its model suffix. The suffix narrows the candidates, and only the machine type picks one. 14AGP11 alone is shared by the IdeaPad Pro 5, the IdeaPad Slim 5 and the Yoga Slim 7. In #67 the Yoga Slim 7 14AGP11 publishes four drivers and an amplifier, where the reporter's IdeaPad Pro 5 14AGP11 has "Stereo speakers, 2W x2".
• Lenovo, next: read the Speakers line of the PSREF static spec PDF, https://psref.lenovo.com/syspool/Sys/PDF/<Family>/<Slug>/<Slug>_Spec.pdf, e.g. .../Yoga/Yoga_7_16IAH7/Yoga_7_16IAH7_Spec.pdf. Use the PDF, not the /Product/… page, which is a JS app and fetches as an empty shell. Judge by whether the line names woofers or tweeters, never by the leading count. The development X1 Yoga reads "Stereo speakers, 2W x2 woofers and 0.8W x2 tweeters": four drivers behind a "Stereo speakers" prefix.
• ASUS: the model's /techspec/ page states it in prose. The #29 Zenbook S14 UX5406 reads "dual front-firing tweeters and dual woofers".
• Other OEMs: find the official spec page. If it doesn't name drivers, record the count as unknown rather than inferring one.

Device-report triage asks

• If the reporter dual-boots, ask for a Windows A/B on the same content, with Dolby processing toggled off and on. It separates device voicing in amp firmware, which the XML cannot reach, from host processing, the surface we translate.

• Check the corpus by SUBSYS before asking for the XML. The collection is keyed by device id and holds no model names, so it can say whether a SUBSYS is in it but not a model. The id is the Codec subsystem: value in --speaker-info: 0x17AA3941 becomes SUBSYS_17AA3941. It is also the first SUBSYS_ token of any filename the reporter quotes. The Controller subsystem: line below it is a different id: the machine's PCI id. It keys SoundWire and Apple filenames instead, reversed: 17AA:2339 becomes SUBSYS_233917AA.

  find "${ATMOS_CORPUS_DIR:-.}" -iname '*SUBSYS_17AA3941*'

  A hit means their attachment would land as a byte-identical duplicate, and the ask buys nothing but goodwill. In #67 it went out twice, asked by model name, for a tuning the corpus already held.

• If the SUBSYS isn't in the corpus, ask them to attach the tuning XML to the issue. A brief online check for the XML is fine and a hunt is not, because availability is OEM-dependent. Look in every audio download the OEM publishes for the model, and extract nested archives.