Stock input-source names (Pro II and Duo, 1.7.6 and 1.7.7)
This page records the evidence for the stock INPUTSOURCE inputId labels, and for whether the firmware has a custom-name field per source. It was written on 2026-11-04 from offline work only. No mixer was contacted.
The main evidence is the pinned-firmware audit in tools/firmware_audit/source_names.py. Public protocol projects and RØDE's guides are used as corroboration.
The audit establishes labels only. It says nothing about whether a source is connected, active, available in the current hardware mode, or usable as an RC2FX processing route.
Result
The "Vendor switch label" column is the literal string in the audited mixer binaries. The audit verifies the 23 labels for IDs 7–29. IDs 0–6 are outside the audit and come from the public enum. StockSource.name is the display string the RC2FX API returns, which shortens some vendor labels.
inputId | Vendor switch label | Public enum (Source) | Current StockSource.name |
|---|---|---|---|
| 0–3 | not audited | Combo1–Combo4 | Mic 1–4 (Duo ID 2: Headset mic) |
| 4–6 | not audited | Combo1_2, Combo2_3, Combo3_4 | Mic 1+2, Mic 2+3, Mic 3+4 |
| 7 | USB 1 Main | Usb1 | USB 1 |
| 8 | USB 1 Chat | Chat | Chat |
| 9 | USB 2 | Usb2 | USB 2 |
| 10 | Bluetooth | Bluetooth | Bluetooth |
| 11 | Sounds | SoundPad | SMART Pads |
| 12 | Game | VirtualGame | Game |
| 13 | Music | VirtualMusic | Music |
| 14 | Virtual A | VirtualA | Virtual A |
| 15 | Virtual B | VirtualB | Virtual B |
| 16–18 | CallMe 1–3 | CallMe1–CallMe3 | CallMe 1–3 |
| 19–20 | Wireless 1, Wireless 2 | — (public enum ends at 18) | Wireless 1–2 |
| 21–29 | RCV Channel 1–9 | — | RCV Channel 1–9 |
IDs 12–15 are Game, Music, Virtual A and Virtual B. The vendor switch, the public enum and RØDE's virtual-device guide all agree on these four. IDs 19 and 20 are established by the firmware audit. The public enum stops at 18.
Firmware audit
tools/firmware_audit/source_names.py works on the official archive offline, using capstone and pyelftools. It stops unless the archive and mixer SHA-256 match the selected profile in config/firmware-profiles.json. It prints PASS only after it has verified that:
- exactly one function references the string
Game\0; - the jump-table dispatch matches
adrp / add / ldrh / adr / add / br, withw0, [x0, w25, uxtw #1]andx0, x1, w0, sxth #2, and the entries are signed 16-bit offsets scaled by 4; - the source field is loaded from
+0xe4(the channel's source enum, which is independent of fader position), indexed assource + 1, and bounded bycmp w25, #0x1e/b.ls; - all 23 literal labels for IDs 7–29 match, including
19 = Wireless 1and20 = Wireless 2.
The audit passes on all four pinned archives:
| Release | Profiles | Function | ldr w28,[x0,#0xe4] | Jump table | Branch base |
|---|---|---|---|---|---|
| 1.7.6 | Pro II, Duo | 0x412f00 | 0x413104 | 0x72b940 | 0x413438 |
| 1.7.7 | Pro II, Duo | 0x415ef0 | 0x416190 | 0x723fa0 | 0x41642c |
Both models share a mixer binary per release: SHA-256 897b9e25… for 1.7.6 and 5151d0eb… for 1.7.7. Each report records the full archive and mixer hashes, the addresses above, and the caseAddress and stringAddress of every label. The reports are at these paths, relative to the repo root:
.deps/stable-1.7.7-check/pro-ii-source-names-1.7.6.json
.deps/stable-1.7.7-check/pro-ii-source-names-1.7.7.json
.deps/stable-1.7.7-check/duo-source-names-1.7.6.json
.deps/stable-1.7.7-check/duo-source-names-1.7.7.jsonReproducing
Install the audit dependencies and run a profile against its official archive:
python -m pip install -r tools/firmware_audit/requirements.txt
python -m tools.firmware_audit.source_names \
--profile rodecaster-pro-ii-1.7.7 \
--archive PATH/1.7.7/os.tar.gz \
--output pro-ii-source-names-1.7.7.jsonRepeat for the other three profiles:
| Profile | Official archive |
|---|---|
rodecaster-pro-ii-1.7.6 | https://update.rode.com/rcp2/1.7.6/os.tar.gz |
rodecaster-duo-1.7.6 | https://update.rode.com/rcp2-duo/1.7.6/os.tar.gz |
rodecaster-pro-ii-1.7.7 | https://update.rode.com/rcp2/1.7.7/os.tar.gz |
rodecaster-duo-1.7.7 | https://update.rode.com/rcp2-duo/1.7.7/os.tar.gz |
The script reads the catalog pins from config/firmware-profiles.json by default. Use --catalog to point at another copy. It creates the output directory and writes the report only after every check passes. Each report states its limits: it covers names only, makes no claim about active modes, and contains no device labels or personal saved-show labels.
Public corroboration
rodecaster-protocol
Yeradon/rodecaster-protocol v0.15.0 (tag v0.15.0, commit 9373ccecb7ec9b03b398b2214515b4b2a0ef3ddb, published 2026-09-11) is an independent reverse-engineering project. Its model-independent Source enum covers 0–18 and labels 12–15 as Game, Music, Virtual A and Virtual B:
- https://docs.rs/rodecaster-protocol/0.15.0/rodecaster_protocol/names/enum.Source.html
src/names/core.rsL202–L282src/names/core.rsL291–L313
The enum shortens some names (USB 1, Chat, Sound Pad) and ends at CallMe3 = 18. It confirms 12–15. For 7, 8, 11 and 19–29 the vendor switch is the source of record.
rodey
seanheiney/rodey commit bc3e70ab3045561ad9d18d39a4b832ff65b7f342 (2026-08-02, firmware 1.7.3) lists 0 = Mic 1, 1 = Mic 2, 7–11 = USB 1, Chat, USB 2, Bluetooth, SMART Pads, and 0xFFFFFFFF for an empty strip. It has no entries for 12–15, so it neither confirms nor contradicts them.
RØDE's Virtual Devices guide
RØDE's guide names the four virtual sources Game, Music, A and B, and describes the names as organisational:
The vendor switch and the public enum call the last two "Virtual A" and "Virtual B". The guide gives no numeric IDs, so it supports the labels but says nothing about the inputId mapping.
StockSource.name and route qualification
StockSource.name is the serialized display name, defined in crates/rc2fx-core/src/inputs.rs. It is set by the saved-show projection in crates/rc2fx-device/src/stock_inputs.rs:
name()maps IDs 0–29 to the API strings in the table above, with Duo source 2 special-cased as "Headset mic".- If an assigned source has no mapped name, the projection uses
Fader {n}, wherenis the fader position. target_idis set separately. It stays empty unlessbinding()finds a qualified configured route. Sources 12–15 have qualified virtual USB bindings. Sources 21–29 have names only and no chains, because their buffers alias other sources.
A source having a name does not make it a processing route and does not mean it is present.
Custom per-source names
The firmware has no per-source custom name field. The public wire format has twelve INPUTSOURCE properties, confirmed from a real Pro II and Duo full sync on firmware 1.7.3 (src/names/input_source.rs L1–L28):
inputId, inputColor, inputType, inputPower, inputMicrophoneType,
inputMicrophoneGain, inputDigitalGain, inputInstrumentGain, inputPhaseFlip,
inputWirelessSN, inputSipCallSlot, inputRcvAudioSourceTypeNone of these is a name, label or title. inputColor is the only per-source display metadata. The crate's naming module also notes that the nodes carry no labels (src/names/mod.rs L13–L15).
The crate's Duo 1.7.4 capture has the same property set: 30 INPUTSOURCE nodes (19 addressable and 11 later ones) each with the same twelve properties, and 10 CHANNEL nodes each with the same 59 properties and no channel name. The capture's sanitizer zeroes state values, inputId included, so it is evidence for the property set only:
The audited 1.7.6 and 1.7.7 switch supplies fixed labels, so RC2FX shows fixed type labels and should not offer custom source names. This conclusion covers the examined releases only. A future firmware, or RØDE's undocumented show-export format, could add a name field, and the current parser would skip it.
Limits
- The audit establishes labels. Live availability, active USB or hardware mode, and audio routing are outside it.
target_idis the field that says whether a source is a qualified route. - The 0–18 vocabulary is shared across models, but its labels are not per-model socket names. Public ordinal 2 is
Combo3, while the Duo's third hardware source is the headset mic, which the API shows as "Headset mic". IDs 3, 5 and 6 are not physical sources on the Duo. - The public enum ends at 18. No public mapping for 19 or 20 was found, so they rest on the firmware audit, and 21–29 (RCV Channel 1–9) are known only from firmware. The 1.7.4 capture contributes the property set and no ID data.
- The audit covers Pro II and Duo 1.7.6 and 1.7.7. RØDE's notes for those releases list compatibility, performance and bug fixes, with no mention of input sources.
References
| What | Where |
|---|---|
| Offline source-name audit | tools/firmware_audit/source_names.py (source checkout) |
| 1.7.6/1.7.7 reports | .deps/stable-1.7.7-check/{pro-ii,duo}-source-names-{1.7.6,1.7.7}.json |
| Catalog pins | config/firmware-profiles.json (source checkout) |
StockSource | crates/rc2fx-core/src/inputs.rs (source checkout) |
| Name map and fallback | crates/rc2fx-device/src/stock_inputs.rs (source checkout) |
| Independent enum, 0–18 | https://github.com/Yeradon/rodecaster-protocol/blob/v0.15.0/src/names/core.rs#L202-L282 |
INPUTSOURCE properties | https://github.com/Yeradon/rodecaster-protocol/blob/v0.15.0/src/names/input_source.rs#L1-L28 |
| Real Duo 1.7.4 capture | https://github.com/Yeradon/rodecaster-protocol/blob/v0.15.0/tests/fixtures/duo/fw-1.7.4/manifest.toml |
| Earlier partial table (1.7.3) | https://github.com/seanheiney/rodey/blob/bc3e70ab3045561ad9d18d39a4b832ff65b7f342/docs/PROTOCOL.md#L194-L203 |
| RØDE virtual devices | https://rode.com/en-us/user-guides/rodecaster-pro-ii/virtual-devices |
| RØDE firmware notes | https://update.rode.com/notes/rcp2/ |