Repository navigation
Device Compatibility
This is the detailed public compatibility reference. Implementation means the
published repository main source; verification means the reported function actually
worked on a particular device/configuration/build. They are different facts.
Updated 20 October 2026. The tested-device lists below record working reports and their sources.
The working lists record devices reported working, with a source where available. If a model is not listed, there is no public report here yet; that does not mean it will not work. Try it and share what happens using Reporting Issues and Logging.
Vendor utilities such as TotalMix, DigiCheck, MixControl, and similar FireWire control apps were built for Apple’s IOFireWireFamily, which is no longer part of macOS. They cannot communicate with devices through ASFireWire. Please do not test or report those apps as ASFW failures.
This does not apply to Storage/SBP-2 devices. Apps that communicate with a device using SCSI use macOS’s SCSI path, which ASFireWire supports; examples include Hasselblad Phocus and VueScan.
DICE devices should retain their reported manufacturer/model names in the app, Core Audio, and hardware reports. “Shared DICE” below describes protocol handling, not a replacement device name. Older “Generic DICE” display-name observations are build-specific history; naming fixes and audio results are recorded separately.
| Host or controller | Implementation / evidence | Limits |
|---|---|---|
| Apple Thunderbolt 3 → TB2 → FireWire 800 | Maintainer-tested reference chain, including M1 MacBook Air; other host results are configuration-specific | Requires actual Thunderbolt, not merely a USB-C connector |
| Intel Macs | Contributor reported playback and analog input with Saffire Pro 40 on an iMac 13,1, OCLP setup, local/ad-hoc 0.5.0 build 14: #193 | Does not validate every Intel Mac, patched OS, or the hosted Apple-silicon install guide |
| Agere/Lucent OHCI PCIe controllers | Current main-service personality matches vendor 0x11c1 with FireWire OHCI class 0x0c0010
|
Matching is not a hardware test; SCSI HBA still matches FW643 0x590111c1 only |
| Other PCIe OHCI vendors | Not automatically matched by those main-service rules | Include controller vendor/device IDs and the connection setup in Reporting Issues |
This table records devices with real-hardware test reports; it is not a prediction or a complete compatibility list. An unlisted device may work, particularly if it uses a supported protocol family, but it needs a real test before it can be listed. Devices are grouped by vendor. DICE is a chip family with multiple ASIC models. The table names the specific ASIC where known; otherwise it simply lists DICE. Linux documents variants including DICE-II, TCD2210 (Mini), and TCD2220 (Junior) in its FireWire sound driver list. Links point to the available test evidence.
| Vendor | Device | ASIC / FireWire protocol | Evidence |
|---|---|---|---|
| Alesis | iO|26 | DICE II JET | Issue #211 |
| Apogee | Duet FireWire | Oxford AV/C | Changelog |
| Avid | Mbox Pro, 3rd generation | DICE | PR #162 |
| Focusrite | Saffire Pro 24 DSP | TCD2210 · DICE Mini | Maintainer testing |
| Focusrite | Saffire Pro 24 | DICE | Issue #159 |
| Focusrite | Saffire Pro 40, TCD2220 | TCD2220 · DICE Junior | Issue #193; contributor reports |
| Mackie | Onyx 1640i, serial 204 | DICE | Gearspace hardware-testing thread |
| M-Audio | FireWire 1814 | BeBoB | PR #143 |
| M-Audio | ProjectMix I/O | BeBoB | Contributor history in issue #99 |
| Midas | Venice F24 | DICE | Community-reported working; source not linked |
| Midas | Venice F32 | DICE | Contributor report by @alicankaralar |
| MOTU | UltraLite-mk3 FireWire | MOTU V3 | Issue #192 |
| PreSonus | StudioLive 16.0.2 | DICE | Contributor report by @klochowicz |
| PreSonus | StudioLive 32.4.2 AI | DICE | Issue #190 |
| PreSonus | FireStudio 2626 | DICE | Issue #189 |
| Prism Sound | Orpheus | BeBoB (BridgeCo) | Gearspace hardware-testing thread |
| Prism Sound | ADA-8XR | BeBoB (BridgeCo) | PR #210 |
| RME | Fireface 400 | RME Fireface protocol | Issue #211 |
| RME | Fireface 800 | RME Fireface protocol | Community report on Discord; PR #196 |
| TerraTec | PHASE 88 Rack | BeBoB | Changelog |
Historical Panasonic result retained from this wiki; the tested ASFW build was not recorded.
ASFireWire can capture DV from MiniDV camcorders to raw .dv files, and drive the tape
transport (play / stop / rewind) over AV/C from the app's DV Capture view.
Note
DV capture and audio receive are mutually exclusive — both use isochronous receive context 0, and starting one while the other runs returns "busy". This is expected behaviour, not a bug.
| Device | Evidence |
|---|---|
| Panasonic MiniDV camcorder | Contributor report by @hoffmabc; handling hardening by @gdemontalivet. DV capture over FireWire 400 on macOS 26 and tape transport control were reported. |
Both SPH=0 streams (consumer camcorders, timestamp in CIP.SYT) and SPH=1 (per-block source packet header) are handled. Reports from other camcorders — especially NTSC models and anything that isn't Panasonic — are wanted.
These applications communicate with the device using SCSI through macOS. This is separate from FireWire vendor control software described above.
| Device | Evidence |
|---|---|
| Nikon Coolscan 9000 (LS-9000) | VueScan · @mhellevang |
| Nikon Coolscan 4000 (LS-4000) | VueScan · @mhellevang; Negpy (nkscan) · COXOS |
| Hasselblad H4D-50MS | Phocus · PR #207 test report. Phocus captured and transferred the original 1:1 image over ASFW’s SBP-2/SCSI path. One capture using the physical shutter hung and required disconnecting power from the camera and FireWire chain. |
| Device / family | Discussion |
|---|---|
| RATOC FR1SX FireWire-to-SCSI converter | #44 |
| FireWire iPod | #34 |
| Tascam family | #81; MOTU implementation does not establish Tascam support |
Use Reporting Issues and attach the relevant exports from Logging. Credit the tester, link the evidence, identify build/revision/configuration, and state which functions were verified. Preserve negative results and earlier successes rather than replacing them with a single “works” label. Unknown builds stay explicitly unknown.