From: Niklas Aldervall <niklas@aldervall.se>
To: alsa-devel@alsa-project.org
Subject: Roland OCTA-CAPTURE (0582:0120) - White Noise/Static on Playback
Date: Sat, 29 Nov 2025 12:44:31 +0100 [thread overview]
Message-ID: <6b623cb6-3a9b-4fe4-a57c-9a165d261692@aldervall.se> (raw)
[-- Attachment #1: Type: text/plain, Size: 4819 bytes --]
# ALSA Bug Report: Roland OCTA-CAPTURE (0582:0120) - White Noise/Static
on Playback
## Summary
Roland OCTA-CAPTURE USB audio interface produces white noise/static
instead of clean audio on Linux, despite being detected and initialized
by snd_usb_audio driver.
## Device Information
- **Model**: Roland OCTA-CAPTURE (UA-1010)
- **USB ID**: 0582:0120
- **Manufacturer**: Roland Corporation
- **Device Class**: Vendor Specific (0xFF) - NOT USB Audio Class compliant
- **Serial**: ISRL4B236F81
## System Information
- **Kernel**: 6.17.9-2-cachyos
- **Distribution**: Arch Linux (CachyOS)
- **Audio Server**: PipeWire 1.2.7
- **ALSA Driver**: snd_usb_audio (built as module)
## Symptoms
1. Device is correctly detected and recognized by ALSA
2. ALSA shows proper capabilities: S32_LE, 10ch playback, 12ch capture,
44100Hz
3. Audio streams can be sent to device without errors
4. **OUTPUT**: Only white noise/static is produced instead of clean audio
5. Issue occurs with both direct ALSA (speaker-test) and PipeWire
6. Issue persists across all available profiles (multichannel, pro-audio)
## Device Detection (lsusb)
```
Bus 001 Device 008: ID 0582:0120 Roland Corp. OCTA-CAPTURE
bDeviceClass 255 Vendor Specific Class
bDeviceSubClass 0
bDeviceProtocol 255
idVendor 0x0582 Roland Corp.
idProduct 0x0120 OCTA-CAPTURE
```
## ALSA Stream Configuration
```
Roland OCTA-CAPTURE at usb-0000:00:14.0-3, high speed : USB Audio
Playback:
Status: Stop
Interface 0
Altset 1
Format: S32_LE
Channels: 10
Endpoint: 0x05 (5 OUT) (ASYNC)
Rates: 44100
Data packet interval: 125 us
Bits: 0
Sync Endpoint: 0x85 (5 IN)
Sync EP Interface: 1
Sync EP Altset: 1
Implicit Feedback Mode: Yes
Capture:
Status: Stop
Interface 1
Altset 1
Format: S32_LE
Channels: 12
Endpoint: 0x85 (5 IN) (ASYNC)
Rates: 44100
Data packet interval: 125 us
Bits: 0
Sync Endpoint: 0x05 (5 OUT)
Sync EP Interface: 0
Sync EP Altset: 1
Implicit Feedback Mode: Yes
```
## Hardware Configuration
- **Digital Input**: OFF (changed from AUTO - no effect)
- **Sample Rate**: 44.1kHz (hardware configured)
- **Output**: Connected to OUT 1/2 (analog)
- **Routing**: Factory reset performed - no effect
## Workarounds Attempted (All Failed)
1. **PipeWire profiles**: Tried pro-audio, multichannel-duplex
2. **Module parameters** (all individually and in combination):
- `lowlatency=0` - Disable low-latency mode
- `autoclock=0` - Force manual clock
- `implicit_fb=0` - Disable implicit feedback
- `device_setup=0,1` - Different device modes
- `quirk_flags=0x8,0x20` - Various quirk flags
3. **Hardware configuration**:
- DIGITAL = OFF (vs AUTO)
- Factory reset via hardware buttons
4. **Different audio paths**:
- Direct ALSA (hw:1,0)
- PipeWire
- All produce same static
## Test Command Used
```bash
speaker-test -D hw:1,0 -c 10 -t sine -f 440 -r 44100 -F S32_LE -l 3
```
**Result**: White noise/static instead of clean 440Hz tone
## Kernel Support Status
- Device added to implicit feedback skip list in kernel 5.11.9 (March 2021)
- Patch: `IMPLICIT_FB_SKIP_DEV(0x0582, 0x0120)`
- No additional patches found for kernels 6.x
## Related Devices
- Roland Quad-Capture (0582:0118) - Similar architecture, better Linux
support
- Roland U-8 (older device) - Has third-party Linux tools (u8ctl)
## Known Issues from Community
- LinuxMusicians forum: "loud white noise sounds generated along with audio"
- Multiple users report same symptom since 2013
- No documented solution exists
## Hypothesis
The OCTA-CAPTURE uses Roland's proprietary USB protocol (vendor-specific
class 0xFF).
The device likely requires:
1. Proprietary initialization sequence not performed by generic
snd_usb_audio
2. SysEx MIDI configuration for internal DSP/routing/clock
3. Vendor-specific control messages for proper operation
Without Roland's cooperation or reverse engineering of Windows driver,
proper support may be impossible.
## Request
1. Is there additional debugging we can enable to trace USB communication?
2. Could vendor-specific initialization be missing?
3. Would Roland device reverse engineering resources help?
4. Should this device be blacklisted if it cannot work properly?
## Additional Resources
- Device manual:
https://static.roland.com/assets/media/pdf/OCTA-CAPTURE_v1.6_OM.pdf
- Kernel patch (2021):
https://patchwork.kernel.org/project/alsa-devel/patch/CAOsVg8psuOkQRoccDs7AZzUO=4ffOm=XfXoY_G0iTqRXHtqu4A@mail.gmail.com/
- LinuxMusicians thread: https://linuxmusicians.com/viewtopic.php?t=17423
## Contact
Best Regards
Niklas Aldervall
[-- Attachment #2: octa-capture-debug-data.tar.gz --]
[-- Type: application/gzip, Size: 3918 bytes --]
next reply other threads:[~2025-11-29 11:47 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-11-29 11:44 Niklas Aldervall [this message]
2025-11-29 18:07 ` Roland OCTA-CAPTURE (0582:0120) - White Noise/Static on Playback Lucas
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=6b623cb6-3a9b-4fe4-a57c-9a165d261692@aldervall.se \
--to=niklas@aldervall.se \
--cc=alsa-devel@alsa-project.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).