Florian Faber wrote:

Pleaseuseafullname,

...

I have been trying to use the hdspm driver with a recently purchased RME Madiface device. The functionality that I require is to play various files through various channels under program (script) control. I am happy to report that it *almost* works - in that I can use my simple userspace programs to play files. However, there are subtle problems - in that the external MADI A-to-D (an Euphonix 703) does not seem to lock into the sampling rate (something that works under windows). Its very close to achieving full functionality (for my needs) - but just that tiny bit away from being fully useable.

Any ideas?

There is no difference on the MADI bus wether you use Windows, OS X or Linux. This seems to be a specific problem with the Euphonix.

I have no reason to believe the MADI bus is different. However, I suspect the stream are different.

I have two Euphonix boxes and they both behave perfectly under windows - but not under linux. The main problem under linux is that they lose sync - at the slightest electromagnetic interference (such as turning off/on the speakers - on the analogue side - even from a distance of 3 metres). The actual detection of the incorrect sampling rate (or "custom") while disconcerting is not really much of a problem if the files are playing fine.

What are the exact parameters you use on Windows (when the Euphonix worked)?

Windows Parameters:

Buffer Size: 256 Clock Mode: Both "Auto sync" and "Master" Madi Out: Both "56" and "64" WDM Devices: 32 96 kHz: both "48 k" and "96 k" Frame System Clock: 44.1 DDS: not selected

The linux parameters:

cat /proc/asound/card0/hdspm

RME HDSPM MADI (Card #2) Rev.d5 Status2first3bits: 1 IRQ: 16 Registers bus: 0xe3100000 VM: 0xffffc20009640000 --- System --- IRQ Pending: Audio=0, MIDI0=0, MIDI1=0, IRQcount=109 HW pointer: id = 0, rawptr = 0 (0->65472) estimated= 0 (bytes) MIDI FIFO: Out1=0x0, Out2=0x0, In1=0x80, In2=0x80 Register: ctrl1=0x1094c9e, ctrl2=0x0, status1=0xc00000, status2=0xc0050701 --- Settings --- Size (Latency): 8192 samples (2 periods of 32768 bytes) Line out: on , Precise Pointer: off Inputsel = Coaxial, SyncRef = MADI ClearTrackMarker = on, Transmit in 64 Channel Mode, Auto Input on Sample Clock Source: Internal 44.1 kHz System Clock Mode: Master Preferred Sync Reference: MADI Sync System Clock Frequency: 44100 --- Status: Inputs MADI=NoLock, WordClock=NoLock AutoSync: Reference= Input not valid, Freq=0 (MADI = 3, Word = 0) Input: Optical, Mode=56 channels

...

Havent had much luck contacting RME and Winfried Ritsch (who wrote/maintains the code).

I maintain the code at the moment.

Good to hear.

Flo

Dee

Sure ...going over to private email.

Florian Faber wrote:

Dee,

...

...

...

Any ideas?

There is no difference on the MADI bus wether you use Windows, OS X or Linux. This seems to be a specific problem with the Euphonix.

I have no reason to believe the MADI bus is different. However, I suspect the stream are different.

The MADI cards work autonomous, they are only controlled by a few parameters. If the parameters are the same, the behaviour is the same. Thus the settings must differ, and we have to find out where.

...

I have two Euphonix boxes and they both behave perfectly under windows - but not under linux. The main problem under linux is that they lose sync - at the slightest electromagnetic interference (such as turning off/on the speakers - on the analogue side - even from a distance of 3 metres).

I don't believe electromagnetic interference to be a reasonable cause for this. What exactly is your setup - 1x 703 and 1x 713? And both lose sync?

I just had a glance over the Euphonix' user manual and had to learn that they cannot loop through MADI and you have to use the wordclock. So, please, be a bit more specific about the setup - and maybe via private mail, because that's of no relevance to alsa development..

Flo