896HD

MOTU is hostile towards Linux

Support Status: 
Reported to work
Manufacturer: 
MOTU

Comments

Through some reading I see that MOTU in fact appears to be anti-Linux or hostile towards it, however I see that a few MOTO devices do in fact work.
Specifically, I see that only a single user reports that they have the 896HD working in a FFADO setup. Does anyone else use a MOTU 896HD with FFADO drivers and JACK? I'd like to confirm what I can before committing to one.

It is true that a few MOTU devices can work through FFADO. However, this is the result of a considerable amount of protocol analysis by myself quite a few years ago. The resulting driver does seem to work reliably for a fair collection of MOTU interfaces. However, because FFADO receives no assistance from MOTU we cannot guarantee that our method of driving the hardware is completely correct. Furthermore, our ability to support the interface into the future is somewhat limited since the only information we have about the 896HD is that which we have deduced ourselves.

The original MOTU driver work was done using the first version of the Traveler, and I still have access to this hardware. Support for all other MOTU devices has come about due to educated guess-work and a lot of trial and error testing by others who happen to own the other interfaces. This has worked only because of the similarities between the different MOTU interfaces.

The FFADO support for the 896HD came about because someone set themselves up to sniff packets to and from their 896HD. They sent me the packet dumps and I was able to get FFADO to work with this interface. This did take considerable time as you would imagine. Debugging hardware that one doesn't have access to is tricky.

It is my understanding that at present at least some of the 896HD functionality works under FFADO. It has been a while since anyone reported using an 896HD under Linux so it's hard to know what the current situation is. While I have no reason to suspect there have been regressions in recent years I have no way of knowing for sure. In the event that something doesn't work the chances of this being fixable are slim because none of the developers have access to the 896HD hardware. Without this, debugging problems is extremely difficult unless the cause is trivially obvious.

As a result of all this, I would NOT recommend that you purchase any MOTU devices for use under Linux. Even though some of their interfaces have been made to work, the motivation for doing this work was mainly to assist people who already had MOTU hardware to give Linux audio a try. Given the continued refusal of MOTU to provide any form of assistance to the FFADO project, we strongly discourage people from purchasing MOTU hardware if their goal at the time is to use it under Linux. It should also be noted that if you encounter trouble with your 896HD under Linux MOTU will not want to know you - you will get the cold shoulder from tech support very quickly once they find out you're using Linux. MOTU have been extremely hostile to Linux users in the past and there's no evidence to suggest that things have changed over the last year or so. In light of this attitude we do not think it worthwhile for Linux users to reward them with fresh purchases. The only language these companies understand is money. If enough people let MOTU know that they missed a sale due to their appalling treatment of Linux users then maybe some day they will start to listen and understand. Don't hold your breath though: it's been nearly 10 years since I first started working with MOTU interfaces and in that time there has been no discernible shift in MOTU's attitude towards Linux.

If you wish to use your audio interfaces under Linux, we would suggest obtaining a unit from a vendor who has actively supported either FFADO or the ALSA project over the years. Such manufacturers include Focusrite, RME, Echo and various others as noted in our device pages (although not that not all interfaces from such manufacturers are necessarily supported at this point in time). If you have a specific set of requirements I would encourage you to post a summary of these in the ffado-user mailing list so people can look them over and provide suggestions on interfaces to consider based on their personal experiences.

I should note that all the above 896HD comments relate to the 896HD. The more recent 896-mk3 is a different unit and as far as I am aware the channel layout has never been confirmed. This means that audio streaming on the 896-mk3 almost certainly won't work without some work. In addition, all "mk3" MOTU units use a completely different mixer/control protocol compared to the earlier interfaces and FFADO does not yet support this. This means that control of the onboard mixer and other settings like phantom power is not yet available through FFADO for all MOTU mk3 interfaces.

To summarise, unfortunately I cannot give any assurances about the functionality of the 896HD under FFADO. To my knowledge it has worked in the past, it may still work, but I can offer no guarantees. Due to the issues noted above I would suggest considering a different interface from a vendor who has supported FFADO instead.

Thank you very much for the detailed response.
I wonder why they dislike Linux so much?

Thank you very much for the detailed response.
I wonder why they dislike Linux so much?

> I wonder why they dislike Linux so much?

We don't know. There seems to be a cultural belief that disclosing programming information for their interfaces will in some way compromise their "intellectual property". This is of course not the case at all, as most other audio companies (not to mention computer hardware companies) have come to realise over the last decade. Unfortunately it has so far proven impossible to have an informed discussion with MOTU about this - and we have reached out to them on multiple occasions over the past decade. Each time we get a scripted response to the effect that they "don't support Linux".

I think at least part of the problem stems from the structure of MOTU. MOTU as a company does not have a great deal to do with the engineering design of the interfaces. While they might generate the external specifications (number of inputs, hardware controls, behavioural description, and son on), the engineering is subcontracted out to an external company. If this is still the case (and I've seen nothing to indicate a change in recent years) it means that there is probably no one at MOTU who really understands what we are asking for when we request programming information. They simply see this erroneously as "we want your intellectual property" and they push back. Regrettably there seems to be no attempt to read extended explanations about how this is not the case; the answer is always the same and they refuse to engage in a conversation.

Likewise, attempts to talk to the subcontractor would be a waste of time. They would have signed NDAs with MOTU and will brick-wall any attempt to discuss technical details as a result.

Another problem is that the only people the general public can access at MOTU are the "customer support" workers. These people work to a script. That script basically says "say 'no' to any request for technical information". They will not read the detailed request and refuse to pass it up the line to others who might be better suited to deal with it.

The practical upshot of this is that it continues to be impossible to get information from MOTU which would permit their firewire audio interfaces being fully supported under Linux. For this reason FFADO continues to recommend that people purchasing interfaces specifically to use them under Linux choose to avoid MOTU and instead look at offerings from other companies who have been supportive of the FFADO project (thereby rewarding those companies who have decided to engage with the Linux community).