MTRX STUDIO + Thunderbolt 3

[Tone_Mechanic]

Does anyone know how the new Thunderbolt 3 interface for the MTRX STUDIO will act with Pro Tools and CoreAudio?

Currently there are two "classes" of playback engine,

(A) is HDX and HD Native, which I would say are "direct" connections with Pro Tools and tend to have very stable operation since they take exclusive ownership of the drivers when used.

(B) is all other CoreAudio devices, which work just fine, but sometimes have problems because other applications on the computer can access the same driver simultaneously causing a hiccup or crash.

Will the Thunderbolt 3 interface for the MTXR Studio act like (A) and give Pro Tools exclusive access to it, or will it just be another CoreAudio device that can be accessessed by multiple applications at once?

(And before anyone mentions it, I know you can use HDX and HD Native as CoreAudio devices in Pro Tools, simultaneously sharing them with other applications, but then you lose that exclusive operational stability and it's a (B) type interface regardless of what processing is on the card).

I mainly bring this up because I recently had a system that was using Dante Virtual Soundcard as their main playback engine, and using it to connect to the MTRX STUDIO. There were definite minor performance issues, and occasional audible clicks with this setup. We changed the playback engine to HD Native Thunderbolt, and noticed a marked improvement in stability.

So I'm wondering if the Thunderbolt 3 will also give the same increase in stability as HD Native did?

[Darryl Ramm]

DAD's Thunderbolt products are CoreAudio. I would be very surprised if any sane engineering organization would produce a non-CoreAudio interface over Thunderbolt. And here regardless of whatever Avid wants these boxes are build by DAD and DAD are pretty dammed smart. I expect/really hope Avid just ships the DAD standard drivers, Avid messing with the drivers would be the last thing I'd want.

The only reasons for funky driver stuff would be if needed to support DSP processing in the box, but I just don't see that coming. It would make much more sense for Avid to release a next gen HDX card, and even then I'm just not sure if Avid has resources to do much.

Any issues with Core Audio are less just the idea that others apps can access the device but rather that the base quality of CoreAudio drivers varies greatly between vendors. From utter crap, vendors shipping not much better than chip vendors sample drivers, drivers done by dodgy contractors and third party teams, etc. to extremely well build driver products shipped by companies with sophisticated internal software engineering teams with deep knowledge of macOS and CoreAudio.

I'd sure argue that say RME (what I use) CoreAudio drivers can provide a more reliable, lower latency environment for Pro Tools than HD Native systems. Yes, not worse, better. RME also provides better customer and engineering support than Avid does. There just is no automatic benefit from operating a HD native system vs. a *well done* CoreAudio system. OTOH there can be disadvantage in Avid lock-in, requiring DigiLink interfaces, etc., and now HD Native is discontinued. The writing is on the wall, besides HDX the Pro Tools world is going CoreAudio/ASIO.

Any Core Audio driver should support multiple-clients, not doing so would be a major mistake. If you find multiple clients accessing a device cause a problem... then simply don't do it. Here I expect DAD to do a good job with their Thunderbolt driver, but the proof will be in the eating...

Vendors moving to DriverKit/user space drivers may affect things, may cause some instability, I'm curious where DAD is in that journey. Something all vendors have to deal with, but still CoreAudio is the correct direction. (I'd love to discuss audio driver DriverKit packaging with anybody working on that, little bit of a documentation mess there Apple. :-().

[LDS]

It is a bit of a faux pas to judge Coreaudio interfaces based on a virtual soundcard like DVS. It does quite a good job given it has no hardware of its own and only costs $30!

Having owned HDX, HDN and RME HDSPe cards, I don't really think there is anything inherently more stable with Avid stuff. It function integrates more closely with Pro Tools software, but that doesn't mean greater stability. In fact, for quite a few years HDX absolutely sucked because it would throw constant CPU overload errors in basically empty sessions! It didn't happen when using Coreaudio interfaces - https://www.youtube.com/watch?v=lRlZRDguJvo


As long as the quality of drivers is fitting for a high end interface like the MTRX Studio or MTRX II, it should be smooth sailing.

[LDS]

Originally Posted by Darryl Ramm

I'd sure argue that say RME (what I use) CoreAudio drivers can provide a more reliable, lower latency environment for Pro Tools than HD Native systems. Yes, not worse, better.

If you are comparing HD Native coreaudio performance vs RME coreaudio performance, I would say yes. But based on recent tests in Pro Tools with HD Native and a HDSPe Madi card, the HD Native card has both lower latency and better CPU performance. It isn't a huge different, but the HD Native card definitely isn't the lesser of the two. I would expect DAD Thundercore stuff to be a drastic improvement compared to both.

The biggest issue with Avid stuff is the driver quality has been a yo-yo for so many years. Even now, HDX is having all kinds of issues when working in Hybrid mode. 'Integration' does always mean 'better'.

[Darryl Ramm]

Originally Posted by LDS

If you are comparing HD Native coreaudio performance vs RME coreaudio performance, I would say yes. But based on recent tests in Pro Tools with HD Native and a HDSPe Madi card, the HD Native card has both lower latency and better CPU performance. It isn't a huge different, but the HD Native card definitely isn't the lesser of the two. I would expect DAD Thundercore stuff to be a drastic improvement compared to both.

The biggest issue with Avid stuff is the driver quality has been a yo-yo for so many years. Even now, HDX is having all kinds of issues when working in Hybrid mode. 'Integration' does always mean 'better'.

No I meant direct HDN vs. CoreAudio, but I was not meaning to imply there is a huge difference. I am more just trying to point out that the sometimes stated belief that DigiLink/HDN is much better performing than CoreAudio/ASIO is not correct. I'd love to see your numbers, and what buffer size/sample rate you are measuring (esp. if these are the FX cards). And any measurements you have with AD/DA converters at each end... yes that can be low overhead, but sometimes not.

We are already talking low latencies, I don't necessarily expect ThunderCore to be lower latency, but I'd love to see what DAD deliver. Just another company that really impresses me with their engineering chops.

Those quality issues you mention, and the slow death of DigiLink/HDN would just have me not using it for anything newish. If there is a studio racked up with DigiLink HDN gear and it's working OK then sure, great, fantastic I'd not touch a thing. For new stuff I'd be looking elsewhere.

[uptheoctave]

I have the AX64 with the Thunderbolt card.
It works as a core audio driver should.
It is way better than DVS.

In terms of latency… this is at 48kHz with 32 sample buffer.

[LDS]

Originally Posted by Darryl Ramm

No I meant direct HDN vs. CoreAudio, but I was not meaning to imply there is a huge difference. I am more just trying to point out that the sometimes stated belief that DigiLink/HDN is much better performing than CoreAudio/ASIO is not correct. I'd love to see your numbers, and what buffer size/sample rate you are measuring (esp. if these are the FX cards). And any measurements you have with AD/DA converters at each end... yes that can be low overhead, but sometimes not.

We are already talking low latencies, I don't necessarily expect ThunderCore to be lower latency, but I'd love to see what DAD deliver. Just another company that really impresses me with their engineering chops.

Those quality issues you mention, and the slow death of DigiLink/HDN would just have me not using it for anything newish. If there is a studio racked up with DigiLink HDN gear and it's working OK then sure, great, fantastic I'd not touch a thing. For new stuff I'd be looking elsewhere.

'ere you go. The easiest way I have found to measure it is with Oblique Audio's RTL measurement tool plus some mathematics. The Oblique app only works with coreaudio/ASIO interfaces, so can't measure Avid hardware directly inside Pro Tools but it makes it pretty easy to work backwards. So, I used the HDSPe Madi card as a baseline then calculated it from there.

All at 48kHz with 32 sample buffers.

HDSPe MADI Loopback = 149 samples
HDSPe MADI through DAD DX32R routing matrix = 155 samples
HDSPe MADI -> DX32R -> HDN through a Pro Tools channel -> DX32R -> HDSPe MADI = 236 samples.

That means:
HDSPe = 149 samples
DX32R = 6 samples per trip
HDN = 75 samples

I am using them with Lynx Auroras attached to the DX32R. They have a fixed latency of 25 samples at all sample rates.

[LDS]

Originally Posted by uptheoctave

I have the AX64 with the Thunderbolt card.
It works as a core audio driver should.
It is way better than DVS.

In terms of latency… this is at 48kHz with 32 sample buffer.

Oblique RTL Utility will give you a firm measurement based on a loopback test, rather than relying on the drivers reported latency - https://oblique-audio.com/rtl-utility.php

It would be interesting to see the different in latency between a digital loop through the DADman routing matrix, and a physical patch between DA output and AD input. I don't think the reported latency in Logic is really able to tell what routing you are using within the AX64.

[Tone_Mechanic]

Originally Posted by Darryl Ramm

DAD's Thunderbolt products are CoreAudio. I would be very surprised if any sane engineering organization would produce a non-CoreAudio interface over Thunderbolt. And here regardless of whatever Avid wants these boxes are build by DAD and DAD are pretty dammed smart. I expect/really hope Avid just ships the DAD standard drivers, Avid messing with the drivers would be the last thing I'd want.

The only reasons for funky driver stuff would be if needed to support DSP processing in the box, but I just don't see that coming. It would make much more sense for Avid to release a next gen HDX card, and even then I'm just not sure if Avid has resources to do much.

Any issues with Core Audio are less just the idea that others apps can access the device but rather that the base quality of CoreAudio drivers varies greatly between vendors. From utter crap, vendors shipping not much better than chip vendors sample drivers, drivers done by dodgy contractors and third party teams, etc. to extremely well build driver products shipped by companies with sophisticated internal software engineering teams with deep knowledge of macOS and CoreAudio.

I'd sure argue that say RME (what I use) CoreAudio drivers can provide a more reliable, lower latency environment for Pro Tools than HD Native systems. Yes, not worse, better. RME also provides better customer and engineering support than Avid does. There just is no automatic benefit from operating a HD native system vs. a *well done* CoreAudio system. OTOH there can be disadvantage in Avid lock-in, requiring DigiLink interfaces, etc., and now HD Native is discontinued. The writing is on the wall, besides HDX the Pro Tools world is going CoreAudio/ASIO.

Any Core Audio driver should support multiple-clients, not doing so would be a major mistake. If you find multiple clients accessing a device cause a problem... then simply don't do it. Here I expect DAD to do a good job with their Thunderbolt driver, but the proof will be in the eating...

Vendors moving to DriverKit/user space drivers may affect things, may cause some instability, I'm curious where DAD is in that journey. Something all vendors have to deal with, but still CoreAudio is the correct direction. (I'd love to discuss audio driver DriverKit packaging with anybody working on that, little bit of a documentation mess there Apple. :-().

Yeah, sorry, I wasn't really trying to open a debate about Avid's proprietary drivers eco system vs CoreAudio.

I've used both extensively, and my statement is very anecdotal evidence, not any sort of scientific tests. For the last several years I've been using CoreAudio almost exclusively, using UAD interfaces, Focusrite interface, etc. Recently I got quite heavily into aggregate I/Os and virtual interfaces (for loopback purposes) on an M1 Mac Studio trying to achieve certain configurations. In all, I found that the more interfaces and drivers I mix together in CoreAudio, the more little problems, ticks and issues I encountered. Vs HDX which just always worked as I expected and I never had any issues, but of course was very limited interfacing with other apps.

I'm sure that there are some fantastic developers that make some rock solid interfaces. I think you are very right that Avid won't make their own driver for the Thunderbolt interface and just use the DAD one. Which is fine for me, I've come to rely on being able to assign my computer audio to my main Pro Tools outputs so I can monitor Soundly and other apps in 5.1

So I hope the DAD driver for ThunderCore really is stable. If so, I'm getting one.

[Tone_Mechanic]

Originally Posted by LDS

It is a bit of a faux pas to judge Coreaudio interfaces based on a virtual soundcard like DVS. It does quite a good job given it has no hardware of its own and only costs $30!

Having owned HDX, HDN and RME HDSPe cards, I don't really think there is anything inherently more stable with Avid stuff. It function integrates more closely with Pro Tools software, but that doesn't mean greater stability. In fact, for quite a few years HDX absolutely sucked because it would throw constant CPU overload errors in basically empty sessions! It didn't happen when using Coreaudio interfaces - https://www.youtube.com/watch?v=lRlZRDguJvo


As long as the quality of drivers is fitting for a high end interface like the MTRX Studio or MTRX II, it should be smooth sailing.

My anecdotal experience is that CoreAudio is slightly less stable due to the increase of multiple drivers interacting with multiple apps. My experience isn't just with DVS, but UAD, Focusrite, Metric Halo and other interfaces too. It's not to say that CoreAudio itself as a technology can't be more stable and I certainly hope that the DAD ThunderCore driver is.

I missed that period with the HDX problem. By then I was already using CoreaAudio exclusively.

And yes, DVS is it's own beast, and I use it a much as I can. The potential is phenomenal in it, I just which it was more stable on an M1 Mac. It seems like they still have a few more bugs to work out of the M1 Native version.