HOME | BENCHMARKS | FORUM| CONTACT |
 
DAWbench - Reference Benchmarks :.

   

Audio Interface - Low Latency Performance : Part III :.

   

Its hard to believe that its been nearly 2 years since the last report of my ongoing adventures of trying to bring some clarity in regards to often misunderstood and continuing confusion over Low Latency Performance of the available audio interfaces on the market. For those that haven't read Part II as yet, it would be a good idea you do so before continuing on , as it will give you a solid grounding of the work so far.

So whats happened in the last 2 years, well we have made some headway in improving the driver performance available to end users on some interfaces , whether it is ever acknowledged as being a direct result of the work here is another matter. There has been a growing awareness of the importance of the driver performance for those requiring greater efficiency at the lower working latencies , which is all very positive. However its not all wine and roses , there are still new interfaces hitting the market with drivers that range from essentially non functioning, to barely workable, to average to O.K , which is something that is increasingly frustrating , and very disappointing , but not overly surising.

This will only be a short(er) summary, lets move on.

 

audio-int

llp-logo-blk

Firewire / USB Controllers and Driver Performance - Redux:.    

In Part II I went into greater detail of the available OEM controllers and core drivers that many of the manufacturers employ to get the interfaces to market, as they simply do not have the R&D resources to develop their own in house.

This leaves the manufacturers at the mercy ( for want of a better word ) of the 3rd parties coding the core drivers , and the end results can vary a great deal across the driver providers, even for the same base OEM controller.

In most cases the manufacturers do little more than fine tune the integration with a custom mixer applet and their own GUI , but the core driver performance is dictated by the supplied 3rd party code.

 

I focused a fair amount of effort on the TC Applied Dice powered units in the last report, as it is one of the most widely used of the OEM FW controllers, and the drivers definately had some room for improvement. Being that both the controller and core driver is developed and supplied by the same developer ( unlike the USB varaints ) , it would make it at least easier to have a unified focus on what was required to improve the performance.

I was very happy to see with the release of the Version 4 driver that the performance had been improved substantially , to the point that it became one of the best performing FW interface combinations on the market.

 

Numerous manufacturers benefitted directly - Presonus, Focusrite, Makie, just to name 3.

Interestingly the focus now has moved away from FW to USB2 and to a smaller degree Thunderbolt ( with added confusion as some implementations are simply a FW/USB protocol using the TB connector and not native PCIe )

With the greater shift to USB2 , and the solutions being further complicated by not having a unified developer, as I noted earlier, we have taken one step forward and 2 steps backwards in some instances.

         

llp-2
llp1

LLP ( Low Latency Performance ) Rating & Data Base - May 2014 :.

Listed above is the current state of play for all of the interfaces tested so far.

Not too much has changed , PCIe interfaces still deliver the best overall LLP , with some notable mentions for FW interfaces with the TC Applied version 4 driver release, and USB2 performance ranging from exceptional to poor.

 

RME continue to maintain their consistency and exceptional LLP across all of the protocols, this is specifically due to their in house development of the controller and drivers at every stage. As noted earlier, the Dice FW powered interfaces have made a significant and well deserved jump up the rankings.

 

The USB2 interfaces are now taking more of the share of new releases in the market as Firewire becomes less prominent. This is a 2 edged sword IMO as the FW interfaces for the most part measurably better at at LLP now, and most USB2 offerings are not anywhere near reaching equivilent performance. RME being the only exception.

         
Conclusion - 36 Months On :

Looking back over the last 3 years since embarking on this project , there has been some definite positives, one being that raised awareness of the area of efficient low latency performance , and what to look for in regards to what the manufacturers are actually peddling when they use the terminology of "low latency" in their blurbs.

There have also been some negatives in regards to some encounters with certain developers who are absolutely clueless in regards to the whole area of LLP , and continue to deliver products on to the market with little concept of the actual end user requirements.


 

I understand that this can be very sensitive and even confronting especially if the developer is being challenged about a poorly performing driver , but I would think it was in their best interest to remain open and communicative , especially when time and energy is being offered to help and improve the driver.

I also understand that some do not have the direct resources at their disposal , but some of the excuses and dismissals I have encountered have been ridiculous, one notable being that the driver was within Microsoft specification, and as far as they were concerned that is all that is required.


 

With the rise in newer boutique interface manufacturers entering the market, some with claims of superior USB controller/driver integration , which have been more than a little, hmmm, grey , I can see this project still has some life in it yet.

Vin Curigliano
AAVIM Technology
May 28 2014

Part I | II |III

     

© AAVIMT 2005-2014