You are here

Audio Interface - Low Latency Performance Data Base :

For anything relating to music-making on Windows computers, with lots of FAQs. Moderated by Martin Walker.

Re: Audio Interface - Low Latency Performance Data Base :

Postby tex » Tue May 06, 2014 12:07 am

While not wanting to knock the valid remarks here about how the manufacturers could improve their drivers it is evident that some (!many)users expect to work at way below real world latency which "physically" cannot be practical. Any stage instruments, from the players and not (even worse) the audience's perspective, has considerable latency. While I appreciate that for processing this is an actual issue at high levels of production FX processing I feel that for most of us rather more conservative levels of latency are more than adequate especially on the sort of computers most users actually use. They can only do so much. Listeners EARS can only do so much, miraculous as they are. AIR can only do so much.
I have a feeling that, like the level wars, zero latency processed music would actually become as annoying to the ear as over compression over time.
Attractive dynamics involve the timing of them as well as the level being attractive.
Thus I feel there's a point at which latency becomes musically meaningless below somewhere around 6ms and more likely around 8ms which is somewhere around the real latency of an actual live instrument onstage. It may also be the case in all but the most powerful computers too.
This could be behind the "sloppiness" of some drivers' latency figures as the manufacturers take that into account and so save some processing effort.
Just a bit of devil's advocacy to balance the books a bit. ;)
tex
Regular
Posts: 361
Joined: Tue Apr 01, 2003 12:00 am
Success is round the corner. It's also round the bend.

Re: Audio Interface - Low Latency Performance Data Base :

Postby TAFKAT » Tue May 06, 2014 1:06 am

? !

You seem to be missing one vital point , in that the real world latency you speak of is added to the interface playback and RT latency once the sound leaves the DA to either a speaker or headphones. Manufacturers are not taking real world latency into account when being lax with driver efficiency, they simply do not have or want to apply the resources to improve them.

Re whether latency becomes musically meaningless at certain levels , well that's up to the individuals , and I know from experience what a lot of my composer based clients require/demand in that department to be able to feel comfortable with playing in real time.

I have always stated clearly that this project is for those that require ultra low latencies in their working environments using VI's and monitoring with FX/Amp simulators, horse's for course's.

I have no idea where you were going with the zero latency processed music becoming annoying, this is all about maintaining musical feel when composing and playing within the working environments and managing the latencies involved. With ASIO Direct or Hardware Monitoring the latency is substantially lower, so going by your definitions, you would find that an even bigger disadvantage ?? !!

V.C
TAFKAT
Regular
Posts: 244
Joined: Wed Jan 08, 2003 1:00 am
Location: Australia
AAVIM Technology
DAWbench.com

Re: Audio Interface - Low Latency Performance Data Base :

Postby TAFKAT » Wed May 28, 2014 3:12 am

Hey All,

Its been a while since the last update , I haven't had a lot of new interfaces across the bench the last 12 months of so , but I have added a few , summary below.

Image

Image

Added :

Mackie Onyx Blackbird : Posted a summary last September on its own, but will repeat here. Like Focusrite and Presonus , etc, using the TC Applied Dice FW controller and base driver , the unit has been updated to also be using the TC applied V4.x driver that has dramatically improved performance as it did for the previously mentioned.

The Mackie Unit has lower latency converters than the tested Presonus and Focusrite units , resulting in a lower RTL , placing the above those in LLP Ratings, just below the UFX Firewire.

Results were a little down perhaps because of the 4.1 version driver , but overall a very impressive result which has elevated the unit IMO to a very attractive option.

Audient iD22 : I know quite a few are interested in this little box. The unit uses the same OEM controller and base driver as the Presonus VSL Series, driver is solid and scales as expected as you ramp up the buffers. Overall Low Latency Performance is O.K , certainly very usable for most environments, those requiring/demanding higher efficiency at the lower buffer settings may find it not as suitable.

Tested but not included.

Prism Titan : I spent quite a lot of time trying to navigate the idiosyncrasies of this unit , too much time actually. The driver simply refused to install on the X58 chipset reference system that I had successfully installed and tested over 2 dozen units prior , and the iD22 since. Communication with their technical support resulted in an ongoing and increasingly frustrating chasing of the tail without resolution. They concluded I was incapable of installing the driver correctly , despite the fact that it installed on the X79 dev system without issue , and despite close to 30 other interfaces installing and working correctly , on the same core image , it had to be something I was doing wrong as far as they were concerned.

So Unfortunately I couldn't get a comparative result for the Database.

However, I did manage to install and attempt to do a comparative shootout on the newer X79 system against one of the better performing interfaces, the RME UCX. The testing had to be aborted as I could not get any consistency out of the driver once a heavy load was placed on it. What I found was that if you saved a resource heavy session and then attempt to reopen the session at the saved buffer setting, or even 2 -3 buffer settings above that originally working , the driver would collapse , ranging from a garbled slowed down mess, to heavily glitching at buffer settings 3 levels above.

I haven't experienced behaviour that extreme on any interface I can remember testing over the years , to say it was disturbing and totally unacceptable on a unit of that calibre, is an understatement.

I have no idea whether the issue has been addressed, nor do I care. The experience from start to finish is not something I care to ever revisit.
TAFKAT
Regular
Posts: 244
Joined: Wed Jan 08, 2003 1:00 am
Location: Australia
AAVIM Technology
DAWbench.com

Re: Audio Interface - Low Latency Performance Data Base :

Postby TAFKAT » Wed May 28, 2014 7:59 am

Posted the updated charts, report and summary to the main DAWbench website :

http://www.dawbench.com/audio-int-lowlatency3.htm
TAFKAT
Regular
Posts: 244
Joined: Wed Jan 08, 2003 1:00 am
Location: Australia
AAVIM Technology
DAWbench.com

Re: Audio Interface - Low Latency Performance Data Base :

Postby robinv » Wed May 28, 2014 10:06 am

Hey Vin,
I mentioned your work in a video i did recently on latency. Although i'm sure you don't approve of much of it I hope i havent misrepresented you :)

http://youtu.be/ojnnP_GXNaM
User avatar
robinv
Frequent Poster
Posts: 595
Joined: Tue Aug 31, 2004 12:00 am
Molten Music Technology - Computers for doing music on
Making Music on the Microsoft Surface

Re: Audio Interface - Low Latency Performance Data Base :

Postby TAFKAT » Thu May 29, 2014 3:14 am

robinv wrote:Hey Vin,
I mentioned your work in a video i did recently on latency. Although i'm sure you don't approve of much of it I hope i havent misrepresented you :)

Hey Robin,

Thank you for including the project in the video, you didn't misrepresent the work at all , much appreciated.

As always, I am really enjoying what I have watched so far , currently getting intermittent dropouts in playback from this side of the pond, which is pretty amusing considering the subject matter...lol

I am attempting to download it so I can watch it all glitch free... :-)

Well this may come as a surprise to you, but we are very much on the same page for a lot if not most of it, except maybe the amount of latency that end users will tolerate comfortably , I know that if I said to my professional composer clients that 10-20ms is OK, I'd get lynched... LOL !!

It needs to be reiterated that the latency experienced when listening , whether headphones or more to the point, studio monitors at a varying distance, is an addition to that provided by the interface drivers , so the efficiency of the quality of the drivers is paramount.

The 2 key areas of importance are the actual resulting latency with all variables in play, and of course the quality and efficiency of the drivers , which you covered very well in the video.

BTW - I got to the lego explanation and was sitting back LMAO at how good it was , and then the buffering stopped. I have to show that one to the kids who are lego mad.

Peace
TAFKAT
Regular
Posts: 244
Joined: Wed Jan 08, 2003 1:00 am
Location: Australia
AAVIM Technology
DAWbench.com

Re: Audio Interface - Low Latency Performance Data Base :

Postby TAFKAT » Thu May 29, 2014 4:14 am

On a side note, watching the video I was looking at the formatting of the DAWbench website being so skewed and wondered WTF. Checked in my IE and it was fine , of course I was running IE9 , updated to 11 and sure enough, its all skewed.

:roll:

Bloody MS and their tiles, its a conspiracy , I tell ya.

I digress , I'll need to sort that for the IE/Win8 crowd.
TAFKAT
Regular
Posts: 244
Joined: Wed Jan 08, 2003 1:00 am
Location: Australia
AAVIM Technology
DAWbench.com

Re: Audio Interface - Low Latency Performance Data Base :

Postby robinv » Thu May 29, 2014 11:07 am

TAFKAT wrote:

Well this may come as a surprise to you, but we are very much on the same page for a lot if not most of it, except maybe the amount of latency that end users will tolerate comfortably , I know that if I said to my professional composer clients that 10-20ms is OK, I'd get lynched... LOL !!

I appreciate that, most definitely. I polled a bunch of people i know in the industry and it varied from "must have zero latency" to the "under 10" level which told me a lot - including that some pro's have no clue what they're talking about :)
I think i was trying to root it in some kind of real world scenario - so for instance if your pro composer was working with an orchestra then it could take ages for him to hear the trombones at the back. That's not to say that's acceptable in a DAW situation but just that latency exists all around us and we're actually quiet good at dealing with it. Really that's for the benefit of non-tech people who are frightened by a word they don't really understand. Otherwise it has to come down to what "feels" right to the user. But you're right, it all adds up in the end.


It needs to be reiterated that the latency experienced when listening , whether headphones or more to the point, studio monitors at a varying distance, is an addition to that provided by the interface drivers , so the efficiency of the quality of the drivers is paramount.

The 2 key areas of importance are the actual resulting latency with all variables in play, and of course the quality and efficiency of the drivers , which you covered very well in the video.

Yeah, most definitely


BTW - I got to the lego explanation and was sitting back LMAO at how good it was , and then the buffering stopped. I have to show that one to the kids who are lego mad.

Peace

That warms my heart :)
I've been trying to explain that idea in books and articles in all sorts of ways over the years and i think i finally cracked it.

Thanks Vin, i feel very encouraged :)
User avatar
robinv
Frequent Poster
Posts: 595
Joined: Tue Aug 31, 2004 12:00 am
Molten Music Technology - Computers for doing music on
Making Music on the Microsoft Surface

Re: Audio Interface - Low Latency Performance Data Base :

Postby TAFKAT » Thu May 29, 2014 9:41 pm

Hey Robin,

I agree that we naturally compensate for latency in real world , and it is interesting talking to the clients of their requirements across a wide range of respective user environments. I have always been very clear that its horses for courses , and for those that require/demand the lower working latencies , they need to be very diligent in interface selection.

The other edge of the sword is that simply having lower latency available , doesn't mean much if the drivers collapse - perfect example was the Steinberg unit you highlighted. I was scathing in my summary of that unit when it first came out, thankfully Steinberg/Yamaha worked with the OEM base driver/firmware and improved it. Same with TC Applied , who's Dice FW solution is now one of the best on the market , which has benefited a large number of manufacturers and end users. One other thing I noticed was the UR28M driver is now reporting the AD/DA to Cubase , Niice.

All of the above has to be a Win/Win.

You did a great job in helping clear the fog in regards to what the manufacturers bundle under the umbrella of low latency, when in many respects they are referring to direct monitoring. I have typed thousands of words on the subject , and some still don't get it, but your video explanation will allow the penny to drop - a picture is better than a thousand words and all that... ;-)

BTW- I managed to download and watch the whole lot ( I have had some other people I passed the link onto locally also have issue with streaming it ), awesome job , but if I may just bring up one point re measuring the latency ( old school way ). I'll give you an A for effort, but seriously Mate, how many have 2 interfaces , 2 computers, and the energy to configure that method ?

Do yourself a favour and use the RTL Utility by Oblique Audio that I use in all my testing, its accurate to within a few samples , and all that is required is a cable from In to Out.

http://www.oblique-audio.com/free/rtlutility

Now to fix the formatting on the website , I have discovered that IE 11 and the latest Safari have both broken the formatting :?

Peace

V.C
TAFKAT
Regular
Posts: 244
Joined: Wed Jan 08, 2003 1:00 am
Location: Australia
AAVIM Technology
DAWbench.com

Re: Audio Interface - Low Latency Performance Data Base :

Postby robinv » Fri May 30, 2014 9:54 am

TAFKAT wrote:

The other edge of the sword is that simply having lower latency available , doesn't mean much if the drivers collapse - perfect example was the Steinberg unit you highlighted. I was scathing in my summary of that unit when it first came out, thankfully Steinberg/Yamaha worked with the OEM base driver/firmware and improved it.
I think it was the first time i was really hit by the driver performance issue, where what you'd been going on about suddenly became very clear. I was planning to do a review and i just couldn't bring myself to do it because it would have been so poor. I was amazed and confused by other glowing reviews published at the time - although the difference in sound quality between this and my Delta 44 was huge so it was really only the drivers that let it down. It's all good now thank god.


... but if I may just bring up one point re measuring the latency ( old school way ). I'll give you an A for effort, but seriously Mate, how many have 2 interfaces , 2 computers, and the energy to configure that method ?
Ha! It's the only way :lol:



Now to fix the formatting on the website , I have discovered that IE 11 and the latest Safari have both broken the formatting :?

There's always something to fix :roll:
User avatar
robinv
Frequent Poster
Posts: 595
Joined: Tue Aug 31, 2004 12:00 am
Molten Music Technology - Computers for doing music on
Making Music on the Microsoft Surface

Re: Audio Interface - Low Latency Performance Data Base :

Postby TAFKAT » Fri May 30, 2014 10:54 pm

Hey Robin,

LOL, lets agree to disagree on measuring the RTL... :-)

The driver performance issue sits under the radar for many , until they encounter it head on as you did , and then suddenly the penny drops. As you noted in the video, I haven't exactly placed myself on some X-Mas card lists of some manufacturers for highlighting the area. In the end it is their responsibility to deliver workable drivers instead of playing Ossie Ostriche , and hoping no one notices. If the work brings some accountability to them to provide a better driver, then it can only be a positive.

Re the reviews, hmmm, its a mixed bad, a lot read like advertorials , some reviewers are brave enough to dig deeper, Sam and Martin are 2 of the later , but it would be a fine line when the manufacturers ads are paying the bills. With reviews that read like ads, the punters have to be informed enough to dismiss them as exactly that , unfortunately they are the majority. Also the shills let loose at various forums are a source of amusement for me as well, as soon as you ask a question re the actual delivered latency and performance , it all seems to get too hard.

Anyhoo,

I am going to have the opportunity to finally test a Roland Capture unit shortly, as I have a client build and configuration using his unit. I will take note of the "Reduce CPU Load" box, also I do hope thy have sorted the issue with the X79 Chipset , otherwise thats going to be an interesting one.

Peace.
TAFKAT
Regular
Posts: 244
Joined: Wed Jan 08, 2003 1:00 am
Location: Australia
AAVIM Technology
DAWbench.com

Re: Audio Interface - Low Latency Performance Data Base :

Postby Goddard » Tue Jun 03, 2014 8:59 pm

Hey V, note that, in their latest "x-Capture" range drivers, Roland changed the label for the "Reduce CPU load" setting tick box to read "Low Latency mode". ;)

Possibly Roland are playing around with the USB I/F parameters in conjunction with toggling the routines their ASIO driver employs at small buffer settings, in which case perhaps Thesycon's USB spy tool utility might shed some light.
User avatar
Goddard
Frequent Poster
Posts: 999
Joined: Wed Apr 04, 2012 12:00 am

Re: Audio Interface - Low Latency Performance Data Base :

Postby TAFKAT » Tue Jun 17, 2014 7:33 am

A quick one with the Roland Studio Capture results and summary

Image

Roland Studio Capture : This is a series of interfaces that I have wanted to qualify for a while , and I finally managed to get one across the bench. I am not sure whether the drivers are universal across the range - i.e Quad Capture / Octo Capture , but this being the latest I would expect the drivers to be the most up to date.

There has been a fair bit of discussion of not only this particular interface but also the others in the Capture range re Rolands claims of superior latency performance and some favourable reviews from some established music tech journos. In the past there has been some confusion navigating the low latency mode , but its clearly marked now, and was duly engaged.

Figures above are for the Low Latency Mode , so I was very surprised to see the double buffering on playback , which is not overly apparent at the lower latency settings, but become significantly more apparent as the buffers are ramped up.

Playback buffers are the largest I have seen on any of the tested interfaces, which brought about some interesting results across the benchmarks , which are based on playback buffer size mostly.

Benchmark results on the surface look very good at the respective latencies in comparison to the reference and most other interfaces, what is not evident unless highlighted is that the results are achieved in some cases at double the actual playback latency per respective buffer setting. The leveller is the RTL value , but the unit still achieves a higher rating due to the faux 032 setting , IMO.

The double buffering on playback forces you to work at a lower buffer setting when requiring comfortable real time playing of VI's , and those requiring tight RTL for Guitar Amp Sims and monitoring with FX will need to be mindful that anything above the 064 buffer setting ramps the RTL quite substantially.

I know Sam Inglis amended the original review due to him missing the low latency mode, but to be honest even with it engaged , the results are not anything I would even remotely consider good , due to the double buffering.

Has the double buffering been introduced with the new driver , or was it present with the older one as well ?

If they got rid of the double buffering on playback this unit would be right up there, but in its current state, I would place it in the tread carefully basket.

V.C
TAFKAT
Regular
Posts: 244
Joined: Wed Jan 08, 2003 1:00 am
Location: Australia
AAVIM Technology
DAWbench.com

Re: Audio Interface - Low Latency Performance Data Base :

Postby Patrice Brousseau » Wed Jun 18, 2014 3:22 am

TAFKAT wrote:

Has the double buffering been introduced with the new driver , or was it present with the older one as well ?

If they got rid of the double buffering on playback this unit would be right up there, but in its current state, I would place it in the tread carefully basket.

V.C


I think your results in pure latency confirm mine. I had something around 7.5 ms (loopback test in Reaper) in Win7x64. In OSX, it's around 7.1 ms (why, I don't know. Same loopback Reaper test).

The test was developed by a Reaper user and built like an addon. It sends a pulse out back to the inputs and calculate the number of samples it takes.

I use the 1.5.1 driver in both Windows and Mac on my MBP (Octa-Capture) and 1.5.0 on my XP DAW with my Quad-Capture.

However, there's a 48 samples setting available, more useful than the "faux" 32. It give 6.3-6.4 ms if I remember well. Also a 96 samples if you want to stay in the 10 ms RTL range.

Thanks for the enlightenment...

Patrice
Patrice Brousseau
Poster
Posts: 20
Joined: Tue Sep 14, 2004 12:00 am
Location: Montréal, Canada
Patrice Brousseau

Re: Audio Interface - Low Latency Performance Data Base :

Postby Pete Kaine » Wed Jun 18, 2014 10:21 am

TAFKAT wrote:
Has the double buffering been introduced with the new driver , or was it present with the older one as well ?

Nope, it was there last time we looked at it last year.
User avatar
Pete Kaine
Frequent Poster (Level2)
Posts: 3171
Joined: Thu Jul 10, 2003 12:00 am
Location: Manchester
Kit to fuel your G.A.S - https://www.scan.co.uk/shop/pro-audio

PreviousNext