1 (edited by speedbird 2012-09-16 23:32:10)

Topic: Avid latency claim compared to USB and Firewire. Is it true?

http://www.avid.com/static/resources/us/images/products/cw_800x700_PTfamilyLatency.png

This is the claim Digidesign makes. Is it true??

I thought Fireface UFX does a lot better than what Digi claims. Anyone?

PS: the chart is here: http://www.avid.com/US/products/Pro-Tools-HD-Native

Re: Avid latency claim compared to USB and Firewire. Is it true?

Seems like a very misleading chart IMO.  Maybe they have a disclaimer that omits RME devices from the results? ;-)

AFAIK, The UCX @ 64 Samples @ 96K (as the chart mentions) over USB would be 64 In + 64 Out + 32 (safety) + 14 (ADC) + 7 (DAC) = 181 Samples @ 96K.  This breaks down to less than 2ms (# of samples / samplerate = 1.88ms).  Otherwords - on par with the HD Native PCIe system!

Maybe Protools is adding some latency of its own to the mix when using a USB device?  I have no clue...

Regardless - Something smells very wrong with that chart!

cool

MADIface-XT+ARC / 3x HDSP MADI / ADI648
2x SSL Alphalink MADI AX
2x Multiface / 2x Digiface /2x ADI8

Re: Avid latency claim compared to USB and Firewire. Is it true?

yep, that's what I thought. My UFX goes lower than some PT rigs. If I were RME I would put up a fight in case this chart that AVID is posting is misleading. They've done things like this before... BS marketing I call it...

Re: Avid latency claim compared to USB and Firewire. Is it true?

Randyman... wrote:

Seems like a very misleading chart IMO.  Maybe they have a disclaimer that omits RME devices from the results? ;-)

AFAIK, The UCX @ 64 Samples @ 96K (as the chart mentions) over USB would be 64 In + 64 Out + 32 (safety) + 14 (ADC) + 7 (DAC) = 181 Samples @ 96K.  This breaks down to less than 2ms (# of samples / samplerate = 1.88ms).  Otherwords - on par with the HD Native PCIe system!

Maybe Protools is adding some latency of its own to the mix when using a USB device?  I have no clue...

Regardless - Something smells very wrong with that chart!

cool

I'm pretty sure they are doing something strange. In my experience with protools (last 3 versions) I have always found I need higher buffer settings to run the same project in PT vs other DAW programs. Even then PT seems to always use more resource than say Reaper or even Vegas!

From a money point of view it would make sense to lock people into their own proprietary hardware, however I'll stick with RME and other programs for low latency, stability etc!!!


even their own interfaces I've owned (mbox 1, mbox 2) were terrible compared to RME in other programs.

Just my $.02.

Isaac P

5

Re: Avid latency claim compared to USB and Firewire. Is it true?

We had discussed this nonsense internally already. While our marketing wanted to use the opportunity to prove how stupid Avid's marketing operates, I am a bit hesitant. We don't need to react on any kind of shit out there. But just for the record: using 64 samples at 96 kHz is not possible as our Windows driver limits the lowest setting to 96 samples. Even that I would not use but recommend 128 samples. This then results in 365 samples = 3.8 ms roundtrip latency. Under Mac the situation is similar. And our FireWire interfaces have similar values as well. So if Avid thinks their chart shows 'Top performing' interfaces then maybe they used their own for that comparison? :-)

Additionally the lowest number in that chart is not the full record path but in - DSP - out, which when done with our units and TotalMix FX results in lower values again. Of course one can't compare TotalMix FX to a ProTools DSP system (that would be silly). But did I put up that (silly) chart? fryingpan

Regards
Matthias Carstens
RME

Re: Avid latency claim compared to USB and Firewire. Is it true?

It's a shame because a lot of people buy that bs that Digi post on their webpage. Everytime I speak with people they really think that only PT can have low latency. Avid is just making this claim over and over and it sticks.

I think RME would do a lot better if people would actually KNOW that their products are as good if not better in terms of latency than any Avid product. But then I'm not trying to run your company... It's your call smile

Steve

7 (edited by Timur Born 2012-09-17 20:46:27)

Re: Avid latency claim compared to USB and Firewire. Is it true?

October 2010 I published the following measurements (done on a Core2Duo Macbook 2010 Pro). Driver updates may have changed the current numbers slightly, but they are still _all_ lower than what Digi claims. And of course the newer UFX and UCX feature lower latency converters and thus lower numbers (than the old published ones) to begin with.

Roundtrip latencies of various RME interfaces at 44.1 kHz with 64 samples audio buffer (rounded), including AD/DA, safety buffers and safety offset:

FF 400 - MAC: 5,37 ms
FF 400 - WIN: 4,65 ms

FF UC - MAC: 4,60 ms
FF UC - WIN: 4,15 ms

HDSPe - MAC: 5,35 ms
HDSPe - WIN: 3,63 ms

Here are two screenshots showing low latency performance of the Fireface UC running *all* I/Os at once while loading the CPU with a demanding Kore patch (several instances of NI Reaktor Space Drone).

Logic Pro 8 running near 100% CPU load at 32 samples, 44.1 kHz:

http://imageshack.us/a/img223/3255/fotofinishlp832.png

Reaper 2 running near 100% CPU load at 48 samples (Vista 64), 44.1 kHz:

http://imageshack.us/a/img20/9249/fotofinishreaper2.png

Re: Avid latency claim compared to USB and Firewire. Is it true?

I'm not sure about Avid/ Pro Tools claims but from my own findings I notice that running a particular performance through Digital Performer 8 with a M- Audio Profire 2626 resulted in having to use buffer sizes of 1024 or higher to ensure all of the tracks with plug ins would play back properly, even then with processor spikes. the buffer setting 2056, although presenting latency issues allowed the optimal playback performance  (12 stereo tracks with 8 -12 plug ins).

The same set of tracks with same plug ins played through DP 8 using my RME Fireface 800 allowed for using buffer sizes of 256 with no processor overload on my Mac. Like Pro Tools, the Avid/ M- Audio gear appears to be far more reliant on processor use than my RME set up. Why is this? I can only attest to what I tried first hand and I am always dubious as to what companies claim to be their test results and what the end user like me actually can achieve with a minimalist set up. For the above tests they must have omitted the use of an RME interface entirely, otherwise they would easily have gotten results similar to mine. cool