a few questions on firewire interfaces

[therage!]

I dont think this is correct. Are you doing 44.1khz? It MIGHT be possible at 96khz but I don't think so.

at 256 samples, even before any sort of processing whatsoever that is 5.8msec in each direction so the minimum possible should be 11.6 msec.

Please recheck and make sure you are at 44.1khz and that there is no dsp mixing going on providing a shorter path

Yep, positive I'm at 44.1, and no DSP.

Would that be round trip? I'm not too up on this so maybe that is one way numbers. I'll have to look at that again.

What I can't figure out is not hearing any latency when monitoring plugins. That's sweet! It's got to have something to do with Cubase integration I think.

 

[pipelineaudio]

Yep, positive I'm at 44.1, and no DSP.

Would that be round trip? I'm not too up on this so maybe that is one way numbers. I'll have to look at that again.

What I can't figure out is not hearing any latency when monitoring plugins. That's sweet! It's got to have something to do with Cubase integration I think.

Im waiting on some expert help but if i run the numbers in calc, I cant even get these numbers theoretically

256 samples @44.1 khz is 5.8 msec, thats in then add the same for out, with ZERO processing and you get a theoretical 11.6 msec. Even for a purely digital to digital chain with no converter latency those numbers are not possible in the real world, but...

I have a theory that yamaha may be doing the honest thing and adding both the input and output buffers to get that 256 sample figure, which would then make those numbers possible.

Kudos to them if they are, I hope they would ALL measure it this way. I would think the centrance utility would catch that, but its possible they expect the number reported to be one way since thats what everyone normally does (looks better, sell more units)

 

[pipelineaudio]

Jim Roseberry weighs in

"As a point of reference, the best PCI/PCIe units deliver a total round-trip latency of 5ms at 64-sample buffer/44.1k.
That 5ms is the sum total of the following:
1. ASIO input buffer - 1,5ms
2. ASIO output buffer - 1.5ms
3. A/D and D/A converters

As you can see, the A/D D/A converters can account for ~2ms of latency.

If the Yamaha mixer was delivering 6ms total round-trip latency at 256-sample buffer/44.1k, it would be delivering 1.5ms total round-trip latency at 64-sample buffer/44.1k.
The A/D D/A converters will introduce that amount of latency.
Thus, it's simply not possible..."

 

[Carter]

You are starting to make my head numb with all those numbers you could be right but in my setups I dont have to deal with the latency so I guess I thats why I question'ed where you recieved those numbers from.

The Firestudio has that crazy ass control panel (the thing that most people have problems with and then call the piece of kit a piece of shit) that took me a while to set up but after that it has been great.

Thanks for the info.

Psss. Those are the true numbers for both of my setups

 

[Carter]

I dont understand, where is the conflict between what I said?

Maybe presonus has finally, after a year or so, provided much lower latency drivers without the safety buffer. That would be cool. And if they dont drop out like mad, I'd have another look.

The Firestudio is the best laid out audio interface I have ever seen. I was stark raving mad over it when they first announced it and got a unit before they were available for retail. I gave it nine months to work, across many systems and even secured offers for other people to write working drivers for it. I wanted this thing to work BADLY

Now I see what your saying, you had a first generation Firestudio? Well I can tell you that the first few batches were crap and so were the drivers. the first one I bought at guitar demon was the messed up one, instead of taking it back I contacted Presonus and discussed the situation. they knew about the driver problem and the heartbeat sound it made.

They decided to ship me a new one and give me a HP-4 with it with a RA number for the broken one. this unit works great. so when I purchased my second one I did the same thing cause I knew It was from that same bad batch at guitar hell.

Now I have the MSR attached to it with the latest driver and its even better last night I ran 21 tracks with some plugs at 64 samples and I could not crash it, I didnt think it could go that low! so that has to say something about the new drivers.

 

[pipelineaudio]

I had a few, and those I also had them send units to had a few. We were aware of the first batch problems (doesnt mean 100% that the replacements we all got still werent first batch). Which is another issue, Presonus' QC, which I doubt you or anyone else will deny, is quite shady. We gave this an epic going thru, with a huge variety of test systems. If it could have gotten stable, then the 64 msec latency would have been acceptable.

I'm not kidding when I say I secured some driver writers for this thing but was denied. You can hopefully guess the names on that list were none too shabby

But back to the latency. Jim Roseberry just confirmed, theres no way those numbers could be correct and the original ones I posted are what he sees as well. I think some of the signal is going thru the DSP mixer. That happens, it also makes it hard to accurately measure the maudio stuff

Now Im not saying you are wrong, just that its tricky to get it right when theres a DSP mixer in the loop. I still cant get an accurate line6 driver measurment because of this issue even when I turn off the tone direct thing

One thing he did say though, was that Focusrite is now letting you change the size of the "safety" buffer. Let me quote his post

"Those figures aren't even possible with the best PCIe hardware. ;-)

The ASIO input/output buffers alone will account for 2.9ms.
What about the A/D D/A converters... and the infamous hidden safety buffer? No matter how small... they're still present.
Presonus needs to do what Focusrite is doing on their new drivers... which is allowing you to tweak the size of the safety buffer. I believe they're calling it the "Firewire Buffer".

Your figures jibe exactly with what I've measured"

So its worth a try probably of running the focusrite drivers.

The only time we got our Firestudio close to stable was when I ran the Alesis IO26 drivers on it...not sure if any damage could happen though

 

[therage!]

Jim Roseberry weighs in

"As a point of reference, the best PCI/PCIe units deliver a total round-trip latency of 5ms at 64-sample buffer/44.1k.
That 5ms is the sum total of the following:
1. ASIO input buffer - 1,5ms
2. ASIO output buffer - 1.5ms
3. A/D and D/A converters

As you can see, the A/D D/A converters can account for ~2ms of latency.

If the Yamaha mixer was delivering 6ms total round-trip latency at 256-sample buffer/44.1k, it would be delivering 1.5ms total round-trip latency at 64-sample buffer/44.1k.
The A/D D/A converters will introduce that amount of latency.
Thus, it's simply not possible..."

Indeed,

I looked at the N12 driver control panel again and it's 6ms input and 6ms output, or send and receive as they call it.

 

[rhythm]

I haven't slammed the door on anything yet, but when a particular unit, chipset, etc. gets a proportionally high number of negative comments, it makes me pretty hesitant.

Regarding the units I mentioned in the initial post, I imagine that I haven't heard too much because the general consensus is that the mic pres and converters are pretty much equal quality among all of them? And again, if anyone would like to weigh in on the DMP3 mic pre vs. the mic pres in these units I'd appreciate the feedback.

Oh, also.....anything on the bundled recording software would be helpful too!

Thanks,
Greg

 

[pipelineaudio]

There are some minor differences between the mic pres and whether the unit offers DI's and stuff.

None of these sound differences will spell the difference between a dud and a hit though

 

[tateros]

It's also telling that the sole ZZounds and Sweetwater review for the TC Konnekt 8 was basically ranting about the poor drivers. Heck, it's bad enough that even some professional reviewers had problems. More to the point, the near complete lack of professional reviews is usually a good indication that something is very wrong with the product, as reviewers don't like to give negative reviews.

Those reviews are from 2006.

The guy in this minimal review calls the driver issue resolved.

http://www.sweetwater.com/store/detail/Konnekt24D/reviews/#anchor

Here's another 2008 review
http://emusician.com/aux_hardware/emusic_tc_electronic_konnekt_macwin/

Aside from the fact that he had to unplug & replug the "USB" cable, it seems like a pretty good review.

So the drivers issue aside, how much noise does this thing produce when recording through the mic preamps?