Advantages of Thunderbolt/USB3 audio interfaces

[jcschild]

i am 100% sure that you CAN NOT get below 32 buffer. IF PCIe cards dont have it
(thats what TB goes thru is PCIe) then no.

at 32 buffer you are pretty much at whats called "real system time"
to get any lower you would need a dedicated OS and non interupts
(another words 100% proprietary system) and then it would not be much less latency. besides who needs less than 64 or 32 buffer? and why?

First of all I think it is better to talk in terms of ms rather than buffer size because that is what really matters, how long it takes from pressing enter to the audio to be hearable. In 44100 sampling rate 32 buffer corresponds to 0.73ms of latency.

I'm actually running everything at 32 samples on my Mac Mini and so far have had no problems. In Logic it quotes me a "roundtrip latency" of 14.5 ms though (at 44100) but I'm not sure if that is the same thing? There is no lower setting though.

That doesn't make any sense. I found appale to be about exactly the same as my windows machine. 44.1 and 64 buffer gives me 6.2ms RT and that is regardless of which machine I'm using (same sound cards)

its settings, not real actual RTL.. how are you mesuaring it anyway? centrance?

in other words 14 buffer on Apple is the same as 32 buffer on windows..

[aMUSEd]

i am 100% sure that you CAN NOT get below 32 buffer. IF PCIe cards dont have it
(thats what TB goes thru is PCIe) then no.

at 32 buffer you are pretty much at whats called "real system time"
to get any lower you would need a dedicated OS and non interupts
(another words 100% proprietary system) and then it would not be much less latency. besides who needs less than 64 or 32 buffer? and why?

First of all I think it is better to talk in terms of ms rather than buffer size because that is what really matters, how long it takes from pressing enter to the audio to be hearable. In 44100 sampling rate 32 buffer corresponds to 0.73ms of latency.

I'm actually running everything at 32 samples on my Mac Mini and so far have had no problems. In Logic it quotes me a "roundtrip latency" of 14.5 ms though (at 44100) but I'm not sure if that is the same thing? There is no lower setting though.

Logic is notorious for absurd built in buffering.. do a logic test and change buffering and it does nothing...

14.5 ms RTL is crazy high.. 4ms is more like it..

That's just what Logic says in the Audio panel - it certainly doesn't sound like 14.5 ms (more like ultra low) - I don't know how it is determining that score

[mbncp]

The audio buffer size used by the host is one thing, most drivers/cards will add some internal latency.
Some drivers reports the extra latency correctly, some don't.
Best way to know is to send audio from your host and recording it back, disabling latency compensation in the host.
This way you get the full latency, in and out.

[glokraw]

My god, that is one of the worst looking shirts I've ever seen.

It's his cult uniform. They used to be the 'moonies'

[hibidy]

My god, that is one of the worst looking shirts I've ever seen.

It's his cult uniform. They used to be the 'moonies'

orly?

Do you get a free bowl of soup with the shirt?

[guarsh]

Sound Devices USBPre 2 has a USB 3 port on it.

j

why no USB 3 interfaces?

thats easy USB2 (with good drivers and thats the key) is more than good enough for low latency and large I/O
case in point RME UFX 30 in and 30 out I/O works @ 48 buffer all day long.
need budget Presonus new AudioBox VSL series rock @ 32 buffer.

Thunderbolt: lets be clear on a few things here.
in of itself it CANT offer any lower latency than we already have for external or internal (32 buffer)
it can offer more bandwidth or sharing ability.

Present TB is NOT full speed and what Intel intended to release and will next yr. thats optical based Lightpeak (the real name for TB)
many devs are holding out for the real deal..
lightpeak just adds yet more bandwidth.

TB/LP i think finally brings the ability for a more "lego" based growth

just plug peripherals in to the TB/LP Hub.

about 10-12 yrs ago this was talked about but more with having the optical connects to a "CPU" (Central Processing Unit) whre just processor and ram resided.

also putting me out of a job
fortunately so far this has not come to pass but i can see it coming down the road..

and FYI firewire is NOT dead yet...

Scott
ADK

[justin3am]

Sound Devices USBPre 2 has a USB 3 port on it.

j

But it's not actually a USB3 device, as far as I can tell. I find it odd that they put a USB3 jack on there but they don't even take advantage of the improved power spec of USB3.

[Ashe37]

look at it this way max speed thru a pipe is say 200 MPH

thru a 1" pipe you get 500Gal per hour
thru a 2" pipe double that.
thru an 8" pipe 8x that.

note never is the water moving any faster just a whole lot more of it per second.

Ok this just keeps bugging me...

your math is wrong here

a 2" pipe would get 4x the flow, not 2x- the 1" is doubled in two dimensions, so it is 1 x 2 x 2=4
a 8" pipe would get 16x the flow of the 2"- 1x (8/2) x (8/2)

And your analogy is incorrect anyway, because firewire, USB3 and Thunderbolt all operate at completely different clock speeds. It isn't just that the pipe is bigger, its that the pipe is bigger AND the fluid is moving faster.

[matros]

..now don't get wrong - i am all about smashing digital audio through my handy laptop machine, BUT...

if y'all this bothered by latency - then you should go get yourself a tidy 2nd hand multitrack tape recorder yo. that shit has got no latency, & sounds pretty good as well.

[hibidy]

..now don't get wrong - i am all about smashing digital audio through my handy laptop machine, BUT...

if y'all this bothered by latency - then you should go get yourself a tidy 2nd hand multitrack tape recorder yo. that shit has got no latency, & sounds pretty good as well.

I'd love to know the psychology of people that join up just to bump really old threads and then be snarky too.

[Shabdahbriah]

^^^ it boils down to a compulsion to say: "yo" ^^^

[Jace-BeOS]

The real problem is the cost vs benefit ratio. How much benefit will it provide to use a newer, more rare and licensed tech? Who will benefit enough to "need" it? That's why developers aren't jumping at building thunderbolt devices.

Hell, what computer even comes with it? Just some Macs (not the Mac "Pro"). My Macs are too old for it. I'm all FireWire.

Well, not ALL: my Virus TI works amazingly well as a plugin considering it's sporting USB 1.1! Korg couldn't even get similar functionality working on the M3 with FireWire. Heh, Korg didn't seem to care much to get it to work, the add-on cost too much (hello to that $500 thunderbolt add-on to Apollo), so i never bought it, neither did most M3 users (who watched the complaints on forums from those that did buy it). Developers don't care one bit unless the income is there to justify it. Call it a catch-22. Apple and Intel need to push thunderbolt harder and with less investment requirement or it will fail. I hope it doesn't fail because it's the right way for peripheral interconnection to go. Except they didn't do the fiber optic version straight off (meaning there will be a confusion state of incompatible devices and cables once it does go fiber).

[Burillo]

just a side note - judging by some comments in this thread, some of you seem to not know what is buffer size and how it translates to latency.

32 samples on 44100 samples per second is 32/44100 = 0.72ms. yes that's right, that's under 1ms. since we do both input and output, 0.72 in + 0.72 out = 1.5ms total latency. simple math. so, "14 samples on mac == 32 samples on windows" is bullshit, at least in theory.

the real culprit is, however, the reality. let's say you have 32 samples @ 44.1K sampling rate. you get your 32 samples on input, and then the most interesting bit happens. that buffer now has to travel through the OS (a couple of ms), then get processed and mixed by your DAW (usually 1-2ms, since a lot of plugins aren't zero latency, plus there's likely quite a few of them), then it travels back from the DAW through the OS to the device (another 1-2ms) and then gets to the output buffer (32 samples buffer on output).

so even if you have a really small buffer, you still can't go under 3-4ms. that's just physically impossible.


on topic - i don't see how lightpeak/usb 3.0 will help in that regard. as was said earlier, it doesn't get the water faster - it just gets more water per run.