Audio Interface Techs / Data Transfer Rates (& what is Gained)

J

Josh Pickeral

New member
USB 2.0 - 480 MB / sec
USB 3.0 - 5 GB / sec
USB 3.1 - 10 GB / sec

Thunderbolt 1 - 10 GB / sec
Thunderbolt 2 - 20 GB / sec
Thunderbolt 3 - 40 GB / sec

FireWire800 - 800 MB / sec




I'm currently using a USB 3.0 AI (Zoom UAC-8) <-- 5 GB/sec Transfer Rate

I guess I wonder about what the difference would be, exactly, between what I'm currently using and, say, "Thunderbolt 3" for example..

Obviously, TB3 has a much higher Data Transfer Rate (40GB/sec).. I wonder, though, would this make my inputted Bass Guitar, for example, sound any better?

Just wondering..

Can anyone provide some insight?



I was thinking, earlier, about the difference between upgrading my Processor versus upgrading the AI tech / Data Transfer Rate (like.. from USB 3.0 to 3.1 or TB 3)

And that lead me to post this question.



Thanks,
-Pick

Josh Pickeral
soundcloud.com/JoshPickeral <-- checkout my tracks! :)

Waldorf, MD
 
Hi,
To be honest the important thing between the modern protocols isn't really the throughput speed; It's the latency.
The slowest you listed there is USB 2 which is still good for, what, 16 simultaneous inputs? More?

Whether you use USB/FireWire/Thunderbolt/Whatever, the sound is going to be the same.

Your audio interface converts analog into digital and that's the critical part of the chain.
People spend big money on converters to get the most accurate representation of the analog audio possible.

Once it's digital it's just a question how much information can get from A to B in what time, but the information is a 1:1 copy. There's no 'quality' with USB/Firewire/TB/etc.

The same is true for your processor or any other component in the computer.
It may impact how much of X you can do or how quickly Y will happen, but this is digital information. There's no degradation or quality involved.

Hope that helps. :)
 
Well written drivers trump raw interface speed. USB2 can handle far more channels than most people need and a company like RME who design their own hardware and driver software can squeeze as many as 128 channels through a USB2 connection.
 
What Steen and James said is true.

You have to make sure you're calculating apples to apples.

USB is 480 megabits/sec or 60 MegaBYTES per second. 24 channels of audio at 96,000/24bit will consume about 55 megabits/sec or about 7 MegaBytes/sec. 128 channels will consume about 5 times that, which is still well under the USB2 limit. That's why James says that well written drivers makes such a difference. Your computer has to regularly poll the USB port to avoid overruns and lost data, and that's the driver's job.

Newer chipsets have improved data handling protocols which is why USB3 and Thunderbolt can improve the handling of the raw data to eliminate errors and dropouts/noise. But in the end, the 1s and 0s are the same regardless of the transmission protocol used. If you're recording a couple of tracks without dropouts, USB2 is perfectly adequate. ALso, if you are thinking about Thunderbolt, you need to make sure that your BIOS and chipset can handle it. Many older ones don't seem to be compatible with some of the Thunderbolt cards. If I was going to upgrade (staying in the PC world) I would make sure the motherboard had TB3 built in. I think AMD had just started making it available in the last couple of months. Intel is putting it on some motherboards. Wait a year and let them get all the bugs worked out and you'll probably be better off.


BTW, Josh, how do you like the Zoom interface? I looked at getting one a year or so back, but ended up with a Tascam 16x08 which is working fine. I do like my R-24 for mobile recording duties, though.
 
It may not affect the transfer rate, but that's not the only thing the processor is doing. Besides pulling data from the USB port, it's off updating the video, checking the keyboard, swapping data to the hard drive, back to the keyboard, back to the USB port, scanning for a virus, back to the USB port.

You get the idea... the faster the processor can do these things, the more often it can get back to check the USB port to look for data to swap into memory. Still, it's not going to make anything sound better or different. If it does, then you are getting corrupt data which is a bad thing with a computer. A couple of corrupt data points and your whole computer dies. It doesn't matter if its the 5th bit of the 12,412th sample from your bass, or a mov bit instruction in the code to read the keyboard. It's gotta be correct.
 
Before the first USB 3.0 interfaces came out I asked the question "why?" several times on two forums and got much the same response all the time.

"USB 3.0 will not give more tracks/lower latency than USB 2.0 because of the way the protocol operates. Firewire and now Thinderbolt work differently and give better performance".

Then RME brought out a USB 3.0/TB interface. Track count was massive and the latency minute AND the USB 3.0 and TB performance were exactly the same!

As has been said, it is the design of the interface and crucuially the drivers that determine perfomance in track count and low latency. Even the computer does not have to be THAT fast, clean and well behaved yes but not Deep Thought.

Dave.
 
You must log in or register to reply here.
Back
Top