Topic: USB 2.0 clarification for UFX II needed

Hello,

Some people consider USB 2.0 on UFX II a step back while the others state it's more than enough for its specs. I understand that the topic might be annoying but could anyone please confirm that my understanding is correct?

I will try to use pure numbers and some assumptions.

Assumptions:

1) We have to transfer data from UFX to PC, say X inputs
Each input gets converted to digital signal so we end up sending X digital channels to PC

2) On the host we record all those inputs to tracks in DAW. At the same time they payback and get mixed with the existent tracks in DAW. I assume DAW/PC sends back just 2 channels back to UFX. Well, may be more than that if we use routing to different speakers/headphones. Let's assume we have Y digital channels that get converted in UFX into Y analog channels.

3) We use 24bit/96KHz per channel, therefore it's 96000 x 4 bytes = 384000 bytes = 375Kb/s per channel. Not 3 bytes as most probably 3 bytes get aligned to 4 bytes internally - this is my assumption).

4) USB 2.0 theoretical bandwidth is 480 Mbit/s = 60Mb/s. Let's divide it by 3 to get some practical limit = 20 Mb/s. So this amount is shared between X input channels and Y output channels since USB is half-duplex.

Since one channel takes 375Kb/s, therefore we can use 54 channels. That seems WAY more than enough for a typical home studio case. Am I correct with my conclusion?

An extreme example:
Inputs for recording:
- bass guitar
- 2 x guitars
- vocal
- 8 inputs for drums
- 10 stereo channels on top of that (not sure how to utilize all of them for a single recording session but just in case)
== 32 input channels

Outputs:
- 1x2 main speakers
- 2x2 headphones
- 8 other speakers (7.1 system just to make things extreme)
==14 output channels

So the total of 46 channels, which is still within the limits. So bandwidth-wise I see no limitations with USB 2.0. Am I correct?

Why then people still complain/resent "RME uses prehistoric USB 2.0 technology in UFX II flagship"?

Or is there another factor that add some sense to their claims? Something I am missing in my considerations?

P.S. Sorry for bringing this perhaps well-known topic.
Thank you!

2 (edited by ramses 2017-10-26 08:18:50)

Re: USB 2.0 clarification for UFX II needed

In terms of BW calculation, this I can only leave up to RME. BTW, AFAIK the whole amount of channels (30in/30out) will instantly be transferred by the driver, no matter how many you use.

There is no limitation, otherwise the UFX would have sucked since over a decade and this is not the case wink
UFX and UFX II have the same amount of channels and for this USB2 is sufficient otherwise it wouldn't have worked for over a decade.

I was using the UFX for ~4 years and also got RayDAT (PCIe) and 2x UFX+, ADI-2 Pro. All RME devices are absolutely reliable and deliver a phatastic performance even with lowest ASIO buffer sizes.

I am also doing latency critical recordings using a guitar plugged to UFX / UFX+ to drive a virtual amp (Kuassa) in some projects and I can use ASIO Buffersizes up to 256 (but that is the "edge" 128 feels more comfortably) without noticing any delay that slows down your playing which is the case when you hear the Amp too late (signal coming from DAW via USB).

I collected the Round trip times that the ASIO driver reports to the DAW in one of my blog articles. From this excel tabular you can see, that USB, Firewire, USB3, Thunderbolt interfaces are ALL absolutely on par with PCIe based products like the RayDAT.

And even when making performance comparison between UFX+ and RayDAT with a 400 tracks Cubase project (consumes 28GB DRAM on my Win7 machine), where each track has 3 inserts (VSTs) and I can use the smallest ASIO Buffersize with UFX+ and RayDAT without noticing any significant / visible difference in terms that "USB requires more CPU resources". This article is also available in english, scroll down, behing the german part.

At the end of this post the relevant links.

> Why then people still complain/resent "RME uses prehistoric USB 2.0 technology in UFX II flagship"?

Could you kindly post a reference, where this has been said ? That would be interesting.
I never saw something like this.

And whats more important, I never experienced any limitation with USB2 of my UFX which I used for a long time.

Here the references:

1. http://www.tonstudio-forum.de/blog/inde … 8-RME-UFX/

In this article about UFX+ you find an excel with the different RTT times.
For the UFX (that I had before the UFX) there are 3 use cases: USB, FW and as Preamp in front of the RayDAT.
RayDAT is a purely digitally working card so for fairness you need to add the time for the D/A conversion part to be compareable with the UFX.

http://www.tonstudio-forum.de/index.php/Attachment/1931-UFX-UFX-RayDAT-Latencies-jpg/

2. http://www.tonstudio-forum.de/blog/inde … cks-de-en/

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub13

Re: USB 2.0 clarification for UFX II needed

I think you have ‘nailed it’ ie USB2 is fine for a UFX II.

4 (edited by ramses 2017-10-26 08:35:07)

Re: USB 2.0 clarification for UFX II needed

And what people also forget, they should be thankful that RME designs the interface to the actual demand and not for "marketing paper" reasons.

USB2 on i.e. the UFX II is a benefit for users:

  • on systems where i.e. the amount of USB3 ports is still a valuable resource (low)

  • on systems where you even do not yet have USB3

  • As you can connect the UFX II to either USB2 or USB3 ports (in USB2 compat mode). So if somebody is having issues with USB on his system (maybe caused by design / BIOS / USB controller / interference with other USB devices), this gives you more options to find a port where the UFX II is the only device behind an USB controller so that audio is not disturbed by the transfers of other devices.

And if you should require to expand your system with an PCIe USB card, because you do not get rid of USB problems on your particular mainboard (which also can happen in rare cases), then you have the choice of either expanding the PC with either USB2 or USB3. I personally would prefer USB3 so that you do not fill a valuable PCIe socket with older USB2 ports. My UFX with USB2 worked i.e. perfectly on an USB3 port which came from the Intel C612 chipset.

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub13

Re: USB 2.0 clarification for UFX II needed

Thank you for you answer, ramses, very, very interesting!

> Could you kindly post a reference, where this has been said ? That would be interesting.
I never saw something like this.

I have seen some posts on gearlutz where people say that "I would never buy USB2 based interface in the Thunderbolt/USB3 epoch", "shame on RME for USB2.0 in UFX" - or something like that, I do not quote them but rather giving the overall impression. If you really really interested to see them I think I shall be able to find that in history and pm you.

I have looked at your table. Well, here is my TC Konnekt 8 (I would like to upgrade to UFX II or UFX+) shows in its TC Near Panel, System Settings:
Sample Rate: 96KHz
ASIO Buffer Size: 512 samples
Latency: 5.3ms

Just out of curiosity I have just set ASIO BUFFER SIZE = 64 samples (the minimum). Latency becomes 0.7ms
But that's apparently the Input latency only.

Now, in Sonar I see the following ASIO latencies:
Input: 5.7ms, 549 samples
Output latency: 7.3ms, 705 samples
Roundtrip: 13.1ms, 1254 samples

However, I have Sonar running about 5 guitar tracks with 8-10 effects in each and recording another guitar track with 8-10 effects as well and I have NEVER noticed even a slightest latency. Never. Ever.

My system is i7-4790K, Asus Z97A, 32Gb RAM + SSDs, TI Firewire controller.

You figures tell me that the roundtrip latency with RME UFX USB2 on 512 samples == 25.374ms, which is twice higher than what I see on Konnekt 8.

Does it mean that UFX with USB2 is THAT slower? That doesn't make sense, does it?

Re: USB 2.0 clarification for UFX II needed

1st of all take same sample rate of 44.1 thanks.

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub13

Re: USB 2.0 clarification for UFX II needed

Right, now it starts making sense.

TC Control Panel
11.6ms, 512 samples

Sonar:
ASIO Reported Latencies
Input: 12.4ms, 549 samples
Output: 16.ms, 705 samples
Roundtrip: 28.4ms, 1254 samples