Thunderbolt is dead. Long live Thunderbolt!

With SONAR being Windows-based only, I'm not sure how much interest there is here re the following news item I received this morning:
 
http://www.zdnet.com/article/thunderbolt-3-how-usb-cooperation-could-lead-to-100-million-connected-computers-soon/
 
Apparently, Intel has come up with a Thunderbolt 3.0 (40 Gbps v 20 Gbps for 2.0 and 480 Mbps for USB 2.0) chip that uses the USB-C (USB 3.0) connector. They expect to see first use by the end of the year.
 
Why might that be important here?
 
How many of us have USB audio interfaces? Yeah, a LOT of us.
 
How many of us wish we had better (or, heck, any) performance from our audio i/f? Yeah, a LOT of us.
 
Would a chip-based 40 Gbps i/f between audio box and PC be, well, awesome? Would for me! So, this was very good news for me to wake up to this morning.
 
Have a great day!

It's about time for an update though. USB 2 has been standard for a long time now and we need to progress. I could see Thunderbolt using the USB-C connection being huge so long as Intel supports it and takes over total control from Apple so it's not so exclusive to Apple products. Up to this point, Apple has been pretty much in control of the software side of Thunderbolt which is why PC has had such rubbish support on it. I'm hoping that now changes.

Um... 5GB/sec (two full length quality movies) is going to quickly find its chokepoint elsewhere. Throw a couple FX on that data stream and those chokes become obvious quickly. Is like trying to render a full length movie; never seen that in any measure of "seconds."

While a lot of people say USB2 will easily handle bandwidth requirements of most typical/smaller applications, since recently moving to an RME-UFX and using USB2 connection, I seem to have more snap, crackle and pop (not from my rice kripsies) and drop-out  issues, simply playing/mixing my recorded tracks than I ever had with my ($300) 10 year old M-audio Delta 1010LT PCI interface card, to the point where I may go back to using it, and thinking I've made a somewhat costly mistake.  I'm not sure what else to attribute this to other than the USB interface, assuming/hoping the RME-UFX isn't a total P.O.C., (aside from apparently the USB interface), perhaps I can still use the mic preamps, directly to outputs without using USB (anyone want a deal on an RME-UFX?). I'm going to now try using the Firewire interface for a while and see if I still have the same issues. Any advice or comments on similar experiences that may help would be appreciated?
 
Cheers  

I use an RME without issue. The only time I have problems are when I need to raise the buffer due to plugins or IO related latency issues.
 
Raising the buffer always solves things for me.
 
If you have an RME, you should be aware that you also get DigiCheck with it. This is a useful tool for all sorts of things.
http://www.rme-audio.de/en/downloads/driver/digicheck.php
 

JohnEgan
Any advice or comments on similar experiences that may help would be appreciated?

 
The great thing about TB (besides it's huge bandwidth, bi-directional capabilities, daisy-chaining, plug & play, etc...)
is the ability to use various drivers over TB (FW, PCIe, Display-port, etc). In most cases, PCIe audio drivers will get you better performance than FW or USB (lower crackle free RTL's under heavy load, no added buffers, lower system overhead, etc..) on virtually any Mac or PC out of the box.
 
It does not end there, as you can add additonal drivers/features to the TB pipeline as needs arise. For instance, UA added it's "Star Clocking" protocol running over TB to their latest 2nd Gen Apollo's. No need for additional cables or spdif/word clock ports. One single cable daisy chained to each UAD TB device (total of 6 and up to 4 Apollo's) that carry separate DSP, audio & clocking streams. Want to add dual 4k monitors,  no problem, just daisy chain to the TB ports. I see lot's of possibilities for future audio/video products that we have not even begun to fathom yet as TB continues to mature (curently moving to TB3 and USB-C connectors). Future computers will have a universal port and cable system for all external devices. The type of device that you connect (USB, TB, Display Port, etc.) will determine it's speed and capabilities. What could be simpler.............this ease of use has been long over due in the computer industry that has had far too many types of chipsets, ports, adapters and cables to keep up with. One port type....one cable for all! Simplicity! 
 
Cheers,
 
Billy Buck
 

Thanks Jim and others, 
I've tried theses things and more, like disabling anything Im not using on PCI bus, except usually I record at 96K, PC probably does need upgrade now, albeit all worked pretty well with the PCI interface card (if that dates my PC), and its a random annoyance. I don't really doubt RME-UFX, but reassurances are good to hear. I'd had been holding out upgrading to get a PC and interface with TB2, but went with the UFX, and USB2 so shouldn't have had to upgrade PC yet. I guess the newer PC's mentioned will now have both options to use, albeit not of real bandwidth value to smaller scale applications, (or to me now), although a lot of newer audio interfaces are now using TB2 I/O's.
Anyway, thanks for the tips, Ill try a few more things suggested, and see how the Firewire works for me.
Cheers         

Larry Jones
Jim - Good to see your well-informed opinions on this. But aside from a USB2 device's ability to manage bandwidth, do you have an opinion/recommendation on a <$400 device with a low enough round trip latency to permit a guitarist (like me) to record tracks one at a time, monitoring through an amp sim? I'm using an old M-Audio PCI card that is probably going to be obsolete under Windows 10. I have no first hand experience with USB interfaces, so have no idea if any of them are "fast" enough. I know I could monitor the interface directly, but I need the amp sim. Thoughts?

Presonus AudioBox VSL series yields 4.9ms total round-trip latency at a 64-sample ASIO buffer size (44.1k).
If you can up the budget to ~$650, the new MOTU Ultralite AVB yields 4.9ms total round-trip latency at those same settings... and has lower noise-floor(~-110dB).

Larry Jones

Jim Roseberry
Presonus AudioBox VSL series yields 4.9ms total round-trip latency at a 64-sample ASIO buffer size (44.1k).
If you can up the budget to ~$650, the new MOTU Ultralite AVB yields 4.9ms total round-trip latency at those same settings... and has lower noise-floor(~-110dB).

Thanks, Jim! I'll take a hard look at both of those.

Avoid the PreSonus AudioBox VSL interfaces if you have an iLok. The VSL drivers will turn the iLok into an unrecognized, malfunctioning USB device. It took me 9 months to connect those 2 dots.
 
That said, the 44VSL I bought is a good piece of hardware. When I connected it as a class-compliant device to my Android tablet, Audio Evolution Mobile found and used it immediately. Easy peasy.
 
If you want to try a VSL i/f, try to acquire it from a place that will let you exchange/return it if you can't get the driver to play nice with your system.

TheMaartian
 
Avoid the PreSonus AudioBox VSL interfaces if you have an iLok. The VSL drivers will turn the iLok into an unrecognized, malfunctioning USB device. It took me 9 months to connect those 2 dots.
 
That said, the 44VSL I bought is a good piece of hardware. When I connected it as a class-compliant device to my Android tablet, Audio Evolution Mobile found and used it immediately. Easy peasy.
 
If you want to try a VSL i/f, try to acquire it from a place that will let you exchange/return it if you can't get the driver to play nice with your system.

I don't have any kind of dongle, so I might be OK with the 44VSL. I've never bought from Sweetwater, but it looks like they have a good return/exchange policy.
 
My real issue all along has been latency: Is the lag short enough that I'll be able to record a direct guitar while monitoring back through an amp sim in Sonar without committing the sim to my track? If I can't do that I will have to change the way I've been working for the past 15 years or so.

mettelus

Larry Jones
 
If I can't do that I will have to change the way I've been working for the past 15 years or so.

I cannot resist temptation here... how did you adapt to the internet and cell phones?

The internet and cell phones made things faster. USB audio interfaces make things slower. Of course I will adapt if I have to, but so far all the options I can find for working with the newer inferior technology seem like poor-choice kludges to me.

Maartian - Sounds like you've got a way to do what (I think) I need to do. Without having any of the gear you describe, I can't quite get my head around how it's done, but I'll figure it out if I need to.
 
I've noticed a lot of people seem not to understand why I feel I need an amp sim. It's because a.) I don't want to mic a real amp and make that much noise in my house and b.) I need at least some overdrive to get the feel I need to play things the way I want to play them. Ergo, amp sim. Doesn't seem that crazy to me.
 
I don't have any experience with external (USB) interfaces, so I was unaware that latency might prevent me from monitoring through the DAW with FX until the past week. I'm like a guy who was given a Ferrari for his 16th birthday and didn't know that not all cars can go 160 mph.

Karyn
Sonar takes the data being streamed from your interface and dumps it directly to disc. It does not process it in any way. The wav in a clip that you just recorded is always raw data.
 
Yes, you can destructively process it later, but not during record. What you're hearing during record is a duplicate of the recorded data being fed into the channel, to which you can add any FX (like amp sims) without affecting what is actually being recorded.
 
As far as Sonar is concerned the only difference between record and playback is the data source for each track, either your interface or the disk.
 
There should be no reason why even the cheapest interface can't get a latency low enough to be able to track guitars with an amp sim loaded in Sonar's FX bin.

I'm going to have to try this. It does make sense to challenge one's workflow after making one or more major changes (like S1 to SONAR). I'm doing it the old way, because? That's the way I've always done it.  Maxima mea culpa!

Karyn

Larry Jones 
My real issue all along has been latency: Is the lag short enough that I'll be able to record a direct guitar while monitoring back through an amp sim in Sonar without committing the sim to my track?

There's no such thing as "committing the sim to my track".
 
Sonar takes the data being streamed from your interface and dumps it directly to disc.  It does not process it in any way.  The wav in a clip that you just recorded is always raw data.
 
Yes, you can destructively process it later, but not during record.    What you're hearing during record is a duplicate of the recorded data being fed into the channel, to which you can add any FX (like amp sims) without affecting what is actually being recorded.
 
As far as Sonar is concerned the only difference between record and playback is the data source for each track, either your interface or the disk.
 
 
There should be no reason why even the cheapest interface can't get a latency low enough to be able to track guitars with an amp sim loaded in Sonar's FX bin.

Karyn - Thanks for your clarification. I guess this wasn't clear. What I meant was that I don't like the "pure" sound of electric guitar plugged straight in. I want the fatness and character that you get from guitar amp circuitry, speakers,  etc., but I don't want to use an actual amp. And so far I haven't had to, because there are a number of good simulators out there in the form of VST plugins.
 
Starting to look for a replacement for my ancient PCI sound card led me to the concern that USB interfaces (the ones I can afford) would introduce enough latency that monitoring back through the DAW/VST would not be practical. So, two interrelated problems: 1.) finding an affordable audio interface fast enough that it wouldn't interfere with my timing as I played, and 2.) getting a decent guitar sound in the monitors to help me play my parts better, while recording the track with no FX. If problem#1 isn't solved, it begets problem #2. And the latency values I've seen (admittedly anecdotal) seem to suggest that I'd better get on with trying to solve problem #2.
 
One potential solution I've looked at is the Steinberg UR-44, which (maybe) comes with an amp sim in the hardware itself. If the sim is any good, I could use it when recording. This might work for me, but I haven't been able to figure/find out if it's possible to monitor with it and not record it. The other possibility would be to hook up something like my Line 6 Pod and record with that. So in these two scenarios, either I would definitely be "committing the sim to my track," (Pod) or I would maybe be committing the effect to the recorded track (UR-44 sim).
 
As for your statement that any interface will do the job, I hope that's true, but I suspect it's not, and I'm trying to find out one way or the other before I make a buying decision. I know I can return merchandise, but I'd rather get it right the first time. So I'm experimentally trying to create longer, measurable latencies on the system I have, so I can see just how much I can tolerate. 

Larry Jones

As for your statement that any interface will do the job, I hope that's true, but I suspect it's not, and I'm trying to find out one way or the other before I make a buying decision. I know I can return merchandise, but I'd rather get it right the first time. So I'm experimentally trying to create longer, measurable latencies on the system I have, so I can see just how much I can tolerate. 

As Jim said above, the actual latency you'll get will depend on the driver and hidden 'safety' buffers that you can't do anything about, but as long as you buy an interface specifically intended for pro or semi-pro recording (not a SoundBlaster or equivalent) then there should be no problem with latency.
 
The differences between the cheap and expensive ones are preamp quality, converter quality, number of channels, number of features, and of course...  branding...