RME Babyface Vs... others.

[Theo Verelst]

The latest Yamaha mixer range has small ones in the low price range, with possible USB audio interface (Add a "U" at the end of the name IIRC), which should be pretty good 192(!)/24 interface, and well adapted to the stage situation, as well as saving a box to bring, if all the keyboards can be mixed with it as well.

 

It's also a consideration that the outputs of the DA converter box must feel ok with the load they're posed with: maybe some mixers aren't high Z enough to make them happy, and it could be if you'd use the DA directly on DI box that is simply will not perform great. Or that certain types of signal deviations are part of the target of the solution, that's another possibility.

 

Yesterday I had read the RME specs, and it's hot on making it possible to sync to wildly varying digital clock signals. That's nice, but it's better to aim for a very stable clock, and dictate that the computer audio clock (over the USB->in) is synced with that very tightly. I don't say it doesn't, but a lot of attention (and supposedly the price) goes out to situations where the connected digital audio source can be tracked even when it's extremely unstable. Probably some people "play" with that idea, but for the use with computer music software, that seems misplaced.

 

T.

[Tusker]

I'd love to get a UAD interface but I don't have $2K sitting around for this purpose.

You and most of us. Fact is, I would like to recommend the Radial USB-Pro I am using, but I can't. The sound quality is good. The form factor is excellent. It solves a lot of problems at once for just $250. But I have a feeling it is contributing to some latency on my I7 8Gig Air. This thread has me intrigued to do some tests ...

 

Would just go stereo 1/4" into stereo mixer which would then go to FOHA.

Gotcha.

 

The right interface could allow you to leave your mixer at home, also ...

[Bobadohshe]

Bobadeath said that he has a MBP. Midi isn't something that we must worry because all the new keyboards has usd/midi and if you have an old synth, like me, you can buy with 10 euros/dollars a simple usb to midi interface...

 

Actually MIDI IS something I want on the interface because I don't want to add the USB / MIDI interface into the equation. I have one and though it works alright, it's just one more thing to plug in and I wonder that it might add latency.

 

The right interface could allow you to leave your mixer at home, also ...

 

It could but I am still harboring fears that if the computer / interface goes down, my entire rig would be toast. If I still mix everything separately I can limp through handily with audio from my CP4 and Motif XS.

[Ashville.Guru]

I read this, and thought "...who the hell buys 60 audio interfaces in less than a year....?"

 

Answer: Dave Weiser

This raises an interesting point. Almost all of us normal folks have a ridiculously tiny experience of the interface market. How many of us have purchased 2 or more interfaces over a 5 year period? And it's not like its easy to pick up second-hand experience "auditioning" dozens of interfaces with different softsynths and DAWS and platforms. Which is why recommendations based on our very anecdotal experiences with one or two interfaces doesn't make a lot of sense to me.

 

Here's an analogy. Say I've never heard software piano sounds, and my only exposure to digital piano is that one entry-mid level DP I purchased few years ago. So when I gush online about how realistic said entry-mid DP sounds, just how valuable is my opinion, really?

 

Just some food for thought, and no offense to those sharing their experiences with good intentions. Apologies for the minor thread hijack...!

 

- Guru

[Six-string-man]

Yeah, AG, but we've got to ask SOMEONE.

 

Otherwise, we're relying on the manufacturer putting the downside of their products on their adverts, and that just ain't gonna happen.

 

 

SSM

[Ashville.Guru]

Yeah, AG, but we've got to ask SOMEONE.

 

Otherwise, we're relying on the manufacturer putting the downside of their products on their ad

I'd like to draw your attention to my previous post in this thread.

 

- Guru

[Six-string-man]

Yeah, AG, but we've got to ask SOMEONE.

 

Otherwise, we're relying on the manufacturer putting the downside of their products on their ad

 

I'd like to draw your attention to my previous post in this thread.

 

- Guru

 

 

Yes, and if we hadn't asked, you wouldn't have posted it!

 

The system works

 

 

SSM

 

 

 

 

[CaptainUnderpant]

I have to stop reading posts about how much better the RME is than my Focusrite, it is making me want to spend money. Ignorance IS bliss!

 

At least I can justify not purchasing it until it is USB 3.0. My wallet better watch out once they make this upgrade.

[Ashville.Guru]

I have to stop reading posts about how much better the RME is than my Focusrite, it is making me want to spend money. Ignorance IS bliss!

For those bemoaning the cost of the RME...

http://brillianceinsight.files.wordpress.com/2014/02/card-2-24-2014-1.jpg

 

The chart that @BurningBusch posted was an old one, before driver updates to many of the devices. Klonk for the latest chart. Interesting part:

 

http://s11.postimg.org/wosffsb8j/FWBabyface.png

 

Take-home message: The $249 Focusrite Saphire Pro (FW) outperforms the $749 RME Babyface.

Other FW devices like the MBox Pro and the Presonus Firestudio are also in the same league as the Babyface.

 

In other words, you don't have to pay through your nose to get a high-performance interface - so long as you have a FireWire port.

 

- Guru

[Bobadohshe]

In other words, you don't have to pay through your nose to get a high-performance interface - so long as you have a FireWire port.

 

What if I go Firewire 800 to Thunderbolt w/ the little converter?

[Tusker]

One question is, what is performance?

 

Is it simply the number of notes one can trigger within acceptable latency tolerances? Or the degree of latency exhibited for a certain number of notes? Or do we broaden the question to include variables such as pre-amp quality and signal to noise?

 

Since our OP has a macbook pro, one question I have about dawbench is the strong steinberg flavor of the tests and relative dearth of mac based hardware manufacturers tested. Would it be accurate to say that the data we have seen thus far were mostly created using nuendo software, on windows operating systems and with windows drivers? :idk

[zephonic]

The latest Yamaha mixer range has small ones in the low price range, with possible USB audio interface (Add a "U" at the end of the name IIRC), which should be pretty good 192(!)/24 interface, and well adapted to the stage situation, as well as saving a box to bring, if all the keyboards can be mixed with it as well.

 

This is what I would do. A sax player I work with just bought that new Yamaha mixer (12-channel version) and it sounds pretty good. Their USB driver may not be the very best in the business, but I'm sure it is more than good enough.

 

In other words, you don't have to pay through your nose to get a high-performance interface - so long as you have a FireWire port.

 

What if I go Firewire 800 to Thunderbolt w/ the little converter?

 

From what I have heard chances are you'll lose buss-powered operation.

 

Take this at face value, but if you really want a dedicated audio interface for your MacBook I'd strongly recommend MOTU. Their drivers are second to none on OSX.

 

If USB 2in/2out is sufficient, I'd go with the aforementioned Yamaha mixer:

http://www.sweetwater.com/store/detail/MG10XU

http://www.sweetwater.com/store/detail/MG12XU

 

 

 

[Bobadohshe]

Take this at face value, but if you really want a dedicated audio interface for your MacBook I'd strongly recommend MOTU. Their drivers are second to none on OSX.

 

I've heard they lagged with OSX updates and that the support is second rate. Can anyone corroborate or deny? If Motu is really reliable it may be a great option for me.

[Ashville.Guru]

One question is, what is performance?

 

Is it simply the number of notes one can trigger within acceptable latency tolerances? Or the degree of latency exhibited for a certain number of notes? Or do we broaden the question to include variables such as pre-amp quality and signal to noise?

Good question. It's important to isolate the features that we're talking about. The DawBench tests say nothing about pre-amp quality, SNR, etc. Happily, they're very clear and specific about what they do reflect.

 

Higher DawBench rating: No pops/glitches even at heavy loads/low buffer sizes plus lowered latency (the two are different).

 

Would it be accurate to say that the data we have seen thus far were mostly created using nuendo software, on windows operating systems and with windows drivers? :idk

Quite accurate. The OP wanted to know what makes one interface different from another, and this is the only data available. No one with a Mac has stepped forward to repeat the tests on OSX (it's a LOT of effort!).

 

I've presented the data, and leave it to the wisdom of forumites to draw conclusions. At one extreme, you could blindly assume that the results are platform-independent. At the other extreme, you could use it as an excuse to completely ignore the data and stick to your favorite manufacturer brand.

 

Or you could adopt a sensible middle path: neither ignore nor trust the data completely. Use it as a starting point, with caveats.

 

One thing's unquestionably clear from the data, though. Brand name is a ridiculous yardstick for choosing interfaces. Take a look at the results chart again:

• The Focusrite Firewire is the best performer under $1000. The Focusrite USB is the worst of all interfaces tested...!
  
• The PCI_based MOTU 424 is one of the best performers, period. Beats even the RME Fireface. The MOTU 828Mk3 USB third from bottom.
  
• Ditto with the ESI Julia Xte (PCI) and the ESI U46SL

The pattern is unmistakable. Brand name means diddly-squat when it comes to performance. It's the technology - Firewire/USB/PCI that really matters.

 

The noteable sole exception to this pattern is RME. The only manufacturer who's consistently at the top, irrespective of tech. That's really something.

 

- Guru

[Ashville.Guru]

In other words, you don't have to pay through your nose to get a high-performance interface - so long as you have a FireWire port.

What if I go Firewire 800 to Thunderbolt w/ the little converter?

I'm cautiously optimistic, from what I've read. On paper, you should get the same performance benefits of Firewire. Caveat: it's a new product, no real-world data, though. My gut feel says it should be pretty good, especially since it's made by Apple.

 

As Zeph points out, you'll need a wall wart PS, though. Not an issue if, like me, you pre-wire everything using something like a SKB Studio Flyer case (IMHO the way to go).

 

- Guru

[Tusker]

Would it be accurate to say that the data we have seen thus far were mostly created using nuendo software, on windows operating systems and with windows drivers? :idk

Quite accurate. The OP wanted to know what makes one interface different from another, and this is the only data available. No one with a Mac has stepped forward to repeat the tests on OSX (it's a LOT of effort!).

 

Companies write better drivers for some platforms than others, hence my concern that these scores are for windows and therefore quite misleading. I agree it is a tremendous amount of work that Vin and his crew have done, and it's certainly interesting to look at.

 

However, I disagree that this is the only data available. It's the only source of a certain kind of data in a certain type of structure. There is an entire industry of publications providing detailed reviews of products, and even though it's more work, we should combine multiple sources of information to make good decisions. For example, in this review of Babyface, SOS makes comparisons (non-numerical ones!) with other hardware as well providing additional clarifying detail about headphone amp output for example. Secondly user experience, even though subjective and anecdotal is often useful and may be partly why the OP began this thread.

 

I bought a $1500 interface and a $3500 computer back when that was the cost of these tools. and I had no idea what I was doing. There was no information, so you talked to somebody who knew somebody who had a glitch free system. I would have killed for the kinds of information that dawbench and periodicals have made widely available.

[Ashville.Guru]

I would venture to add that there is an entire industry of publications providing detailed reviews of products, and even though it's more work, we should combine multiple sources of information in making decisions. None of use would choose between a Honda Accord and a Toyota Camry based solely on consumer reports benchmark scores, as useful as they may be. I think you are saying the same thing.

Absolutely. The Dawbench scores say nothing about pre-amp, transparency, signal-to-noise ratio, jitter and other indicators that make up 'pristine audio quality'. Truth be told, I don't know much about benchmarks for those, and I was hoping someone else can chime in with objective info.

 

- Guru

[Theo Verelst]

As long as audio softw. and interface manufacturers do not specify the exact product properties, nor "invent" the next "new" wave of, after reintroducing U/Li-nux, real-time OS components, it's always going to be a long string of wibble nose know-it-betters of a non-descript type of unmovableness that want to become conductor at the hip train. Or so it seems.

 

In music s.w. on the modern I platforms there so much going on that competes for IO and memory interfaces, various level of cache access, memory page management, kernel time for switchting thread and process context, that combined with a lot of self written contentions in the software, it is impossible for laymen to predict the number of hidden or clear drop-outs that come from setting certain software buffer sizes. The need for buffering audio in PC software has little tp do with the IO requirements of an interface.

 

If some of you would use an o.s. and varied audio software that walks the edges of what works and what doesn't, and reports when things aren't ok (like often on Linux), you would know that. But, admittedly, OS design and software design that has linux (r OSX variation of unix) probably requires serious undergrad univ. EE courses to really get to the bottom of the real issues.

 

Until that time, consider again that low latency as some people like to make nice tables of is in the range that it isn't *that* low that the variation in the small latency is negligible, so that I repeat my comment about it being more useful to define and implement a test which decides on the variation of the latency. Because a few milliseconds in variation is a half-wave in the normal chord range fundamental frequency, and I don't like my piano to respond that inaccurate or I don't even like play on it. Unless you play some very lame examples, or course.

 

And then the following: if you'd check out how DAC works (agreed, that's also lot easier with an undergrad EE level under your belt..), you'd find out that the reconstruction filter in a decent modern Digital to Analog converter chip, when set to "normal" and not to "short" latency, for which there is good reason, may well take about 5 mS of actual, measurable latency at 44.1kHz sampl. freq. .

 

T.

 

[CaptainUnderpant]

As long as audio softw. and interface manufacturers do not specify the exact product properties, nor "invent" the next "new" wave of, after reintroducing U/Li-nux, real-time OS components, it's always going to be a long string of wibble nose know-it-betters of a non-descript type of unmovableness that want to become conductor at the hip train. Or so it seems.

 

In music s.w. on the modern I platforms there so much going on that competes for IO and memory interfaces, various level of cache access, memory page management, kernel time for switchting thread and process context, that combined with a lot of self written contentions in the software, it is impossible for laymen to predict the number of hidden or clear drop-outs that come from setting certain software buffer sizes. The need for buffering audio in PC software has little tp do with the IO requirements of an interface.

 

If some of you would use an o.s. and varied audio software that walks the edges of what works and what doesn't, and reports when things aren't ok (like often on Linux), you would know that. But, admittedly, OS design and software design that has linux (r OSX variation of unix) probably requires serious undergrad univ. EE courses to really get to the bottom of the real issues.

 

Until that time, consider again that low latency as some people like to make nice tables of is in the range that it isn't *that* low that the variation in the small latency is negligible, so that I repeat my comment about it being more useful to define and implement a test which decides on the variation of the latency. Because a few milliseconds in variation is a half-wave in the normal chord range fundamental frequency, and I don't like my piano to respond that inaccurate or I don't even like play on it. Unless you play some very lame examples, or course.

 

And then the following: if you'd check out how DAC works (agreed, that's also lot easier with an undergrad EE level under your belt..), you'd find out that the reconstruction filter in a decent modern Digital to Analog converter chip, when set to "normal" and not to "short" latency, for which there is good reason, may well take about 5 mS of actual, measurable latency at 44.1kHz sampl. freq. .

 

T.

 

What?

[Bobadohshe]

I think I've decided upon a course of action. Because NAMM is just around the corner and the current Babyface is just a little long in the tooth, I'm going to go with the Saphire Pro for now and see what emerges at NAMM. There have to be more thunderbolt and USB 3.0 interfaces on the way. Maybe even RME will debut something new. And even if nothing new comes out, and if in a couple months I find myself limited by the Saphire, I'll upgrade and sell the Saphire. At $250 it's not a crazy investment.

[Ashville.Guru]

Looking forward to hearing your experience with it.

 

- Guru

[zephonic]

I think I've decided upon a course of action. Because NAMM is just around the corner and the current Babyface is just a little long in the tooth, I'm going to go with the Saphire Pro for now and see what emerges at NAMM. There have to be more thunderbolt and USB 3.0 interfaces on the way. Maybe even RME will debut something new. And even if nothing new comes out, and if in a couple months I find myself limited by the Saphire, I'll upgrade and sell the Saphire. At $250 it's not a crazy investment.

 

I doubt you'd go wrong with this. I have a 1st-gen SaffireLE (firewire) and it has served me well since 2007, although I don't use it nowadays. Super-dependable.

 

[Ashville.Guru]

@Bobadeath - My laptop doesn't provide bus power for Firewire either. And I'm a bit anal when it comes to wiring and setup time.

 

So I purchased on of these for $8 on eBay: Firewire Power Injector. Basically gives you a powered FW port with the aid of a DC adapter.

 

 

http://s28.postimg.org/x9tgun74t/image.jpg

 

The wall-wart for the interface doesn't bother me since I have a power strip velcroed to the same case which has my laptop mounted. Everything's pre-wired, so I have only one FireWire and one USB connection to make during setup. My setup time is less than that of most hardware rigs (*cough* Kronos *cough*).

Hope this helps.

 

- Guru

[Bobadohshe]

^^^ That is a curious contraption. We'll see how crazy I go having to deal with the Wall Wart time after time.

 

My Saffire Pro 14 just arrived today during the Chargers game. Sunday delivery via USPS. I love Amazon Prime.

[Reezekeys]

Please let us know your experiences & opinion on it. I'm seriously looking at either the Saffire Pro 14 or the MOTU Microbook II.

 

Curiously there was a Saffire Pro 14 with built-in DSP for on-board reverb and eq but it looks like it's been discontinued.

[Ashville.Guru]

Please let us know your experiences & opinion on it. I'm seriously looking at either the Saffire Pro 14 or the MOTU Microbook II.

I was really excited about the Microbook II when it came out, I think I've posted about it here. The form factor and design are just perfect for mounting on top of a board. Nice LED VU display, big handy knob like my TC Konnekt.

 

But MOTU devices perform surprisingly poorly on the DawBench tests. On several counts:

• MOTU is the only brand who's Firewire device performs worse than other USB devices.
  
• Their USB devices are in the bottom quartile of everything tested.
  
• Relatively unknown brands like the ESI Maya, Midas Venice and the Audient iD22 outperform the MOTU
  

This is surprising, given the word-of-mouth reputation the MOTU brand seems to enjoy. Of course, these tests were on Windows, and it's *possible* that on OSX, the results would be dramatically reversed. I haven't seen any evidence to that effect, though.

 

- Guru

[zephonic]

@Ashville:

 

While the DAWbench test is very useful, it is nowhere near definitive:

 

 

- Windows 7 only

- The interface test is done with Cubase only (Logic/Mainstage or ProTools is a VERY different kettle of fish)

- System specific results (all published results were done on Vin's personal X58 test machine)

 

I ran the original MOTU PCI324/2408 on a Windows 98SE system, and it was hell. At my former job we ran a PCI-X424/24IO system on a G5 that was on 24/7 and it was heaven. We had another G5 with a 896HD (FW) and again, flawless performance 24/7.

 

I can assure you that their FW devices work excellently on OSX.

[Bobadohshe]

Been up and running since last night with the Saffire. I haven't been able to put it through its paces extensively yet --- because I spent about 4 hours trying to get to the bottom of this super obnoxious issue:

 

Software instruments responding monophonically on MIDI Channels higher than 1

 

I thought it was the Saffire's MIDI. I then thought it was my long in the tooth MOTIF XS. After systematically and slowly eliminating all of these as possibilities I knew it HAD to be Mainstage itself. So some Googling got me to the above solution.

[Ashville.Guru]

While the DAWbench test is very useful, it is nowhere near definitive:

 

- Windows 7 only

- The interface test is done with Cubase only (Logic/Mainstage or ProTools is a VERY different kettle of fish)

- System specific results (all published results were done on Vin's personal X58 test machine)

Let me reiterate: Yes, I fully acknowledge the possibility that the DawBench guys just so happened to pick that singular OS/DAW/hardware combination, wherein MOTU performs poorly, and almost every other manufacturer does not.

 

- Guru

[Tusker]

Yes, I fully acknowledge the possibility that the DawBench guys just so happened to pick that singular OS/DAW/hardware combination, wherein MOTU performs poorly, and almost every other manufacturer does not.

That's a fair comment. It helps to keep an industry structure perspective. Vin and crew didn't pick the environment for the test out of a hat. They grew up in Nuendo/Cubase Windows, so that's what they test with. By contrast, MOTU grew up in the Mac environment, and that's what they became good at. I used MOTU's Performer software on a Mac system in 1986 ...

 

SOS indirectly references MOTU's Windows problems in this review of the track 16 interface.

 

"As is often the case with Windows interfaces, buffer sizes and sync settings are made using a separate utility called MOTU Audio Console. The Track 16's Windows ASIO driver offers a choice of buffer sizes from 1024 samples all the way down to 64. However, there seems to be a non-documented safety buffer at work as well. Using Oblique Audio's Round Trip Latency utility and the Track 16's USB drivers, I measured the actual input-to-output latency as being just over 900 samples (20.5ms) with a 256-sample buffer size. At the largest 1024-sample setting I measured 3128 samples (70ms), while at the 128-sample setting, I recorded 522 samples of latency (11.9ms). This suggests that the true latency is almost twice what it theoretically should be at low buffer sizes not an uncommon state of affairs. I got marginally better results when I tested the Track 16's Firewire drivers: 863 samples of latency (19.9ms) at a buffer size of 256 samples and 479 samples (10.9ms) at 128.

 

Perhaps as a result of this padding, the Track 16 performed well in Vin Curigliano's DAWbench test. I was able to run 90 instances of Tube-Tech's Classic Channel at the 256-sample setting, and 86 instances at the 128-sample setting. This was exactly the same on Firewire and USB. By comparison, Presonus's AudioBox VSL1818 would only let me run 56 instances at 256 samples, and wouldn't play back audio at all at 128 samples but it should be remembered that the AudioBox's buffer settings are much more accurately reflected in the actual round-trip latency it achieves (15.6ms at 256 samples against the Track 16's 20.5ms)."

 

I imagine that if Apogee went cross-platform, their Windows performance might be inferior to their OS X performance also.