What's new

Low latency of Apollo vs. RME card on Mac only.

Headlands

Senior Member
This is ONLY for Mac (or Hackintosh)-- Windows users need not respond unless you have direct Mac experience with what I'm about to ask:

How would you compare RME devices to UAD Apollo as far as low latency performance and drivers on Macs or Hackintoshes? I only want to know about that -- no comments needed on mic pres, TotalMix vs Console or anything else. This is for a Hackintosh, and it would be a PCIe RME card vs. an Apollo 8 2nd gen TB2. I did a search but most comparisons are for Windows, and I'd like to the most up-to-date opinions.

I needed to post where I'd get more views and opinions than on the Computers/Hardware forum, so that's why I'm posting here.
 
This is year old data (nearly that anyway), and oddly it was a Hackintosh, and even stranger, the debate was between an Apollo Twin (TB) and an RME, can't recall the exact model, not sure it matters.

There was a difference in latency, but it was subtle, not at all what I expected. I probably would not make this my primary point for decision making. Or at least I would not have back then.

By the numbers the RME did report lower latency, and it was stable at smaller buffer sizes, but it just wasn't an issue. I could comfortably play in to a track, MIDI or Audio, with no problems.

It has been my experience that RME is a little better at maintaining their drivers, so if I did not want access to the UA Unison plugins I'd use RME. But I do love the Unison plugins, so it looks like I'm sticking with a UA interface for the now.

You almost certainly know this, but with the exception of the Unison plugins you do not need a UA audio interface to run the rest. I mention this because I know a couple folks that were unaware of the other options for UA DSP plugins.
 
This is year old data (nearly that anyway), and oddly it was a Hackintosh, and even stranger, the debate was between an Apollo Twin (TB) and an RME, can't recall the exact model, not sure it matters.

There was a difference in latency, but it was subtle, not at all what I expected. I probably would not make this my primary point for decision making. Or at least I would not have back then.

By the numbers the RME did report lower latency, and it was stable at smaller buffer sizes, but it just wasn't an issue. I could comfortably play in to a track, MIDI or Audio, with no problems.

It has been my experience that RME is a little better at maintaining their drivers, so if I did not want access to the UA Unison plugins I'd use RME. But I do love the Unison plugins, so it looks like I'm sticking with a UA interface for the now.

You almost certainly know this, but with the exception of the Unison plugins you do not need a UA audio interface to run the rest. I mention this because I know a couple folks that were unaware of the other options for UA DSP plugins.

Thank you - great information.

I'm almost at the point where I don't need UAD plugins anymore (I've gotten there on purpose), so that wouldn't be an issue. I already have the Apollo 8 and was considering an RME card for my new Windows machine (would run it through the Apollo for its converters and Unison), but now have decided to make it a Hackintosh (I came from Mac before the new Windows machine, and I've decided I much prefer Mac after over a month on Window) and was curious since the RMEs had such stellar ratings and reports for stability and latency.

On Windows the Apollo wasn't good at low latencies, which is an acknowledged issue by users, so I was going to order the RME because it's supposed to be far better performance-wise. Now that I'll be back on Mac I'm curious about its performance vs. the Apollo.
 
I just sold my RME Fireface 400 after using it for over ten years and replaced it with an Apollo x6. I do really like the UAD plugins and have been using these since the Mackie days. My experience after two months with the Apollo is that, if not for the much improved latency with using UAD plugins, I'd happily go back to RME. For some reason, RME feels a bit more solid, however, I can't say I've had any problems with the Apollo. In fact, the converters on the Apollo are much improved and there is a clarity that I'm now hearing in my room (note that we are comparing two devices from different decades). What bugs me is that I often use different audio related apps at the same time set at different sample rates, for example RX and QT might be 44.1, and Logic Pro or Soundminer, etc... at 48kHz or 96kHz - all open on my desktop at the same time. When I go between applications, it causes the interface to hiccup and physically switch sample rates (it even makes a little sound on the box). I don't remember having this experience with RME but I worry that this physical switching could cause problems after a number of years. I feel like my old RME box will literally last forever. Just to reiterate, I do like the Apollo though and plan on keeping it at this time... Additionally, I am curious why RME still sells mostly USB. I was certainly influenced by UAD's move to update all their interfaces to Thunderbolt 3. Not sure if anyone has thoughts on this, but why is RME mostly committed to USB? For cross platform compatibility? Is it even an issue from a bandwidth perspective?
 
You said it: cross-platform compatibility. RME does offer Thunderbolt 2 on the UFX+ but really, with USB3 the bandwidth is not an issue. And while latency might be seen as an issue, RME’s USB drivers are so good that the difference seems negligible.
 
I just sold my RME Fireface 400 after using it for over ten years and replaced it with an Apollo x6. I do really like the UAD plugins and have been using these since the Mackie days. My experience after two months with the Apollo is that, if not for the much improved latency with using UAD plugins, I'd happily go back to RME. For some reason, RME feels a bit more solid, however, I can't say I've had any problems with the Apollo. In fact, the converters on the Apollo are much improved and there is a clarity that I'm now hearing in my room (note that we are comparing two devices from different decades). What bugs me is that I often use different audio related apps at the same time set at different sample rates, for example RX and QT might be 44.1, and Logic Pro or Soundminer, etc... at 48kHz or 96kHz - all open on my desktop at the same time. When I go between applications, it causes the interface to hiccup and physically switch sample rates (it even makes a little sound on the box). I don't remember having this experience with RME but I worry that this physical switching could cause problems after a number of years. I feel like my old RME box will literally last forever. Just to reiterate, I do like the Apollo though and plan on keeping it at this time... Additionally, I am curious why RME still sells mostly USB. I was certainly influenced by UAD's move to update all their interfaces to Thunderbolt 3. Not sure if anyone has thoughts on this, but why is RME mostly committed to USB? For cross platform compatibility? Is it even an issue from a bandwidth perspective?

How does the low-latency performance compare between the two for you?
 
I actually have not noticed a difference so far, latency-wise. I'm running at a 256 buffer setting. This has been fine for large midi-template orchestral programming. I haven't had an opportunity to record live acoustic instruments yet (besides voice and hardware synths, where I didn't hear any latency at all).
 
Sorry to be so cranky, but why the f is latency the only thing people talk about?

Well, I know the answer - because of marketing disinformation. :( It's a disservice to musicians.

But sound quality is the most important consideration. That's what people should be talking about.
 
Sorry to be so cranky, but why the f is latency the only thing people talk about?

Well, I know the answer - because of marketing disinformation. :( It's a disservice to musicians.

But sound quality is the most important consideration. That's what people should be talking about.

Others have different concerns -- totally up to the individual. I would be using the Apollo converters either way here, and its sound quality is great. Sound quality is hugely important for me, and latency is also incredibly important as well. But PLEASE let's not continue with this because it's irrelevant to my original post.
 
I actually have not noticed a difference so far, latency-wise. I'm running at a 256 buffer setting. This has been fine for large midi-template orchestral programming. I haven't had an opportunity to record live acoustic instruments yet (besides voice and hardware synths, where I didn't hear any latency at all).

Ok cool, good to know. I might just stick with the Apollo then.
 
Others have different concerns -- totally up to the individual. I would be using the Apollo converters either way here, and its sound quality is great. Sound quality is hugely important for me, and latency is also incredibly important as well. But PLEASE let's not continue with this because it's irrelevant to my original post.

Nope, I'm not going to comply. It's totally relevant to your original post. Sorry, you've been sent on a wild goose chase by silly marketing.

Mike_solar is running his sequencer at a 256-sample buffer, which in Logic is roughly 7ms output (at regular sample rates). I usually run Logic at a 64-sample buffer, or 3.3ms - on my 20-year-old FireWire interface (Metric Halo 2882) and 11-year-old Mac.

The difference in latency between high-quality interfaces is very small these days. What makes the difference is your computer and what you're running (i.e. how much you're stressing it, whether you need to raise your buffer). And most people - not everyone, most people - won't feel the difference even then.
 
Nope, I'm not going to comply. It's totally relevant to your original post. Sorry, you've been sent on a wild goose chase by silly marketing.

Mike_solar is running his sequencer at a 256-sample buffer, which in Logic is roughly 7ms output (at regular sample rates). I usually run Logic at a 64-sample buffer, or 3.3ms - on my 20-year-old FireWire interface (Metric Halo 2882) and 11-year-old Mac.

The difference in latency between high-quality interfaces is very small these days. What makes the difference is your computer and what you're running (i.e. how much you're stressing it, whether you need to raise your buffer). And most people - not everyone, most people - won't feel the difference even then.

Now that's useful information - thank you! Your first post didn't go into these details. :)
 
MIDI takes what, 10ms (off the top of my head)? to get from the keyboard into the sequencer and trigger the V.I. (Not only is that a guess, but the actual point in the key travel where the note-on is generated is hard to know - although I"m sure someone here who knows how they work does know.)

The main thing is that most people... well, that's not scientific - I, at least - can tell the difference between the latency with a 128- and 256-sample buffer. But it's pretty subtle, and only when A/B-ing. And only with percussive sounds, of course.

For perspective, I'm not a real actual keyboard keyboard player, but I do play enough keys to use them as a tool.

Look, latency definitely is a real factor. It's not like one shouldn't care at all. A 512-sample buffer is enough to be annoying if you have to work like that. The problem is what I've been ranting about: people have been trained to ask about an interface's latency before they ask how many ins and outs it has!
 
By the way, I've only been talking about keyboards. But I have no latency issues playing percussion parts on my BopPad percussion controller, never mind my EWI.

Oh, and we're talking about playing V.I.s. If you're recording audio, direct monitoring is your friend.
 
Nick,
Thank you for clarity on midi. My current I/O for Logic Pro is 128 I/O buffer on orchestrated V.I. samples. Anything higher and I have difficulty added more tracks that are in sync (I rarely play to a click track). I have purchased a new Mac Pro which will make my existing Zoom TAC2R thunderbolt 1 interface useless (no drivers for Catalina). Do you believe the interface (usb or thunderbolt) makes a difference when playing samples only?
 
Not sure, but my engineer says his Avid HDX I/O blows them both away, legit mac and hackintosh.

He’s correct. However, if you want latency free Unity preamps to suddenly record a vocal or Guitar part whilst in session with a high buffer, and using the best plugins in the world, it’s hard to ignore the Apollo. Especially if you hook one up in conjunction with a HDX system.
 
Top Bottom