Question for RME HDSP-TotalMix users

I have an HDSP AES32 card. After reading the manual and trying to work with it, I do not want to use the TotalMix mixer. If you are using TotalMix, what is your reason? It seems like a big hassle compared to using everything within Cubase like the Control Room and the Cubase Mixer. Was there a “latency point” where you had to use it? If so, what was it?

It’s very helpful, if you need to send different mixes to the different outputs, for example for different recording rooms. You can use “Studios” for this, in the Cubase, but TotalMix looks more flexible, for me.


I can see this if you need more than 4 cue mixes. On the other hand, I find the Cubase Control Room Mixer much more intuitive, particularly for things like Inserting Reverb for Cue mixes. I just find it more intuitive, but I don’t need lots of cue mixes. So, I can see its limitation. I wonder if there’s a workaround if you need more than 4?

Um…you’ve lost me–you don’t know why someone would use a hardware mixer rather than the software for cue mixing? Because they arent using an analog mixer to provide the cue mixes.

I’ve gotten Cubase down to 64 samples…unacceptable–not that my machine could drop into a big project at that buffer level anyway, but I know a newer one will.

But, then, I also think converter latency is off putting though generally acceptable, so…I keep a little analog mixer for cue feeds of overdubs.

Also, while I don’t use RME…I can assure you that using my Echo’s mixer IS consistently ADA latency only…using “direct monitoring”, which SHOULD do exactly the same thing does not. Something about the bigger and more complex the project gets, the latency varies…which is utterly unacceptable. Is that Steiny or Echo’s drivers? Don’t know. But, it’s a frustrating thing. So, someone cutting a whole band full of feeds? If it can’t handle properly DM for overdubs, I can’t imagine it does better with a complicated cue mix.

If you can do everything you need to within Cubase and you’re happy with that then I see no reason to bother with TotalMix other than to initially set up your I/O, but I don’t think of TotalMix as a mixer (despite the name), and certainly not as a replacement for an analogue mixer – I regard it more as a powerful digital audio router. In that funtion there is nothing to compare with it in terms of flexibility and power, and it’s when you get outside of Cubase that you begin to appreciate that, for example incorporating external hardware or audio streams from other applications.

Thanks for your response popmann. My studio is being reconstructed, and I wondered about latency issues. . So, maybe I’ll have to use TotalMix for latency free monitoring. I have a Lynx Aurora 16 with the RME HDSP AES 32, which is pretty good hardware.

Thanks for your response Mr. Soundman.

I am planning on using an API 2500 on the Master bus, with a pair of of Distresssor’s on the drum bus. I just bought the gear and am renovating my studio space. So, that’s why I’m asking these questions now. Are you saying I’ll have to use TotalMix if I want acceptable latency for cue mixes and monitoring?

The signal path though TotalMix is digital (obviously) so there will always be some latency because of the A/D-D/A conversions, but you also have to consider that TotalMix can bypass the CPU completely as well as allowing e.g. to add reverb via an external processor to a vocal monitor mix with no drain on CPU. Going via TotalMix for monitoring as opposed to going through Cubase will reduce latency, but if you want no latency, then you need an analogue monitor mix.

TotalMix is king if you need to run a big session with several different headphone mixes. Incredibly flexible. Also to incorporate outboard hardware.

On Mac, there is no Direct Monitoring, and therefore no way to get latency free headphone mixes through Cubase. In my view, even very low latency is unacceptable, and even if you start off a project with low latency, it always creeps up - VSTis, FX everywhere - what if you need to do a vocal overdub just when you’re all set up to mix?

Much better to lose the latency altogether, either with hardware (analogue) monitoring, or TotalMix (or CueMix if you have MOTU).

Although there is a tiny amount of latency because of the AD/DA, it really is too small to notice, and it stays constant, which Cubase latency does not.

Thanks for this great thread!

Can you explain please - if there are no reverbs or VSTi’s or anything like that activated in Cubase - what latency are you referring to going into Cubase and then straight out again? I guess from what you are saying, it is not instantaneous, but I can’t quite understand why not …?

Thanks much for any further teaching!

Thanks again MrSoundman.

I’ve been thinking about options. I have a Mackie 1604 VLz mixer. So, could I have my vocalist sing through a mic into my vintech preamp to the A/D and then have her monitor the D/A of the mix on two channels of the 1604- and her performance on 1 channel from the 1604’s- from the headphone jack of the 1604? And could I add some reverb on her playback channel to give her some warmth, all in real time? Would that be less latency and more efficient than going through TotalMix and a headphone amp? Unfortunately, the singer would be hearing the Vintech output being altered by the 1604’s preamps though.

Although there is a tiny amount of latency because of the AD/DA, it really is too small to notice, and it stays constant, which Cubase latency does not.

Not to mention the noticeable CPU bump Control Room adds when active.

The latency caused by the buffer- / samplerate settings of your soundcard.

Please keep in mind, that some VST Plugins are producing an additional latency to the ASIO latency.
You can sort this out under Devices/ Plugin Information.

Some VST Plugins like the Fluxx VST Plugins can produce more than 100ms additional latency.

But probably not, if


Sure thinkingcap…

that is why I wrote this additional info…if you wish to add some VST Plugins for the headphone mix…


Cubase (or any software) aside for the moment, let’s imagine you have standalone A/D and D/A converters with say, AES3 or S/PDIF, and you connect the output of the A/D to the input of the D/A; a signal applied at the analogue input of the A/D will appear not instantaneously, but after a short (read: very short!) delay at the output of the D/A. This delay is primarily made up of the time the A/D conversion process takes, plus the time it takes to transmit the digitized information between the two units, plus the time for the D/A conversion process, and is directly related to the clock rate, but can be regarded as constant.

Many audio interfaces have the A/D and D/A all in one box or on the same card, so there’s the possibility of connecting the analogue input signal from before the A/D process directly to the analogue output – this is the only true zero-latency form of direct monitoring possible and is the equivalent of using an external mixer, but you lose a lot of flexibility. I personally would not be calling the lawyers if a manufacturer claimed this was also “zero-latency direct monitoring”, but it still pays to be aware of the phenomen as it can lead to issues with phasing even if the delays are not perceptible to the human ear.

In the case of RME TotalMix, the signals you “see” have already been digitized, and the signal routing takes place completely in the digital domain, which is precisely what gives you the unbridled flexibility. There will be some more delay once the routed digital signal reaches its destination D/A converter, and then the converter will add some more. Nevertheless, the combined delay is still miniscule compared to the delay added by software processing, and may be acceptable, but cannot be eliminated from any digital audio system.

For a further authoratative discussion of the wider system implications, have a read of: Living With Latency by Sam Inglis in Sound On Sound. For what it’s worth, once you add your DAW software into the mix, the only way to know exactly what latency you have with any specific configuration (as defined by the delay between the input signal and the signal that appears at the output) is to actually measure it (yes, with a 'scope!), and not to trust what any software tells you!

For foldback to one artist the Mackie would be fine.

You would need to put the output of the mic preamp into the Mackie, for foldback, and into the AD of the soundcard to record it in Cubase. What you shouldn’t do is monitor the singer through Cubase. If you do that you will hear the latency as a delay.

If you want 4 or more separate headphone mixes to, say, a band, Totalmix is much more powerful and flexible. Don’t worry about the nominal AD/DA latency with TotalMix, it isn’t an issue.

Does the RME TotalMix pretty much mean one can skip the Cubase 6 Control Room without missing any features?

Thanks again for your response MrSoundman.

The 1604 can have 4 submixes. So, couldn’t I submix 4 artists and send out to a headphone amp with 4 headphones?
I say this because I already have the 1604 and the headphone amp. Though I do agree, if you have more than 4 band members, TotalMix has tons of submixing options.

It’s so simple with an analog mixer.

As for Totalmix, the manual leaves a lot to be desired in trying to figure out the routing. This has a lot to do with the interface. If their interface would have emulated an analog mixer, which their routing emulates, they would have hit the ball out of the park.