Maschine Plus latency with audio over USB

jolly Dodger

Is anyone getting bad 20+ms latency when using audio over usb?

I have a Maschine+ and a Nord Drum, the Maschine+ is 19ms behind the Nord, when both are sequenced externally from the same device.

The signal path is:

• Maschine+ -> usb -> babyface(CC mode) -> analogue -> mixer
• Nord Drum -> analog -> mixer

If I go Maschine+ -> analog -> mixer then the offset between them is ok. i.e. it's too small to hear. I tested going out of the baby face with ADAT, and the issue persists. So the delay is between the m+ and the babyface

The RME gear should be less than 1.5ms. The m+ is 5.8ms. So I should be getting no more than 9ms, but I have 19ms, and it's clearly audible.

PK The DJ

I'm not quite sure I understand why you're involving the Babyface.

If you're using it just as an alternative audio output for the M+ then my question would be - why?

jolly Dodger

https://community.native-instruments.com/discussion/comment/133838#Comment_133838

Sorry I failed to say explain that I'm using the ADAT out of the Babyface. But for testing/explanation purposes I used the above example. But to answer your question why...

• I can then mix/fx outside of the m+ with 8 rather than 4 channels.
• manage cpu in m+, as greedy fx is external.
• bus sends with external gear

DeepThumb

Analog is the solution for many digital issues, particularly when looking at latency.😉

I would use the analog inputs of the Babyface or - as PK The DJ already mentioned - connect M+' line level outputs directly with your analog mixer.

jolly Dodger

https://community.native-instruments.com/discussion/comment/133845#Comment_133845

Thanks deepThumb, yeah latency in the box is often a mare.

In this situation I expected the latency to be better when going out of the USB on the M+ (on the Force and Mpc live, using USB rather than analog reduces latency). My previous experience has been that if I have one digital to analog conversion at the end of my audio chain, I normally maintain fidelity and reduce latency.

Looking for conformation if latency is worse or better for you when using USB out rather than analog out?

Kubrak

I do not have M+. Have you set the audiobuffer for Babyface in M+? I am not sure, if it is possible.... Maybe it is huge and so introducing big latency. Or simply CC in M+ does not work very well.

PK The DJ

@jolly Dodger Thanks for the details.

Next question 😁 - if you need/want more channels and to use "greedy FX" then why use the M+ standalone? Put it in controller mode and use Maschine software on your computer.

DeepThumb

https://community.native-instruments.com/discussion/comment/133849#Comment_133849

Sorry, out of several known reasons, I have not or will ever use USB for audio or MIDI. Therefore I can't compare it. However with analog signal chains I have not experienced latency issues at all.

Kubrak

But OP needs digital out. Probably to be processed by digital effect unit....

darkwaves

I have M+ connected to a motu 8a. Audio is sent to my desktop interface (RME UCX) via adat. MIDI is handled via a iConnectivity MioXM (connected to my computer over the network). My system is at 44.1k and I normally have the buffer on my desktop interface at 128 samples. I moved it down to 48 for the following:

I'm basically sending that midi track to both the maschine software (as a plugin) and to M+. They're playing the same sample (seen in the image) and I'm recording to audio. I struggled to find any samples in the maschine library that were trimmed well, so this is what you get.

So I'm getting about 12 milliseconds of latency if I attempt this kind of workflow.

M+ reports a total latency of 8.7. My DAW says 3.5. Seems to be in the ballpark I guess. I've never looked too deeply into it simply because I don't use M+ in this way. It certainly wouldn't be workable if trying to perform live.

I'm too lazy to do it now, but I'm curious how noticeable the delay would be if monitoring directly from my interface.

jolly Dodger

Your a legend MrDarkwaves... ta++

So yep something wrong on my side, maybe the Babyface(gen1) in class-compliant mode is frack'd - it's a closed box with only a kHz menu.

But yah, 12ms is still a lot, anything over 7 is likely to mess with ones timing when playing live. I assumed that the reported m+ latency is a round trip in & back out, so that might explain how the analog out is only 3.5ms.

I also wondering if the m+ usb implementation or some other issues are contributing to the 12ms when going over USB for you.. i.e. difference between midi paths, or the ADAT to RME-UCX to computer audio paths.

tetsuneko

Maschine+ is essentially a linux computer, and any class compliant audio device might perform very differently in it than with a mac or a pc. You are basically opening a can of worms here. Therefore my advice is to use analog outs whenever possible, if low latency performance is vital in your standalone workflow when it comes to Maschine+

Kubrak

@jolly Dodger I guess, you have to set buffer size (and so the latency) in M+. There is the driver that operates Babyface. If you set it low latency will be low, but there is the danger of dropouts..... M+ has rather weak CPU, so everything depends on how CPU demanding project M+ runs in your case....

I am not M+ user, so I only guess, it is possible to set audio buffer size somewhere in M+....

tetsuneko

https://community.native-instruments.com/discussion/comment/133989#Comment_133989

there is no way to adjust buffer size in Maschine+ standalone, unfortunately.

jolly Dodger

https://community.native-instruments.com/discussion/comment/133987#Comment_133987

As you said can of worms, i.e. when there is a firmware update on a groove-box, the compatibility/glitch-factor of a different usb audio interfaces often changes somewhat randomly. Dev'ing usb-audio must be a real challenge. ADAT I/O is a safer bet for groove-boxes.

FYI: I've done a measurement of my latency (i.e. not just my ears), and it's 15.5ms not 20ms as previously stated.

Can anyone else check please their latency over their usb interface, so we can verify if the M+ (v1.5.1.0) over USB is practical for finger drumming / live performance?

reffahcs

Seems like it would be way easier to use the Babyface analog inputs then route them to ADAT out in Totalmix. I'm not too familiar with the Babyface so I don't know if it will save routing info for standalone use (ie without needing totalmix running). I know the higher end devices like the UFX will.